COSTE53 ConferenceWarsaw Presentations All PDF

Quality Control
for Wood
and Wood Products
October
15th/17th, 2007
Warsaw
Poland
COST E53
The First Conference
Faculty of Wood Technology

Warsaw University of Life Sciences
Action E 53
THE FIRST CONFERENCE
Quality Control
for Wood
and Wood Products
October
15th/17th, 2007
Warsaw, Poland
Edited by
Marek Grześkiewicz
The organizers of the first COST E53 conference “Quality Control for Wood and Wood Products”
and the editor of the proceedings, would like to thank all the authors of presentations and papers.
We hope that the meeting in Warsaw will be informative and interesting for the Conference
participants, as well useful for their future work and cooperation.
Organizers in Warsaw:
Ewa Dobrowolska
Marek Grześkiewicz
COST E 53 Quality Control for Wood and Wood Products
Conference 15-17 October 2007 Warsaw
Table of contents
Relevance of wind-induced compression failures regarding 1

bending strength and stiffness of spruce structural timber
Martin Arnold, René Steiger
Development of quality management in Polish wood industry 7
Jan Bagiński, Łukasz Babuśka, Marcin Klimczewski
Bark measurements with X-ray technology 13
Rafael Baumgartner, Franka Brüchert, Jörg Staudenmaier, and Udo Hans Sauter
Dimensional and shape quality control. Modern equipment 17
Camelia Boieriu, LuminiŃa Brenci
Ultrasonic evaluation of mechanical properties of wood 21
in standing trees
Witold Dzbeński, Tomasz Wiktorski
Measurement of moisture content profiles in coated and uncoated 27
Scots Pine using Magnetic Resonance Imaging
Jan Ekstedt, Anders Rosenkilde, Stéphane Hameury, Magdalena Sterley , Henrik Berglind
Assessment of early stage fungal decay of wood by FT-NIR- 33
-spectroscopy
Karin Fackler, Barbara Hinterstoisser, Manfred Schwanninger, Cornelia Gradinger,
Ewald Srebotnik, Kurt Messner
Fingerprint traceability of sawn products using x-ray log scanning 39
and sawn timber surface scanning
Jens Flodin, Johan Oja, Anders Grönlund
Influence of drying technique and process conditions on the drying 43
quality of beech wood (Fagus silvatica L.)
Željko Gorišek, Aleš Straže
Mass customized production and design process for the parquet 49
flooring industry
Hans Häuslmayer, Manfred Gronalt, Alfred Teischinger
Tensile proof loading to assure quality of finger-jointed 55
structural timber
Georg Jeitler, Gerhard Schickhofer
Effects of heat treatment of small clearwood samples 63
on equilibrium moisture content and deformation
Björn Källander, Pierre Landel
Modeling the properties of strength graded timber material 69
Jochen Köhler, Markus K. Sandomeer
Wood and wood products quality evaluation in Slovakia 75

Rastislav Lagaňa, Stanislav Kurjatko, Alena Rohanová, Ján Dubovský, Anna Šuriková
Evaluation of the French oak for structural use 81
Jean Denis Lanvin, Didier Reuling, Costrel Y , Ducerf J
Indisputable key – a research and demonstration project 87
aiming at an improved quality through traceability systems
Anders Lycken, Jarl-Gunnar Salin
Variation in yield of strength graded timber due to board origin 93
Audun Øvrum
Aspects of timber quality in the United Kingdom 97
Tim Reynolds
Variables affecting the performance of British grown Sitka spruce 103
Tim Reynolds
MC based or time based drying schedules? 109
Calculation of the influence on final MC variation
Jarl-Gunnar Salin
Approach for an efficient control of grading machine settings 115
Markus K. Sandomeer, Jochen Köhler
Timber grading machine using multivariate parameters 167*
based on ultrasonic and density measurement
Jean-Luc Sandoz, Yann Benoit
Bending strength and stiffness of aspen sawn timber 121
Harald Säll, Bo Källsner and Anders Olsson
Visual grading in wood flooring sector in Greece 127
Michalis Skarvelis, Antonios Papadopoulos
Improved log sorting combining X-ray and 3D scanning- 133
- a preliminary study
Johan Skog, Johan Oja
Development of twist on boards of Norway spruce 141
during kiln drying
Aleš Straže, Robert Kliger, M. Johansson, Ž. Gorišek
Online wood industries database 149
László Szalai, János Boros, Ákos Bausz
Quality assessment of old-growth Scots pine stands in Poland 153
Stefan Tarasiuk, Grzegorz Jednoralski, Krzysztof Krajewski
Pre-grading of sawn timber in green condition 161
Helene Unterwieser, Gerhard Schickhofer
* - paper was send to organizers in Warsaw 12.10.2007. when the other papers were printed
15th - 17th October 2007, Warsaw
COST E 53 Conference - Quality Control for Wood and Wood Products
Relevance of wind-induced compression failures regarding bending strength and

stiffness of Spruce structural timber
Martin Arnold, René Steiger
Empa, Swiss Federal Laboratories for Materials Testing and Research
Wood Laboratory
CH-8600 Dübendorf, Switzerland
martin.arnold@empa.ch rene.steiger@empa.ch
Keywords: compression failures, grading, storm damage, structural timber, strength class
ABSTRACT
Compression failures (CF) are a well-known phenomenon mainly in lower density softwood trees exposed to frequent
and/or strong winds. They are induced by large stem deflections, which cause a buckling of the wood fibres on the leeward
side of the stems after exceeding the axial compressive strength of the wood. The distorted fibres are weak points in the
wood structure, which can lead to brittle fractures already at a relatively low stress level in bending or tension. According
to many grading rules, CF are therefore not permitted in construction timber.
In an extensive research project started after the violent winter storm 'Lothar' (1999), open questions regarding the extent
and location, the causes, the detection and the consequences of compression failures have been studied. 30 blown-down,
broken or standing mature spruce (Picea abies) trees were harvested from a heavily storm damaged stand and converted
into construction timber. During the visual grading procedures acc. to the Swiss standard SIA 265/1 the sawn timber was
scrutinized for compression failures. Afterwards MOE and MOR of 563 beams (2200mm x 110mm x 95mm) were
determined in bending tests acc. to EN 408. A statistically significant reduction of the moduli of rupture and elasticity
in bending was found. The characteristic values of the mechanical properties however still exceeded the limits for the
strength classes of visually graded timber (acc. to SIA 265/1). Since posing a severe safety risk because of the very brittle
fracture behaviour it is nevertheless recommended not to use timber with detected CF in load bearing structures especially
when members are subjected to tension or bending stresses.
INTRODUCTION
Defects in the wood structure in the form of buckled cell walls of the wood fibres are a well-known
'natural' phenomenon and are observed quite frequently in lower density softwoods such as spruce
(Picea abies) [4, 9, 11, 19, 20]. These co-called compression failures (CF) may be wind-induced in the
standing trees, if the stems are bent so much by frequent or strong winds that the proportionality limit
of the wood in axial compression is locally exceeded on the inward (leeward) side of the bow. The size
of CF may range from small deformations in the cell wall to wide bands of several millimetres in
width, which can affect more than half of the stem's cross-section. The geometric structures of CF are
complex, have more or less fuzzy boundaries and appear in a broad range of intensities. CF are usually
difficult to detect, particularly in rough sawn timber. This poses a problem regarding processing and
grading of such timber.
The distorted and damaged fibres can lead to brittle fractures in structural timber already at a relatively
low stress in bending or tension. CF therefore are regarded as unwanted structural defects and
particularly after heavy storm damages in the forests, questions regarding their influence on the
mechanical properties of timber from the salvaged trees arise anew.
A reduction of the mechanical properties (mainly in bending and tension) at the fibre level [18] and in
small clear wood specimens is generally acknowledged [9, 15], but the effect is less clear with
structural timber. There the effect of CF may be confounded by the presence of other defects such as
knots or grain deviations [7, 8]. However, because of their potential safety risk, many grading
standards explicitly [14] or implicitly ('mechanical damages') exclude CF from timber elements in load
bearing structures. The presented work is part of an extensive research project in Switzerland started
after the violent winter storm 'Lothar' in December 1999, aiming to collect more information regarding
the extent and location, the causes, the detection, and the consequences of wind-induced CF [2].
1
MATERIALS AND METHODS [1]

Sample material
The sample material was taken from 30 fully-grown spruce (Picea abies) trees of heights between 35
and 43 m and diameter at breast height between 39 and 72 cm, which were harvested from a mature,
even-aged, heavily storm-damaged forest stand near Zurich, Switzerland. 3 to 5 logs of 5 m length
were cut from each tree, graded and visually inspected for CF after partial debarking. All logs were
sawn into boards of 100 mm or 55 mm thickness, acc. to a systematic sawing pattern adjusted to the
previously marked main wind direction of the storm. Subsequently the sawn boards were cut into full-
size structural timber as well as small clear specimens for the assessment of various mechanical
properties in bending, tension and compression. This paper is focusing only on the bending tests
of structural timber. Results of the tests with small clear specimens from the same sample material
have been published already earlier [15].
The sample material included a wide range of wood quality and growth ring orientations. All sawn
pieces were kiln-dried to a wood moisture content of 15% and planed to their final dimension. Finally,
a sample of 563 squared timber beams (2200⋅110⋅95 mm3) was available for bending tests.
Grading and detection of CF
Each beam was visually graded into 4 strength classes (I, II, III, ungraded) according to the
supplementary specifications of the Swiss standard for the design of timber structures SIA 265/1 [14].
These visual grades are linked to the EN 338 [5] strength classes C24 (I+II) and C20 (III). Due to
practical reasons, visual grade I is currently assigned to C24 instead of the theoretically possible
strength class C27. Deviating from the grading rules, beams with detected CF were not excluded.
The wood density and the axial ultrasonic speed were measured additionally, but these parameters
were not used to derive the strength classes (Table 1).
All beams were inspected for CF and the detected CF were marked. A reliable detection of CF
is difficult and depends on the light conditions, the angle of observation, the surface structure, and the
experience of the observer. All macroscopically visible CF on the longitudinal faces of the beams were
assessed. Because the inspection was done on planed surfaces, even rather fine CF could be detected.
For each beam the 'intensity' of CF was recorded as the total number of identified CF and the 'size'
of the largest CF (Table 1). The 'size' was assessed by a system simplifying the complex geometric
structure of the CF to a one level defect plane, defined by its maximum axial 'width' (CFMAX, Table
1) and the visible length on the circumference of the beam (CFLEN, Table 1). This procedure resulted
in 2 sub-samples of beams without and with CF (CFIND, Table 1).
Table 1. Parameters recorded during grading and testing procedures
Parameter group Code Parameter Unit Details
3 calculated from mass and volume of whole beam at
DENS wood density [kg/m ]
15% MC
USMIN axial ultrasonic speed [m/s] min. value of 2 measurements (device 'Sylvatest')
KMAX diameter of largest knot [mm] between loading heads (inner third of span)
maximum sum of knot diameters (2 or more knots)
KCLU size of largest knot cluster [mm]
within 150 mm length
Wood quality
CW compression wood [%] affected cross-section (0, 5, 10, 20, 30, …, 100%)
discolouring fungal attack in affected cross-section by blue stain or red stripe
FUNG [%]
sapwood (0, 5, 10, 20, 30, …, 100%)
I, II, III, ungraded (i. e. unfit for structural
SCSIA strength class purposes); by visual grading according to Swiss
standard SIA 265/1 [14]
binary indicator variable for
CFIND 0 = without CF, 1 = with CF
presence of CF
CF NCF number of (single) detected CF
CFMAX axial 'width' of largest CF [mm] Classes: 0.1 mm, 0.5 mm, 1.0 mm, > 1 mm
CFLEN circumferent. length of largest CF [mm] visible length on tension edge and side faces
MOR modulus of rupture [N/mm2] calculated for nominal cross section of 110⋅95 mm
Mechanical
MOE modulus of elasticity [N/mm2] (h⋅b)
properties
DMAX total deformation at max. load [mm] measured at neutral axis over 1980 mm span
2
Bending tests
Bending tests followed the procedure given in EN 408 [6] (4-point bending). The average time to
failure was 210 s with 238 beams having times to failure shorter than 180 s due to particular brittle
fracture behaviour with a small deformation at maximum load. Deformation was measured on both
side faces at the neutral axis over the total span as well as within the central gauge length and averaged
over the two faces. The critical section with the expected failure location (e.g. knots, CF) was
positioned between the inner loading points. In order to reproduce the original loading situation in the
tree, the beams from the leeward side of the stem were loaded in the wind-direction and vice-versa.
Thus, present CF were positioned in the majority on the tension edge and thus the beams were loaded
in their most critical orientation. Parameters modulus of rupture (MOR), modulus of elasticity (MOE)
and total deformation at maximum load (DMAX) were recorded (Table 1).
Data analysis
In a first step the grading characteristics of the sample material was compiled. Then analysis of
bending test results was carried out using two different approaches: (1) focusing on the statistical
significance and (2) the practical relevance of the effect of CF on the mechanical properties. The effect
of the 'size' of the CF and the fracture behaviour was regarded as well.
The statistical significance of the effect of CF on the mechanical properties was assessed with a
multiple regression approach. The same main factor model without interaction terms was used for both
dependent variables MOR and MOE as defined in formulas 1 and 2, thus allowing a comparison of the
respective influence of the same factors in both models:
MOR = a0 + a1⋅DENS + a2⋅USMIN + a3⋅KMAX + a4⋅KCLU + a5⋅CW + a6⋅FUNG + a7⋅CFIND + ε1 (1)

MOE = b0 + b1⋅DENS + b2⋅USMIN + b3⋅KMAX + b4⋅KCLU + b5⋅CW + b6⋅FUNG + b7⋅CFIND + ε2 (2)
a0-a7 and b0-b7 are regression coefficients and ε1 and ε2 are error terms in the models 1 and 2
respectively. The abbreviation codes for the regression variables are listed in Table 1. Interaction terms
were not included in the models because they were not significant at the 5% level in most cases.
All 563 tested beams have been included in this analysis.
The practical relevance of CF regarding the mechanical properties was assessed by comparing the
characteristic values of the visually strength graded beams (including both beams without and with
CF) as determined according to prEN 384 [12] with the limits of the given strength classes according
to SIA 265 [13] and EN 338 [5]. 'Ungraded' beams were not included in this analysis.
RESULTS AND DISCUSSION

Properties of sample material
Only very few beams (24) were visually graded as strength class I, 145 and 197 beams were graded as
strength class II and III respectively and a rather high number of beams (197) were graded unfit for
structural timber ('ungraded') (Table 2). Big knots and severe compression wood (CW > 20% in 127 of
the beams) were the most frequent reasons for downgrading. 200 (36%) of the 563 tested beams
contained CF in various 'intensities'. The proportion of beams containing CF increased in the lower
strength classes. More information on sample material properties is given in [1].
Table 2. Number of tested beams grouped by presence of CF (CFIND) and strength class (SCSIA)
Frequency [%] Strength class SCSIA
1) Total
CFIND I (C24) II (C24) III (C20) ungraded
0 23 (4.1%) 114 (20.3%) 120 (21.3%) 106 (18.8%) 363 (64.5%)
1 1 (0.2%) 31 (5.5%) 77 (13.7%) 91 (16.2%) 200 (35.5%)
Total 24 (4.3%) 145 (25.8%) 197 (35.0%) 197 (35.0%) 563 (100%)
1)
Note: According to SIA 265/1 [14] visual grade I is assigned to C24 (see 2.2)
3
For the interpretation of the results it must be noted that the sample material consisted of timber from a
heavily storm-damaged forest stand. Compared to earlier studies with structural timber [7, 8], we have
tested a notably larger number of specimens with CF and with a higher 'intensity' of CF.
Effect of CF on mechanical properties
In Table 3 selected sample statistics for wood density (DENS), modulus of elasticity (MOE), modulus
of rupture (MOR) and total deformation at maximum load (DMAX) are given. MOR and MOE are in
the usual range observed for spruce timber in Switzerland [16], but somewhat higher than in other
recent studies [7, 8]. Variability within the sub-samples without and with CF is very similar. The mean
values of MOR and DMAX are distinctly lower in beams with CF, MOE differs only slightly (Fig. 1).
Table 3. Sample statistics for bending test results grouped by presence of CF (CFIND)
CFIND n Mean Std Min Max
DENS 0 363 472 34 367 570
[kg/m3] 1 200 506 38 424 602
MOR 0 363 53.0 9.9 23.1 76.3
[N/mm2] 1 200 45.8 11.4 12.2 71.6
MOE 0 363 12451 2149 4550 16877
[N/mm2] 1 200 12200 1971 6863 18041
DMAX 0 363 55.7 17.6 19.9 106.4
[mm] 1 200 39.3 14.7 12.4 81.5
Figure 1. Effect of the presence of CF (CFIND) Figure 2. Relationship between predicted and observed
on MOR and MOE. Box plots for sub-samples MOR for the beams with CF. (The thin solid and dotted lines
without and with CF. (Boxes show the median are the linear regression line and the 90% confidence bands for
together with 25th and 75th percentiles, whiskers individual predicted values respectively of the regression model
extend to 5th and 95th percentiles.) built on the sub-sample of the beams without CF. The thick
dotted line is the linear regression line for the beams with CF.)
Both multiple regression models corresponding to formulas 1 and 2 (see above) show a high overall
statistical significance with a coefficient of determination (R2) of 0.45 for MOR and 0.69 for MOE.
Except for compression wood (CW) in the MOR model, all included factors are significant at the 5%
error level. The diameter of the largest knot (KMAX) and the presence of CF (CFIND) are the most
important factors in the MOR model, while wood density (DENS) and axial ultrasonic speed (USMIN)
are particularly dominant in the MOE model. A negative effect of CF on both MOR and MOE could
statistically be proven. For MOR the presence of CF reduces the bending strength 'on average' by 8.1
N/mm2 (95%-confidence limits: 9.7 and 6.5 N/mm2). Related to the mean value of the beams without
CF this corresponds to an average reduction of MOR by 15%. The estimate for the effect on MOE is -
526 N/mm2, which is a 4% reduction related to the mean value of the beams without CF. The
identified more dominant effect of CF on MOR compared to MOE agrees well with earlier studies [8,
15]. As expected, the reduction of the mechanical properties of structural timber is slightly lower than
with small clear specimens [15].
4
In an alternative approach, a multiple regression model (formula 1 without the indicator variable
CFIND), was built on the sub-sample of beams without CF. The resulting regression coefficients were
then used to calculate predicted MOR values for the beams with CF (without explicitly taking the
visually detected CF into account), thus simulating a specifically calibrated machine stress grading
system. An analysis of the deviations of the predicted and the observed MOR values is shown
in Fig. 2. This approach disregarding the visually detected CF clearly predicts too high MOR values
for many beams containing CF. A practical conclusion of this result is, that CF will not reliably be
detected by machine stress grading relying on the dynamic or static assessment of the MOE, as
reported already in earlier studies [3, 15, 17]. Machine stress grading of timber containing CF without
an additional visual inspection may therefore lead to an overestimation of expected MOR and wrong
strength grades.
Characteristic values for MOR, MOE and DENS are listed in Table 4. The sub-samples of beams
without and with CF were combined, but stratified by strength classes. The characteristic values all
exceed the limits in the given strength classes C20 and C24. This means that in spite of the general and
statistically significant reduction of the mechanical properties, visually graded structural timber even
with CF meets the requirements with an adequate safety margin. (During the usual visual grading
process in sawmills, some of the beams with CF would have been detected and probably discarded.)
Table 4. Verification of characteristic values of strength classes according to prEN 384 [12]
Test results 2) Adjustment to reference conditions 3) Characteristic values 4)
Strength kMC kh adjusted ks char.
n Mean Std Min 5%-P limit
class 1) (Factor) (Divisor) values (Factor) value
2
MOR [N/mm ]
I+II / C24 169 56.6 9.0 22.2 39.1 36.8 0.89 32.7 24
− 1.064
III / C20 197 48.7 10.6 12.2 30.1 28.2 0.90 25.5 20
MOE [N/mm2]
I+II / C24 169 13464 1683 8608 10397 14272 14272 11000
1.060 − −
III / C20 197 12432 1839 7362 9279 13178 13178 9500
Density [kg/m3]
I+II / C24 169 475 33 367 422 468 / 33 414 350
0.985 − −
III / C20 197 480 35 395 427 473 / 34 416 330
1)
Strength class based on visual grading (without considering the presence of CF) according to the Swiss standard SIA 265/1 [14]
2)
Abbreviations of sample statistics: n = sample size, Mean = arithmetic mean, Std = standard deviation, Min = minimum value,
5% P = empirical 5th percentile. The input values for the calculation of the characteristic values are printed in bold.
3)
Conversion to reference conditions according to prEN 384: kMC: Correction factor wood moisture content 15 → 12%,
kh: Correction factor timber size / height of beam (150/110)0.2, adjusted values: property values at reference conditions. No
adjustments to a pure bending MOE were made, which leads to a more conservative estimation of the characteristic values of MOE.
Also no adjustments were made for the density as determined from mass and volume of the whole beams.
4)
Verification of characteristic values: kS: correction factor for sample size, char. value: characteristic value calculated from tested
sample, limit: expected characteristic value according to EN 338 [5] for the given strength classes
Effect of the size of CF

Multiple regression analysis failed to show evidence for an improvement of the model fit by the
inclusion of the recorded CF 'intensity' variables NCF, CFMAX and CFLEN instead of the binary
indicator variable for the presence of CF (CFIND). This is interpreted as a general difficulty to
describe the 'damaging' dimensions of CF. The correlation between the 'size' variables of CF and MOE
is even lower. Based on these findings, the use of allowable 'size' limits for CF in grading procedures
seems neither safe nor practical. This result is in contrast to some earlier reports, where such limits
have been proposed [7, 8].
Fracture behaviour
If CF were involved in the mode of failure, frequently abnormally brittle and short-fibred fractures
have been observed. Fracture occurred often suddenly and without any prior indications. In some cases
the beams were broken completely over the whole cross-section. The low-strain failure mode is also
apparent in the total deflection at maximum load. Average total deflection at maximum load was only
39 mm in beams with CF compared to 56 mm in beams without CF, which corresponds to a reduction
5
of 30% (Table 3). The beams with the lowest MOR all contained CF and exhibited very low
deformations. This brittle fracture behaviour has been observed already in earlier studies [8, 9, 20].
Moreover, CF have been reported to be particularly sensitive to impact loads (e.g. impact bending
tests), which showed a high strength reduction by CF [9, 15].
CONCLUSIONS
The following conclusions regarding the influence of wind-induced CF on the mechanical properties
of spruce structural timber can be drawn from the results of the study:
• Bending MOR and MOE of squared timber beams containing CF are statistically significantly
lower compared to beams without visible CF. (MOR more pronounced than MOE)
• Despite the general reduction of strength and elasticity, the limits for the characteristic values
of the strength classes C20 and C24 of visually graded structural timber (acc. to the Swiss standard
SIA 265/1 [14]) are still exceeded. Considering the decreasing influence of other structural defects,
this may however not be the case in the higher strength grades.
• Because the MOE is only slightly affected by the presence of CF, machine stress grading methods
relying on the dynamic or (low stress level) static assessment of the MOE are not able to reliably
detect CF. Machine stress grading of timber containing CF without an additional visual inspection
may therefore lead to an overestimation of expected MOR and wrong strength grades.
• The macroscopically visible appearance ('size') of the CF is only a weak indicator for the potential
reduction of MOR and MOE. It is therefore impossible to distinguish between 'benign' and
'malignant' CF and to define allowable 'size' limits for CF for visual grading procedures. Only a
strict exclusion of CF seems practical.
• Timber containing CF frequently fails abnormally brittle with low-strain, short-fibred fractures.
• Because of the potential safety risk and the difficult prediction of their strength reduction, detected
CF should be excluded from load bearing structural elements stressed in tension or bending and
explicitly addressed in the relevant grading standards. Timber containing CF should only be used
in compression loaded or not load-critical applications.
ACKNOLEDGMENTS
The presented work was funded partly by the Swiss Forest Agency.
REFERENCES
[1] Arnold M., Steiger R. 2007. The influence of wind-induced compression failures on the mechanical properties of spruce structural timber. Materials
and Structures 40:57-68.
[2] Arnold M. 2003. Compression Failures in wind-damaged Spruce Trees. In: Proceedings International Conference 'Wind Effects on Trees',
16. – 18. September 2003, University of Karlsruhe, Germany, (Ed. B. Ruck et al., Karlsruhe), 253-260.
[3] Bernasconi A., Boström L., Schacht B. 1999. Detection of severe timber defects by machine grading. Proceedings CIB-W18, 32. meeting, Graz,
Austria, Paper 32-5-2.
[4] Delorme A. 1974. Über das Auftreten von Faserstauchungen in Fichtensturmholz. Forstarchiv 45(7):121-128.
[5] EN 338:2003. Structural timber – Strength classes.
[6] EN 408:1995. Timber structures – Structural timber and glued laminated timber – Determination of some physical and mechanical properties.
[7] Glos P., Denzler J.K. 2004. Einfluss von Faserstauchungen auf die Festigkeit von Fichtenbauholz. Schweiz. Z. Forstwes. 155(12):528-532.
[8] Hoffmeyer P. 2003. Mechanical properties of timber from wind damaged Norway spruce. In: Proceedings of the 'Second International Conference
of the European Society for Wood Mechanics', Stockholm, Sweden, 25.-28. May 2003 (ESWM, Ed. L. Salmén, 2003).
[9] Koch G. 1999. Sekundäre Veränderungen im Holz dynamisch beanspruchter Fichten (Picea abies [L.] Karst.) aus immissionsbelasteten
und windexponierten Hochlagenbeständen. Mitteilungen der Bundesforschungsanstalt für Forst- und Holzwirtschaft, Hamburg, Nr. 192.
[10] Kucera L.J., Bariska M. 1982. On the fracture morphology in wood. Part 1: A SEM-study of deformations in wood of spruce and aspen
upon ultimate axial compression load. Wood Sci. Technol. 16:241-259.
[11] Mergen F., Winer H.I. 1952. Compression failures in the boles of living conifers. Journal of Forestry 50:677-679.
[12] prEN 384:2000. Structural timber – Determination of characteristic values of mechanical properties.
[13] SIA 265:2003. Holzbau. Schweizerischer Ingenieur- und Architektenverein, Zürich.
[14] SIA 265/1:2003. Holzbau – Ergänzende Festlegungen. Schweizerischer Ingenieur- und Architektenverein, Zürich.
[15] Sonderegger W., Niemz P. 2004. The influence of compression failure on the bending, impact bending and tensile strength of spruce wood
and the evaluation of non-destructive methods for early detection. Holz als Roh- und Werkstoff 62(5):335-342.
[16] Steiger R. 1995. Biege-, Zug- und Druckversuche an Schweizer Fichtenholz. Forschungsbericht ETH/IBK Nr. 207, Birkhäuser, Basel.
[17] Steiger R. 1996. Mechanische Eigenschaften von Schweizer Fichten-Bauholz bei Biege-, Zug-, Druck- und kombinierter M/N-Beanspruchung;
Sortierung von Rund- und Schnittholz mittels Ultraschall. Forschungsbericht ETH/IBK Nr. 221, Birkhäuser, Basel.
[18] Terziev N., Geoffrey D., Marklund A. 2005. Dislocations in Norway spruce fibres and their effect on properties of pulp and paper. Holzforschung
59:163-169.
[19] Timell T.E. 1986. Compression Wood in Gymnosperms, Volume 3, Chapter 15.2.1.5, Springer, Berlin.
[20] Trendelenburg R. 1940. Über Faserstauchungen in Holz und ihre Überwallung durch den Baum. Holz als Roh- und Werkstoff 3(7/8):209-221.
[21] Wilkins A.P. 1986. The nomenclature of cell wall deformations. Wood Sci. Technol. 20:97-109.
6
15th – 17th October 2007, Warsaw, Poland
DEVELOPMENT OF QUALITY MANAGEMENT IN POLISH WOOD

INDUSTRY
Jan Bagiński1, Łukasz Babuśka2, Marcin Klimczewski3
1
Warsaw University of Technology, 02-524 Warsaw, Narbutta 85, j.baginski@wip.pw.edu.pl
1
Warsaw University of Technology, 02-524 Warsaw, Narbutta 85, lukasz.babuska@wesola.3.pl
2
Warsaw University of Life Sciences, 02-776 Warsaw, Nowoursynowska 159, marcin_klimaczewski@sggw.pl
Keywords: ISO 9001:2000, Quality Management Systems, Wood Industry
ABSTRACT
Recently, quality management systems concept is very well-liked however this approach, within polish as well as
international wood industry, is applied very rarely. On the other hand, lately, development of polish wood industry
has taken a place and it allowed for increasing of quality management systems certification number. The core ideas
of quality management system implementation are: management which it is base on facts and data, process approach,
people development what in consequence provides to achieve benefits easier.
DEVELOPMENT OF QUALITY MANAGEMENT SYSTEMS AROUND THE WORLD

The ISO 9001:2000 and ISO 14001 standards (1996 and 2004 versions) are implemented
by some 887 770 organizations in 161 countries [www.iso.org].
The ISO 9001 family is among ISO's best known standards ever. The ISO 9000 family addresses
"quality management". This means what the organization does to fulfill:
− the customer's quality requirements, and
− applicable regulatory requirements, while aiming to
− enhance customer satisfaction, and
− achieve continual improvement of its performance in pursuit of these objectives” [www.iso.org].
700000
600000
500000
Numer of certificates
ISO 9001/2/3:1994
400000
300000
200000
100000
0
2000 2001 2002 2003 2004
Year
Figure 1. Number of certificates around the world according to the ISO 9001:2000 standard
(WIŚNIEWSKA S., 2007)
7
The rapid growth of quantity of certificates is caused by customer requirements and not only final
customer but as well and maybe the most important an internal customer among corporation or even
among plant or department.
DEVELOPMENT OF QUALITY MANAGEMENT SYSTEMS IN POLAND
Permanent changes at the polish market causes increasing of the organization number with certified
Quality Management System. Figure 2 shows how many organizations in Poland have implemented
certificate of Quality Management System according to ISO 9001:2000.
7000 6618
Numer of certified organizations
6000 5482
5070
5000
4086
4000
3091
3000
2157
2000 1515
997
1000 600
109 301
0
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Year
Figure 2. Growth of certification of Quality Management Systems according to ISO 9001:2000

(WIŚNIEWSKA S., 2007)
In the last decade number of organization with certificate of Quality Management System has
increased sixty times and number of organization in general has increased only three times.
Organizations implement Quality System due to different reasons. The most popular reasons are
as follow:
- marketing tool,
- cost saving,
- customer requirement,
- corporation requirement,
- market requirement (for instance automotive industry or aircraft industry).
In Poland, till now, the most important was to utilize Quality Management certificate as an
advertisement tool. However nowadays companies changing their point of view on the market as well
as they are changing it into customer approach. Organizations try to meet customer requirements
and Quality Management System is one from the best, helpful instrument for this.
It is good to underline that Figure 2 concerning only ISO 9001:2000 so the most popular of Quality
Management System and id does not concern high-specialized, industry standards and regulations,
which presently are popular as well. Different Quality standards are base on process and system
approach therefore it is easy to make a mix or make a compilation of each other.
Quality System base on processes it has not to be certified, but this fact makes harder an identification
of company which use this quality system.
8
POLISH WOOD INDUSTRY

Central Statistical Office in Poland – GUS, classifies wood industry according to the EAC in two
major sectors: 6 – manufacturing of wood and wood, straw, and wicker products (except furniture)
and 23 – manufacturing of furniture and manufacturing n.e.c. The numbers of officially registered
enterprises in these sectors are shown in table 1.
Table 1. Industrial enterprises by number of employees in I quarter 2007 (Central Statistical Office - GUS data).
SECTOR total 0-9 10-49 50-249 250-999 1000 and more

Manufacture of wood and wood,
40454 37501 2476 417 53 7
straw and wicker products
Manufacture of furniture;
42550 39617 2265 556 98 14
manufacturing n.e.c.
The first sector includes:

▬ sawmills,
▬ veneer production,
▬ timber frame constructions,
▬ windows and doors production,
▬ flooring material,
▬ wood based panel production (plywood, chipboard, MDF, HDF and fiberboard
form wet method production).
The second sector mainly covers furniture production.
Table 2. Sold production, employment and average monthly gross wages and salaries in industry by sections and
divisions in I quarter 2007 (Central Statistical Office GUS - data).
Average paid Average monthly gross

Sold production
employment wages and salaries
SECTIONS AND DIVISIONS
. I quarter I quarter
I quarter in zl
in mln zl in thous. 2006=100 2006=100
2006=100
TOTAL 198902,2 113,0 2542,1 103,2 2777,98 108,3
MANUFACTURING 168174,6 116,2 2164,1 104,0 2541,98 108,7
Manufacture of wood and wood, straw and 5510,7 119,1 98,3 101,2 1932,35 112,8
wicker products
Manufacture of furniture; manufacturing n.e.c. 7607,9 118,9 163,3 104,0 1964,35 109,5
A current condition of wood industry in Poland in 2007 was presented in table 2.

Since 1989 the polish wood industry is rapidly growing. The growth rate is much faster then average
in national economy. The dynamics of that process is strongly dependent on the type of industry
sector. In a case of wood based panel industry it was aligned with privatization of state owned
companies and huge investments in new production lines. In furniture production sector intensive
growth was a continuous trend for many years. During last 15 years Poland becomes one of the biggest
furniture producers in the EU (6thin 2006) and worldwide (10th in 2006) (Adamowicz, Wiktorski
2006). Considering the period between 1989 and 2005, the furniture export increased over 30 times
reaching the value of $5030 millions (fig.3).
9
Figure 3. TOP 20 world-exporting countries in 2005 (export value in € millions) (Adamowicz, Wiktorski 2006)
The furniture production becomes very important sector of national economy. As shown on table 3
the value of furniture production as a share of GDP is more then two times higher in Poland
than EU average.
Table 3. Importance of furniture industry for polish economy (Adamowicz, Wiktorski 2006)
Description UE – 25 Poland*
Gross Domestic Product (GDP) 10 527,0 bln € 268,7 bln €
Value of the furniture sold production ** 93,8 bln € 5,6 bln €
Furniture production share in GDP 0,9% 2,1 %
Notes:
*1 € = 3,8951 PLN – average annual exchange rate in 2006 r. based on NBP
** Estimate values for companies employing more than 9 people. Based on: UEA, GUS, OIGPM
Dynamics of sold production in 2006 and 1 quarter of 2007 for two main wood industry sectors
is presented at fig. 4. The base for indexes calculation was the monthly average value of sold
production in year 2000.
10
MANUFACTURING TOTAL
Manufacture of wood and wood, straw and wicker products
Manufacture of furniture; manufacturing n.e.c.
250
200
150
100
monthly average of Y 2000 = 100
50
0
I II III IV V VI VII VIII IX X XI XII I II III
Fig.4 Volume index of sold production of wood industry in 2006 and 1 quarter 2007
(Central Statistical Office –GUS data)
DEVELOPMENT OF QUALITY MANAGEMENT SYSTEMS

IN POLISH WOOD INDUSTRY
With the rapid grow and relatively shallow internal market polish wood industry is based on export.
To sustain strong competition on western, highly demanding markets, producers must focus on quality
of their products and best in class efficiency of production process. New trends, as outsourcing,
enables development of strong production competences and brings into consideration the need for B2B
communications. This situation results with a change in approach to the ISO 9001 standard. For many
years main reason for implementation of QMS standard was based on companies marketing strategy.
Nowadays it is seen more likely as a tool for establishing efficient management system within
the company and supply chain.
In wood based panels industry since many years QMS was a market standard requirement. Currently
all enterprises in that sector are ISO 9001:2000 certified.
The opposite situation is within sawmilling sector. Almost none of such companies are certified
according to the requirements of ISO 9001. This reflects global trends in that sector where the main
focus is on environmental issues and FSC certification.
Among doors and windows production sector the ISO 9001 certificates are popular, mainly because
major part of the production is made for export. The other reason is that statutory requirements
for those products are forcing companies to establishing formal and documented quality control
systems. This situation promotes ISO 9001 standard as one of the best practices in that subject.
Situation within furniture production sector according to OIGPM research was presented at fig. 5.
There is strong correlation between size of the company and its interest in ISO 9001 certificates.
One of the reasons that ISO certificates are not very popular within smaller companies is the cost
of implementation, certification and maintenance of such QMS system. Apart form the costs connected
to certification process, the need for formal management standards and best practices is much stronger
among bigger companies.
11
100%
90%
80%
70%
60%
50% OTHER
40%
ISO 9001
30% Certified
20%
10%
0%
>249 100-249 50-99 25-49 10-24 <10 TOTAL
Fig.5 ISO 9001certified industrial enterprises in furniture production sector by number of employees in 2004
(OIGPM)
Significant impact on QMS implementation in Poland has activity of world biggest trade companies
in furniture market – IKEA. IKEA’s buying policy is established on the basis of the Supplier Quality
Assurance Program - the QMS that covers industry specific requirements. All IKEA suppliers are
forced to implement it. The standard has four levels of requirements: entry level – QMUST, basic level
– QWAY and advanced level − 4SIP, the most advanced − 4th level requires fulfillment of 4SIP
requirements and 3rd party ISO 9001 certificate. Because IKEA (and SWEDWOOD which is owned
by IKEA) are cooperating with significant numbers of suppliers within polish wood industry,
this standard becomes the most popular.
CONCLUSIONS
With its intensive growth, polish wood industry is searching for good management standards.
Implementation of QMS based on ISO 9001:2000 standard becomes “best practice” among
the enterprises in this sector. The process approach brings in to company’s management system
a valuable tool for continuous improvement and high efficiency of production process. The other
important advantage is impact on marketing strategies where ISO 9001 certificates are recognized
as an entry requirement for international markets.
REFERENCES
1. ADAMOWICZ M., WIKTORSKI T., 2006. Condition and Development Prospects of the
Polish Furniture Industry. Annals of Warsaw Agricultural University. Warsaw 2006
2. Biuletyn informacyjny OIGPM., Czerwiec 2006: http://www.oigpm.org.pl/biuletyn.php
3. FACTS Report., 2005. UEA. Brussels 2005.
4. International Standards Organisation, The ISO Survey of Certifications 2004, Geneva 2005.
5. WIŚNIEWSKA S., 2007 Sposoby zapewnienia jakości w
polskich przedsiębiorstwach, Przegląd Techniczy nr 1/2007,
6. Reports of Central Statistical Office (GUS) www.stat.gov.pl
7. www.iso.org
12
Bark Measurements with X-Ray Technology

Rafael Baumgartner1, Franka Brüchert1, Jörg Staudenmaier1, Udo Hans Sauter1
1
FVA, Wonnhaldestr. 4, 79100 Freiburg, Germany
Rafael.Baumgartner@forst.bwl.de, Franka.Bruechert@forst.bwl.de, Joerg.Staudenmaier@forst.bwl.de,
Udo.Sauter@forst.bwl.de
Keywords: bark, spruce, x-ray
ABSTRACT
Automatic measurement of log dimensions at the gate of medium and large scale sawmills in Germany and other countries
in Central Europe is successfully established for many years. The scanner technology measures the dimensions of the logs
at the debarked surface precisely. On this basis the wood volume is calculated without bark as a reference for the payment
of the resource delivered by the forest owners. The use of this method in sawmills requires to debark the logs directly after
the logs arrive at the wood yard of the mill due to the requested payment after short time. For some wood species and
during the summer months it can be critical to store debarked logs because of fast degradation caused by discolorations
due to fungi, cracks and insect damages. One possibility to overcome these problems of early debarking to date is to
measure the logs over bark and subtract a bark estimate derived from a fixed table afterwards. This does not require
debarking but on the other hand leads to a loss of precision with respect to the variation of bark thickness.
Application of automatic detection and measuring of the bark using x-ray technology allows to control and update
the discounts in the bark tables, but more importantly to measure exactly the real volume of the wood at the mill site
without prior debarking.
In this investigation, 96 logs of Norway spruce were tested to measure wood volume using x-ray technology. The logs were
fully barked except for defined, exactly measured, rectangular shaped, debarked stem sections to test the system
for accuracy of bark detection. Data were gathered by an industrial scanner with two x-ray sources installed in a sawmill
in Austria. Algorithms extracting the bark thickness based on the form of the signal change at the edges of the logs were
developed and tested. The manual reference values were obtained by measurements at four positions around the log each
at the butt end, half length, and the top end of the logs. High correlations between the values computed by the algorithms
and the manually measured values were found.
INTRODUCTION
Automatic measurement of log dimensions at the gate of medium and large scale sawmills in Germany
and other countries in Central Europe is successfully established for many years. The scanner
technology measures the dimensions of the logs at the debarked surface precisely. On this basis
the wood volume is calculated without bark as a reference for the payment of the resource delivered
by the forest owners. The use of this method in sawmills requires to debark the logs directly after the
logs arrive at the wood yard of the mill due to the requested payment after short time. For some wood
species and during the summer months it can be critical to store debarked logs because of quick
degradation caused by discolorations by fungi, by cracks, and insect damages. One possibility
to overcome these problems of early debarking to date is to measure the logs over bark and subtract
a bark estimate derived from a fixed table afterwards. This does not require debarking but on the other
hand leads to a loss of precision with respect to the variation of bark thickness.
Application of automatic detection and measuring of the bark using x-ray technology allows to control
and update the discounts in the bark tables, but more importantly to measure exactly the real volume
of the wood at the mill site without prior debarking.
MATERIALS AND METHODS

Wood material
This investigation used 96 Norway spruce logs cut from 27 trees taken from two stands in the Swabian
Alb in southern Germany. The logs were handled with such care to keep them in bark as much as
possible during transport. After arrival at the mill each log was debarked at defined positions (small,
rectangular areas) which were exactly measured in position, size, and shape. These areas were used to
13
test the system for accuracy of bark detection. Between felling and all measurements there were
at most five days in which the logs were stored to prevent them from drying, keeping moisture loss
to a minimum.
Manual measurements
From each log circular shaped bark samples (Figure 1) with a diameter of 3 cm were taken at the butt
end, half length, and the top end of the log. At each of these positions four samples were drawn evenly
distributed around the log. The thickness of these samples was measured with a calliper immediately
after extraction. From the twelve thickness values an average was calculated and used as reference
value for the automatic procedures.
Figure 1. Circular shaped bark sample next to the position it was drawn
Automatic measurements
Data for the automatic procedure were gathered by a Microtec Tomolog®, an industrial scanner with
two x-ray sources and corresponding detectors, installed in a sawmill in Austria.
From the raw data an attenuation image was computed and smoothed in longitudinal direction of the
log. This resulted in an image where each row corresponds to the attenuation profile of the log
at a certain position in longitudinal direction. The background was set to zero. A line with constant
slope was fitted to the signal ascent on one side of the log and a second line to the descent on the
opposite side of the log. From the length of the ascent/descent of the fitted line a first value for the
bark thickness was derived. A constant offset was found during validation of the values for debarked
areas. This offset was subtracted from every bark thickness value. Figure 2 shows the shape of a log
with bark (continuous line) and the shape without bark (dashed line) which was calculated from the
bark thickness derived by the method described.
Figure 2. Shape of the log (continuous line) and of the recognized solid wood (dashed line); missing parts in the
downer lines are caused by the conveyor where neither the outer shape nor the shape of solid wood can be detected
14
Due to knots or other defects outliers were produced in the calculation. To eliminate these outliers two
methods were applied: bark thickness values over a certain threshold were eliminated, followed by
the application of a median filter.

With the methods described in the last chapter, the data gathered by the x-ray scanner were analysed
and the results were compared to the manually derived data.
Statistical distribution of single bark measurements

The distribution of the automatically derived bark thickness for every position in every log was
compared to the distribution of the 1152 (96 times twelve) manually gathered values. The main
parameters are shown in table 1.
Table 1. Main parameters of the statistical distribution of the single bark measurements
mean [mm] std. dev. minimum 1st quartile median 3rd quartile maximum
[mm] [mm] [mm] [mm] [mm] [mm]
manual 6.0 1.5 3.2 4.9 5.7 6.9 14.1
automated 5.8 2.3 0.0 4.0 5.5 7.5 28.0
The mean value of all results obtained by the automatic procedure is similar to the manual one(5.8 mm
compared to 6.0 mm), but the results have a higher standard deviation (2.3 mm in comparison
to1.5 mm). One reason for that is that the range of the values in the automated measurements
is broader which can be seen from the lower minimum and the higher maximum (0.0 mm and 28.0 mm
for automated, 3.2 mm and 14.1 mm for manual measuring). But also for the central part of both
distributions differences can be found. The range between first and third quartile is wider for
automated measurements than for the manual measurements (between 4.0 mm and 7.5 mm compared
to 4.9 mm and 6.9 mm). For the median likewise the mean a similar value for automatic and manual
results could be achieved (5.5 mm / 5.7 mm).
One reason for the wider spreading of bark thickness values is that x-ray data includes errors especially
in the rough conditions in a sawmill. Another reason is that in the automatic procedure every position
is measured except for the conveyor positions and some eliminated values because of the threshold.
This sums up to about 540’000 positions in contrast to 1152 positions of manual measurements.
An increase of robustness against some wrong measurements is the consequence, but also a larger
variety of values, because of the inhomogeneity of bark.
Measurements for whole logs

By averaging all values for one log, a single value (“automatic bark thickness”) was obtained
for the automatic procedure. From the twelve manually gathered values per log the average was
computed, too (“manual bark thickness”). Figure 3 shows the plot of manual versus automatic bark
thickness.
15
Figure 3. For each log manual versus automatic bark thickness is plotted
Each point corresponds to the average measurement for one log. The dashed black line is the optimum
where manual and automatic values are equal, and the continuous grey line is the best linear fit.
The quadratic mathematical correlation (R²) between the manually and the automatically derived
results is 0.78 (where total correlation leads to a value of one; no correlation to zero). The average
absolute difference between manual and automatic bark thickness is 0.49 mm.
CONCLUSIONS
This investigation shows that for the observed logs with x-ray technology a bark measurement can be
carried out which is on average less than half a millimetre different from the manual reference value.
Up-to-now the procedure was only applied to data recorded on one day and logs from two stands.
A verification with an independent test is in progress.
REFERENCES
1. Leenen M. (2005): Vermessung von Rundholz in Rinde mit Röntgentechnologie. Final report
of the identically named project for the Holzabsatzfonds, Bonn
2. Marshall H. D., Murphy G. E., Lachenbruch B. (2006): Effect of bark thickness estimates
on optimal log merchandising. Forest Products Journal Vol. 56, No. 11/12, pages 87-92
16
Dimensional and Shape Quality Control. Modern Equipment

Camelia Boieriu1, LuminiŃa Brenci2
1
Transilvania University of Braşov, B-dul Eroilor 29, 500036 Braşov: cboieriu@unitbv.ro
2
Transilvania University of Braşov, B-dul Eroilor 29, 500036 Braşov: brenlu@unitbv.ro
K ey w o rd s : processing accuracy, measurement methods, modern equipment.

ABSTRACT
Nowadays, the product quality in furniture industry is more and more required for the competitiveness of the products on
the market. The quality is defined through a high processing accuracy, high quality of the raw materials, good design and
high quality of the finished surfaces and of the mounted product. That means a controlled production process, a controlled
supply chain, good equipment and high quality tools. The accuracy of the finished product is the result of the accuracy
of the technological process and it is defined by the machine parameters used for the process operation and also by the
tool type and quality. In the present paper, the authors present new methods and modern equipment used to measure
the processing accuracy.
INTRODUCTION
The European Union common regulations have as a result the introduction of a DISCIPLINE IN THE
MANUFACTURING PROCESS of the producers, in order for them to guarantee the quality of the
products sold in the European Union. That means, among the other things, to assure the accuracy level
of their production. This guarantee can not be granted unless periodical analyses of the “production
activity results”, one of the analyzed parameters being of course, the quality of the product.
To define if the company has good activity results it is needed, for example, to test the quality of the
product, which is also the result of its dimensional and shape precision. The dimensional and shape
precision are also the result of the fabrication process, due to the machines and equipments used for the
process. The dimensions measurement gives indication about the processing accuracy of the machines
used into the process. The measurements can be done with classical instruments with the condition of
being very accurate ones (gauged and certified by specialized bodies).
In wood industry field, the dimensions can vary from 2 mm to 2 200 mm. In this case it is difficult to
measure with the same instrument the small and the big dimension, that instrument being also very
accurate one. So, in the last years, when the knock – down furniture became more and more requested
by the market, the dimensional accuracy of the parts was compulsory for the interchangeability
purpose. So, the needed for the equipment able to measure small and high dimensions in an easier way
went to the concept of measuring equipment for wood industry purpose.
Thus, in the frame of the Wood Industry Faculty from Braşov, has been set up on the 23-rd of April
2004 the most modern laboratory of testing the processing accuracy of the wooden and wooden based
components resulted from the furniture and finished wooden products sector, unique in Romania. The
laboratory was set up with the help of the German in cooperation with SAN Software Company from
Romania and the German Company HECHT ELECTRONIC AG.
SAN Software has offered the support necessary for the endowment of the laboratory with a modern
computer network which transfers the drawings obtained with CAD program to the measuring
equipment. The German Company HECHT ELECTRONIC AG is the producer of the measurement
equipment and the investment has been done through a project initiated by SEQUA Foundation from
Germany.
OBJECTIVES OF USING MODERN MEASURING EQUIPMENT
Fabrication of the wooden products as engineering processes means the presence and uses of a
system between machines, tools, devices, installations, etc. In an industrial process, the fabrication
process introduces different concepts than in prototype or craft industry and they are:
batch processing, the dimension of the batch being established as function of the production
capacity of the plant, the order size and delivery schedule, aiming to optimize their size in order to
use the rational engineering site and inventory space, the optimum changeover and processing
times, etc;
17
to assure the total interchangeability of the parts in a batch, condition achieved only through
assuring the same processing quality for all the batch components (dimensional processing
accuracy, shape and position accuracy) and through assuring a permanent processing control (using
dimensional, shape and position checking devices – for all important operations or between groups
of operations – for parts, complex parts and sub-assemblies);
to assure a permanent maintenance activity to keep the work parameters of the system of machines,
tools, devices, checking devices, etc.
Performances of an engineering process impose in fact a high level of processing accuracy at
each workstation, including also a high level of processing accuracy of the production sector or of the
plant.
Because of the topics mentioned before, the paper intends to present new equipments of
measuring the processing accuracy and the specific measuring methods.
EQUIPMENT AND MEASURING METHOD

CADesQ Equipment
CADesQ equipment, presented in figure 1 is composed of a measuring table of 2500 x 1250, mm, a PC
Touch-screen and a measuring device placed on a moveable bridge, which has a translation movement
on X axis. The measuring device is composed of six measuring heads, each one being positioned for a
separate measurement type (length or width, depth, hole diameter, hole depth, etc.). The equipment has
a measurement accuracy of 0.01 mm.
Figure 1. CADesQ measuring equipment
The drawing of the part which is intended to be measured is a .dxf files (developed by AutoCAD).
Based on the software developed by the manufacturer, all the measurements to be done are introduced
in the measuring program. They may be as follows: linear sizes (on the three coordinate axis X, Y and
Z), holes diameters and distances between them, hole depth, together with the accepted tolerances
according to the technical specifications or to the customer requirements. The interface of the program
with the operator is shown in figure 2. Through the command: ”New meas. point”, the type of
measurement (distance on X or Y axis, depth on Z axis, diameter of the hole, depth of the hole, etc)
and the accepted tolerances for those dimensions are introduced into the measuring program. “Read
measure from .dxf” will be selected than and the cursor will choose from the drawing imported as .dxf
file the start point and the end point that define the line or the point that define the hole position.
After defining the measurement points and the imposed tolerances, the file will be transferred through
network on the measuring table PC, where the operator will measure the dimensions according to the
program.
The measured values are transferred to a .txt file, in which it is specified if the measured dimension is
included in the tolerances field or not. Than, statistical parameters and measurement uncertainty can be
calculated in order to describe the process.
18
Figure 2. Interface of the measuring program on CADesQ equipment

OptoDesQ Equipment
OptoDesQ equipment presented in figure 3 is composed of a granite table, a video CCD camera fixed
on a mobile bridge and a computer. The measurements are contact less performed, with a precision of
0.01 mm, using an interface program (OpoDesq Interface) shown in figure 4, defined by the operator,
based on the drawing or client requirements. First the overall dimensions (length, width and thickness)
are introduced in the program.
Figure 3. OptoDesQ equipment

For example, when testing the panel flatness, „Vertical hole” is introduced as an option and the
coordinates of the points in which we measure the dimension on “z” axis together with their tolerances
are introduced.
Figure 4. Interface on OptoDesQ program

The data will be automatically processed by the software and than generated by it in an interface as
shown in figure 5 and in the drawing of the measured part when critical points appear in red color.
After the measuring action is completed, the data will be transferred in a .txt file. This will allow
statistics of some parameters and also the measurement uncertainty. OptoDesQ allows to measure the
19
following parameters: straightness, linear measurements (length, width, thickness), the angular
deviations of the edges, flatness, the hole centre position and the hole diameter, distances between
holes and the edges linear deviations.
Figure 5. Results generated by OptoDesQ program

LaminatDesQ Equipment
LaminatDesQ equipment is characterized through high precision measurements and it is recommended
to be used for measuring the laminated products: parquet and floor panels, veneered or laminated
panels having a thickness of maximum 18 mm. LaminatDesQ allows the measurement with a precision
of 0.02 mm of the following characteristics: linear sizes (length, width, thickness), parallelism,
flatness, rectangularity, tongue and mortise sizes according to DIN EN 13329. The obtained results can
be seen on the display. The results can be printed.
Figure 6. LaminatDesQ equipment

CONCLUSIONS
In order to allow the transfer of the wooden and wooden based products on the European market, as
mentioned before, it is necessary for the manufacturer to get certificates of conformity for the products
they are producing. According to standards 17050-1:2005 and 17050-2:2005, the declaration of
conformity issued by a certified body/laboratory (according to the requirements of SR EN ISO/CEI
17025:2005), proofs the conformity of an object with the specified requirements, no matter of the
sector of production. Thus the object of any declaration of conformity can be a product, a process, a
management system, a person or a specific body. The equipment we have presented offers the
advantage of a high measurement precision and modern measurement methods, not standard yet, but
which can be validated, in order to assure the producers about the accuracy of the results, certifying
them the precision of their manufacturing process.
REFERENCES
1. *** (2003). Manual of instructions for CaDesQ equipment. Firma Hecht Electronic A.G.
2. *** (2003). Manual of instructions for Laminat DesQ equipment . Firma Hecht Electronic A.G.
3. *** (2003). Manual of instructions for OptoDesQ equipment. Firma Hecht Electronic A.G.
4. *** SR EN ISO/CEI 17051-1:2005 –Conformity assessment. Declaration of conformity issued by
the supplier. Part 1: General requirements.
20
Ultrasonic evaluation of mechanical properties of wood in standing trees

Witold Dzbeński, Tomasz Wiktorski
Warsaw University of Life Science, Faculty of Wood Technology, ul. Nowoursynowska 159, PL 02-776 Warsaw
witold_dzbenski@sggw.pl tomasz_wiktorski@sggw.pl
Keywords: mechanical properties, non-destructive testing, standing trees, timber, ultrasound
ABSTRACT
Methods based on propagation of stress waves phenomena indicate particular usefulness in diagnosis and assortmentation
of raw material. It has been demonstrated that ultrasonic measurements made both on sawn timber and on standing trees
produce satisfactory results in predicting certain mechanical properties of sawn timber obtained from the analysed raw
material. The highest correlation has been observed between the velocity of ultrasounds along trunks and the modulus
of elasticity of sawn timber obtained from them. A similar correlation with the bending strength has also been satisfactory.
To ensure the correct interpretation of the results, it is recommended that studies should be conducted to detect and locate
inner rot in standing tree trunks. The studies can be performed by taking measurements of ultrasonic wave propagation
transversely in trunks in a potentially defected area and an area free from defects.
INTRODUCTION
Development of non-destructive testing made it possible to use the technique for the purposes
of diagnosis and assortmentation of raw material and in a forest taxation work. Acoustic methods
based on the phenomena accompanying propagation of longitudinal stress waves and transverse
vibration demonstrate their particular usefulness for such testing. Based on measurements of round
sawn timber [Dzbeński 1981], it was found that the velocity of transverse ultrasonic waves
propagation in pine trunks could be a good indicator of mechanical properties of sawn timber obtained
from them, e.g. aircraft timber. Similar results based on acoustic methods were achieved among others
by Jayne /1959/, Marra, Pellerin and Galligan /1966/, Kaiserlik and Pellerin /1977/, Aratake and Arima
/1994/, Sandoz /1996/, Ross, McDonald, Green and Schad /1997/, Wang and others /2000/.
The results of the above-mentioned tests suggested a possibility of detecting hidden defects in round
wood, e.g. inner rot in telecommunications poles and masts [Breeze, Nilberg 1971], and in standing
tree trunks [Sandoz, Lorin 1994, Sandoz 1999]. Using acoustic methods, Ross, De Grott, Nelson
/1994/ were detecting biological degradation of wood in tree trunks, while Dzbeński and Wiktorski
/2004, 2007/ - in pine and spruce trees in Polish forests.

Mechanical properties of sawn timber in the light of defectoscopic testing of sawn timber raw
material.
Ultrasonic methods, used already thirty years ago e.g. at the Faculty of Wood Technology, Warsaw
University of Life Science demonstrate particular usefulness for assortmentation of saw logs.
They successfully replace the testing conducted up to now with the use of traditional (destructive)
methods on trees specimens. Traditional methods required a long time to perform tests and provide
only an approximate evaluation of a large batch of material on the basis of testing of a small sample
population. 100% measurements were conducted on pine logs using ultrasonic waves propagated
transversely in the trunk (by equipment of 541 type produced by UNIPAN, Warsaw). The bark was cut
through with conic or needle concentrators extending gauge probes (heads).
It was found that the velocity of ultrasound propagation can be an important (r=0.588 … 0.647)
indicator of mechanical properties determined longitudinally in sawn timber obtained from the round
raw material where waves with the velocity of c⊥ travel in sapwood with a high moisture content over
a heartwood area with knots (or inner rot) grouped in a core wastewood zone. The testing results
are presented in table 1.
21
Table 1. Relationship between velocity of ultrasounds conducted transversely in trunks and selected mechanical
properties of (air dried) pine sawn timber obtained from them.
Type of sawn timber properties Regression equation f(c⊥) Correlation coefficient (r)
Bending modulus of elasticity
Eg = 0.039 c⊥ + 44.3 0.647
Eg [MPa*102]
Static bending strength
Rg= 0.230 c⊥ + 551.1 0.599
Rg [MPa*10-1]
Longitudinal compressive strength
Rc = 0.113 c⊥ + 290.9 0.588
Rc [MPa*10-1]
Mechanical properties of bending wood from standing tree trunks tested with the ultrasonic
method.
The ultrasonic technique made it possible to test several dozen spruce trees (Picea abies Karst.)
and pine trees (Pinus sylvestris L.) growing in Puszcza Augustowska (Augustowska Primeval Forest),
Puszcza Romnicka (Romnicka Primeval Forest) and in Experimental Forests of the Warsaw University
of Life Science in Rogów. Thus, a quick, repeatable and quasi-non-destructive (a need to make small
drills to install gauge probes on the trunks) method was used. Gauge probes were made by connecting
piezo-electric and electro-dynamical transducers [Dzbeński, Wiktorski 2004, 2007]. The method
has shown promise for becoming commonly used in wood science metrology.
6000
5900
5800
Wave velocity in sawn timber VtL [m/s]
5700
5600
VtL = 0,7267 VL + 2310,7

5500 2
r = 0,709
r = 0,842
5400
5300
5200
5100
5000
4000 4100 4200 4300 4400 4500 4600 4700 4800 4900 5000
Wave velocity along the standing tree trunks VL [m/s]
Figure 1. Correlation of ultrasonic wave velocity in standing tree trunks and sawn timber.
This paper presents results of a preliminary testing with the use of equipment of the Sylvatest Duo type
from Concept Bois Technologie – Concept Bois Structure, as in Sandoz’s studies, measuring selected
trunks of pine trees longitudinally and transversely. After cutting the trees, their butt ends (up to 1.5 m
long) were used to obtain little beams (specimens 1.2 meter long and 50 x 50 mm cross-section
according to the PN-77/D-04227 standard, along radii corresponding to four geographical directions,
separately from the near-circumferential and pith areas). Modulus of elasticity and static bending
strength were tested according to the PN-EN 408 standard. The results of sapwood testing were
correlated with ultrasonic velocity along the trunk VL (0.8 m long section), while of sapwood and
heartwood together – with the transverse ultrasonic velocity V⊥ in the trunk with a diameter
up to 50 cm. Above this allowable trunk diameter, the energy of the transmitted signal, measured by
a millivoltmeter, reaches a critically low level, and the receiving probe registers only random
disturbances. The measurement results are presented in tables 2 and 3.
22
6000
Wave velocity in sawn timber VtL [m/s]
5800
5600
5400
5200
5000
VtL = 0,3438 VL + 3886,2
4800 2
r = 0,067
4600 r = 0,259
4400 Wood with resinosis
Wood without resinosis
4200
4000
4000 4200 4400 4600 4800 5000
Figure 2. Correlation of ultrasonic wave velocity in standing tree trunks and sawn timber – including wood with
resinosis.
Table 2. Relationship between velocity of ultrasounds conducted longitudinally in trunks and selected
mechanical properties of (air-dried) pine sawn timber obtained from them, from the near-circumferential zone.
Type of sawn timber properties Regression equation f(VL) Correlation coefficient (r)
Bending modulus of elasticity Eg [MPa] Eg = 5.996 VL + 14214 0.867
Static bending strength Rg [MPa] Rg= 0.029 VL + 63.716 0.663
Table 3. Relationship between velocity of ultrasounds conducted transversely trunks and selected mechanical
properties of (air-dried) pine sawn timber obtained from them, from the near-circumferential zone.
Type of sawn timber properties Regression equation f(V⊥) Correlation coefficient (r)
Bending modulus of elasticity Eg [MPa] Eg = 12.829 V⊥ - 6460.9 0.711
Static bending strength Rg [MPa] Rg= 0.0501 V⊥ - 6.9027 0.796
The studies demonstrated that the presence of wood with resinosis has a significant impact
on the character of the regression equation. In the testing described above, the correlation coefficient
of ultrasonic velocity propagated along the trunk and air dried sawn timber obtained from them was
only 0.259 in case of population with specimens of wood with resinosis (fig. 2) and 0.842 in case
of population with specimens of wood with no resinosis (fig. 1). A parallel correlation with modulus
of elasticity (r = 0.867, fig. 3) and bending strength (r=0.663, fig. 4) looks similarly. Negative impact
of excessive presence of resin (the so-called resinosis) caused by a fungal infection suggests that
an attempt should be made to detect and locate rot in trunks of the tested trees. In case of transverse
ultrasonic measurements in the trunks, the confidence level of the presented correlation coefficients
is less than 0.90 due to the smaller sample group (5 trunks).
Detection of rot inside trunks of coniferous trees by means of the ultrasonic technique.
Longitudinal and transverse measurements of pine and spruce tree trunks were made using equipment
of the Sylvatest Duo type, having modified the method described by Sandoz [Sandoz, Lorin 1994,
Sandoz 1999]. Both the transverse heterogeneity of wood with rot (the „Z” coefficient of the trunk
degradation) and anisotropy of elasticity properties (coefficient „A” of anisotropy of ultrasonic
velocity - longitudinally and transversely in the trunk) were taken into account.
23
16000
Modulus of elesticity of sawn timber Eg [MPa]
15000
14000
13000
12000
Eg = 5,9959 VL - 14214
11000
r2 = 0,752
10000 r = 0,867
9000
8000
4000 4100 4200 4300 4400 4500 4600 4700 4800 4900 5000
Figure 3. Correlation of modulus of elasticity of sawn timber and ultrasonic velocity along standing tree trunks.
85,00
80,00
Bending strength of sawn timber Rg [MPa]
75,00
70,00
65,00
Rg = 0,0294 VL - 63,716
2
r = 0,44
60,00 r = 0,663
55,00
50,00
45,00
40,00
4000 4100 4200 4300 4400 4500 4600 4700 4800 4900 5000
Figure 4. Correlation of bending strength of sawn timber and ultrasonic velocity along standing tree trunks.
Using the phenomenon of usually non-symmetric distribution of defects (rot, hollows, knot groups,
etc.) in a cross section it is possible to describe structural transverse heterogeneity, determining
the socalled „Z” coefficient of the trunk degradation, in the following way:
Z = (V⊥ref – V ⊥)/V⊥ref * 100 % (1)
Where: V⊥ - transverse wave transmission time in the trunk above the root collar;
V⊥ref – reference velocity (in wood with no defects, usually at the altitude of 2.0 meters above
the ground).
The coefficient of transverse heterogeneity of material should be determined from the following
equation:
∆V⊥ V⊥1 − V⊥ 2
K= = × 100% (2)
V⊥ ref V⊥ ref
24
Where: V⊥1 and V⊥2 - registered wave transmission velocities in two perpendicular directions, across
trunks.
The bigger the difference between the above-mentioned velocities, the higher the transverse
heterogeneity of the material. The lower the measured velocities, the bigger the inner defect will be.
The bigger the values of the „K” coefficient from the equation (2), the more the rot tends to move
in relation to the pith. Examples of the measurement results are presented in table 4.
The percentage of U⊥’ [%] of the rot area (or another inner defect) of the D diameter in a cross-section
of the trunk depends on the propagation time (τ) and transverse wave velocity (V⊥) in the trunk with
an average diameter of (d), decreased to the value of (V⊥’ = d/τ’) as a result of longer time it takes
for the signal to pass through the inner rot area (τ’), i.e.:
∆τ ∆V⊥ D
= = ≈ U ⊥' (3)
τ V⊥ d
The percentage of the rot area U⊥’ can be also estimated by means of anisotropy of the centre elasticity
method (np. wood with density of ρ and modulus of elasticity of EVL) based on the following
proportion:
V V
A = L ≤ L' (4)
V⊥ V⊥
1,35 × EVL
Where: VL = [m / s ] having taken account of the Poisson number [Dzbeński 1984]
ρ
Table 4. Quality characteristic of tested spruce and pine trunks (selected examples).
Transverse Transverse
Reference velocity Degradation coefficients
Percentage of rot velocity of fibres heterogeneity
area U’⊥ [%] Z1 Z2 coefficient
V1⊥ref
[m/s]
V2⊥ref
[m/s]
V⊥1
[m/s]
V⊥2
[m/s] [%] [%] ∆V⊥ K = ∆V⊥ / V⊥ref [%]
a) b)
0 1269 1239 1329 1353 0 0 24 1.9
a) b)
0 1548 1408 1559 1548 0 0 11 0.7
a) b)
3 1440 1436 1510 1466 0 0 44 3.0
b)
6,5 1305 1250 1293 1318 1,0 0 25 2.0
17 1305 1250 1235 1070 5.4 14.4 165 12.6
31 1377 1354 1054 1075 23.5 20.6 21 1.5
c)
35 1171 1065 978 1089 16.5 - 111 9.5
42 1341 1579 837 1012 37.6 35.9 175 13.1
44 1463 1420 1049 1086 28.3 23.5 37 2.5
61 1538 1517 1322 1276 14.0 15.9 46 3.0
a) b) c)
V1⊥ref = V⊥1 accepted after verification; V2⊥ref = V⊥2 accepted after verification; Result rejected after verification
CONCLUSIONS
1) Velocity of ultrasonic wave propagation in coniferous standing tree trunks (pine, spruce) is a good
indicator of mechanical properties of sawn timber obtained from them. Satisfactory correlations
were achieved between longitudinal wave velocity (r=0.867) and transverse wave velocity
(r=0.711) in standing trees trunks, and modulus of elasticity of sawn timber obtained; parallel
correlation coefficients relating to bending strength (r=0.663; r=0.796), and compressive strength (r
= 0.588 > 0.5139).
2) An important correlation (r=0.842) between longitudinal ultrasonic velocity in standing tree trunks
and sawn timber obtained from them provides the basis for the above-mentioned correlations.
3) Strong resinosis of wood caused by a fungal infection results in a significant discrepancy
of the relationships indicated in conclusions (1) and (2) and makes it necessary to conduct studies
relating to presence and location of inner rot in coniferous tree trunks.
25
4) Irregularity of a shape or non-centric distribution of inner rot can be determined by a cross-wise

measurement of heterogeneity in transverse ultrasonic conduction in the trunk („Z1” and „Z2”
degradation coefficients); the higher their value, the more the rot tends to move in relation
to the pith.
5) Low transverse ultrasonic velocity in fibres is a direct proof of the inner rot presence. The size of
the rot can be determined on the basis of the „K” transverse heterogeneity coefficient in transverse
ultrasonic conduction in the trunk, i.e. by comparing the wave velocity outside the rot area
(e.g. at the altitude of 2.0 meters) to the wave velocity in a defected area of the butt end. The rot
dimensions can also be described by means of the „A” anisotropy coefficient in longitudinal
and transverse ultrasonic conduction in the trunk.
REFERENCES
1. Aratake S., Arima T., (1994): Estimation of modulus of rupture (MOR) and modulus of elasticity
(MOE) of lumber using higher natural frequency of log in pile of logs. II. Possibility of application
for sugi square lumber with pit. Mokuzai Gakkaishi. 40(9): 1003-1007
2. Breeze J.E., Nilberg R.H., (1971): Predicting by sonic measurements the strength of logs and poles
having internal decay. Forest Products Journal. No. 5: 39-43
3. Dzbeński W., (1981): Versuche zur Anwendung der Ultraschall-Messtechnik bei der
Gütesortierung von Konstruktionsholz nach seiner Festigkeit. Holzforschung und Holzverwertung.
No. 6:105-110
4. Dzbeński W., (1984): Nieniszczące badania mechanicznych właściwości iglastej tarcicy
konstrukcyjnej wybranymi metodami statystycznymi i dynamicznymi. (Non-destructive testing of
mechanical properties of coniferous construction sawn timber using selected statistical and
dynamic methods). Wydawnictwo SGGW-AR in Warsaw
5. Dzbeński W., Wiktorski T., (2004): Influence of some structural wood defects on ultrasonic wave
spectrum. Ann. Warsaw Agricult. Univ. SGGW, For. and Wood Technol. No. 55:156-163
6. Dzbeński a W., Wiktorski T., (2007): Wykrywanie zgnilizny wewnątrz pni drzew Ŝywych za
pomocą techniki ultradźwiękowej (Detection of rot inside living tree trunks using ultrasonic
technique). in. Technologia drewna – wczoraj, dziś, jutro. Studia i szkice na jubileusz profesora
Ryszarda Babickiego (Studies and papers in connection with professor Ryszard Babicki’s
anniversary). Instytut Technologii Drewna. Poznań. p.225-234.
7. Jayne B.A., (1959): Vibrational properties of wood as indices of quality. Forest Products Journal.
9(11):413-416
8. Kaiserlik J.H., Pellerin R.F., (1977): Stress wave attenuation as an indicator of lumber strength.
Forest Products Journal. 27(6):39-43
9. Marra G.G., Pellerin R.F., Galligan W.L., (1966): Nondestructive determination of wood strength
and elasticity by vibration. Holz als Roh-und Werkstoff. 10:460-466
10. Ross R.J., De Groot R.C., Nelson W.J., (1994): Technique for nondestructive evaluation
of biological degraded wood. Experimental Tech. 18(5):29-32
11. Ross R.J., McDonald K.A., Green D.W., Schad K.C., (1997): Relationship between log and lumber
modulus of elasticity. Forest Products Journal. 47(2):89-92
12. Sandoz J.L., Lorin P., (1994): Standing tree quality assessment using ultrasound. Proc. Of. First
European Symp. on Nondestructive Evaluation of Wood. Hungary Univ. of Sporon. Vol.2:493-502
13. Sandoz J.L., (1996): Ultrasonic solid wood evaluation in industrial application. 10th International
Symposium on Nondestructive Testing of Wood Proceedings. s.135. Lausanne
14. Sandoz J.L., (1999): Standing tree assessment using acusto-ultrasonic. Acta Hort. (ISHS) 496:269-
278 http://www.actahort.org/books/496/496_34.htm
15. Wang X., Ross R.J., McClellan M., Barbour R.J., Ericson J.R., Forsman J.W., McGinnis G.D.
(2000): Assessment of standing trees using a nondestructive stress wave technique. Forest Products
Laboratory. Research Paper 585.
26
Measurement of moisture content profiles in coated and uncoated Scots Pine using
Magnetic Resonance Imaging
Dr Jan Ekstedt 1 , Dr Anders Rosenkilde 1 , Dr Stéphane Hameury
M.Sc Magdalena Sterley1 , Dr Henrik Berglind
1
SP Technical Research Institute of Sweden, PO Box 5609, SE-11486 Stockholm, Sweden, magdalena.sterley@sp.se
Keywords: Moisture buffering capacity; Moisture profiles; Magnetic resonance; MRI Stray field imaging, SPI.
ABSTRACT
Non-invasive spectroscopic method of accessing information about moisture profile in wood is presented. The technique
used is proton magnetic resonance imaging (1H MRI) with a low-field Benchtop Imager Magnet and an effective Single-
Point-Imaging sequence (SPI sequence) and with a high-resolution orthogonal permanent magnets system and multi-echo
acquisition. Different examples of application of this method in wood science are presented. One example is the interaction
between indoor air fluctuations and Scots pine. SPI images of bound-water distribution in Scots pine with a spatial
resolution on a sub-millimeter scale were acquired when one of the orthotropic directions of the wood material was
exposed to typical indoor day-to-day moisture fluctuations. Another example is cyclic sorption experiments with coated
Scots pine exposed to day-to-day relative humidity step changes and water absorption and evaporation through the wood
coating. Even other examples with higher resolution moisture profiles describing water interaction with wood, coatings
and adhesives are presented.
INTRODUCTION
In a biogenetic and heterogeneous material like wood the action of water and moisture plays an
important role in the performance of the material as such but even more where wood is in combination
with other materials. These material combinations exhibits interfaces that sometimes are polar and
high energy interfaces causing accumulation of water. The spatial distribution of water in wood has
been studied by 1H MRI (proton Magnetic Resonance Imaging), which is a widely used versatile non-
invasive tool. The spatially resolved MRI of the moisture content (MC) in wood during drying using
Hahn spin echo acquisition has been demonstrated by several authors (Araujo et al., 1992).
The present paper presents research where the action of water and moisture are studied. The research
was part of doctoral thesis of Stéphane Hameury (2006) at KTH Stockholm and also part of an EU
funded project MARWINGCA (MAgnetic Resonance Imaging of Wood at the INterface with Glues,
Coatings and Air) (Berglind et.al 2006).
Moisture buffering capacity

Wood is a porous and hygroscopic material constantly undergoing transient changes in its moisture
content while trying to reach a state of hygrothermal equilibrium with its surrounding environment.
The moisture gradients developed within wood in response to climate fluctuations are accompanied by
e.g. shrinkage and swelling, moisture-induced stresses and mechano-sorptive effects, which may
ultimately lead to cracking or loss of load-bearing capacity. This affinity of wood to liquid water and
water vapour is usually perceived as a drawback by the building community, but it can also be an
advantage regarding the moisture buffering capacity of the building materials facing the indoor
climate. This capacity may be used as a passive way to regulate the relative humidity fluctuations
together with an efficient ventilation system (Osanyintola & Simonson, 2006). An increase in the
amount of wood facing the indoor environment will help to regulate the indoor relative humidity.
Measuring of moisture buffering capacity requires accurate method which follows moisture changes of
the exposed surface in the very shallow depth. With resistive sensor it is not possible. Another way to
study the moisture buffering effect of wood might be to use a Computed Tomography scanner (CT
scanner) (Wiberg, 2001, Rosenkilde and Arfvidsson, 1997, Lindgren, 1992;) which show capacity of
27
measuring moisture profiles in wood with a resolution of ca. 240 µm. In the present paper the moisture
profiles of uncoated and coated wood in response to dynamic moisture cycles obtained with MRI are
presented (Hameury and Sterley 2006, Hameury 2006).
Measurement of wood surface layer drying (MARVINGCA project)

Wood drying is an important industrial process in the sawmilling industry since it has a great impact
on both the product quality and the production economy manufacturing costs. Hence, several scientists
have aimed to improve the industrial drying process by developing theoretical models with the aim
to analyse the drying process. These models need relevant experimental measurements in order to
verifying their results. A weak point in some of those models has been the description of the moisture
transport above the fibre saturation point, FSP, and the behaviour at the surface interface. Therefore,
the moisture transport behaviour has been studied near the surface interface between wood and air
based on high resolution measurements during drying using MRI (Rosenkilde et al 2004).
Water accumulation at the coating/wood interface (MARVINGCA project)

The use of surfactants is essential for controlling the colloidal stability of the dispersion during
synthesis, storage, application and film formation of surface coatings on wood, especially water-borne
coatings. During film formation, phase separation occurs and these surfactants in the water phase may
be mobilised and transported into the wood substrate. The surfactants can also be released by rain after
film formation and, either washed away, or transported into the wood substrate. The chemical nature
of surfactants, with a hydrophilic and a hydrophobic part in the molecule, causes them to accumulate at
interfaces, where high concentrations may occur. In order to study the moisture dynamics at the
wood/coating interface and especially the influence of surface-active substances, which under some
circumstances accumulate at interfaces measurements has been performed with the MRI technique.
Water evacuation from a curing glue line (MARVINGCA project)

Wood gluing can be split into three phases: glue spreading, assembly and pressing. Knowing the rate
of water evacuation from the glue line through evaporation to the air and absorption into the wood
during the different process steps with high accuracy is essential for modelling of the curing behaviour.
In order to quantify the velocity of these individual processes, evaporation and diffusion, high spatially
resolved MRI has been used to follow glue line curing for various substrates.

Images were obtained from the samples of Scots pine (Pinus sylvestris L.) shaped as small cylinders
with a diameter of 12 mm and a length of 18 mm to 22 mm and placed in NMR glass tubes. Details are
presented in Hameury and Sterley (2006) and Hameury (2006). The samples sealed with a wax varnish
on all sides except on the surface to be exposed to the climate fluctuations and were exposed to
periodical step changes in relative humidity typical of the daily indoor. The samples were placed for 8
hours a relative humidity of 79.5 % and thereafter the samples were then exposed for 16 hours at 24
% RH and 22ºC. The cyclic procedure was repeated for 4 days to reach an acceptable quasi-steady-
state and moisture profiles were then recorded every hour during the 5th day.
The measurements were made using a low-field MARAN Benchtop Imager Magnet operating at
24.5 MHz from Oxford Instruments Ltd. The 1-dimensional gradient set provided a maximum gradient
strength in the magnetic field of approximately 3 T.m-1 .The RF probe used was an 18 mm diameter
solenoid probe driven by a 300 W RF output amplifier. The measurement sequence was the Single-
Point-Imaging (SPI) also known as Constant-Time Imaging (CTI). The SPI sequence was chosen
because of its efficiency to measure bound water in wood under fibre saturation point. The spacial
resolution was 127 µm.
28
MARVINGCA project
The MRI measurements were performed at University of Surrey, School of Physics. Guildford, U. K.
with a high-resolution orthogonal permanent magnets system. The measurement sequences were multi-
echo acquisition similar in form that of Carr and Purcell. Profiles of the samples were measured using
the thin film coil using the standard pulse sequence (αx-τ(αy-τ-echo-τ)n).The magnetic field strength
on the sample was 0,7 T with a magnetic gradient strength of 17 Tm-1. The spatial resolution for the
measurements is in the order of 15-20 µm.

Figure 1 and 2 present the one-dimensional MRI profiles of bound water in Scots pine measured in the
two orthotropic directions along the moisture flow, i.e. the tangential and radial. These profiles are
presented as moisture concentration profiles. The profiles shown in Figure 1 for the tangential
direction indicate that only a narrow region interacts with the surrounding environment. The periodic
penetration depth is approximately 2 mm and after 9 mm no moisture changes occur with time. We
call this depth the Active Buffering Depth (ABD) to differentiate it from the penetration depth. In
Figure 2, where the axis of the sample is along the radial direction, the profiles match the annual
growth rings shown in the superimposed digital picture. In this case, it is not possible to define a
periodic penetration depth as was possible in the tangential direction because of the presence of
fluctuations induced by the annual growth rings. However, the ABD is still measurable and is
approximately 9−10 mm. Secondly, the exposed surface of the samples does not reach equilibrium
immediately. This feature may be explained by a considerable moisture surface resistance but it is
certainly not the only cause and we suspect that a retarded sorption process occurs in the wood. The
MRI technology and the sequence used in this paper could be of help to further analyse this
phenomenon. Moreover the Moisture Buffer Value calculated from the moisture profiles was almost
exactly the same as that determined by the gravimetric method (Hameury and Sterley 2006).
70
8h 100
65
6h 80
RH [%]
60 60
3h
40
55
1h 20
Moisture Concentration [kg.m ]
-3
50 0
8h 30min
0 4 8 12 16 20 24
45 Time [h]
9h 30min
40
11h
35
24h 30min
30
24h
25
20
15
10
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
Distance [mm]
Figure 1. 1D SPI magnitude profiles in the tangential direction of Scots pine at different times. The surface exposed
to cyclic humidity changes is to the left. The resolution is ca. 127 µm.
29
100
70
80
8h
RH [%]
60
40
60 3h 20
0
1h 0 4 8 12 16 20 24
Time [h]
Moisture Concentration [kg.m ]
-3
50
40 11h
19h
30 8h 30min
24h
20
10
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
Distance [mm]
Figure 2. 1D SPI magnitude profiles in the radial direction of Scots pine at different times. The surface exposed to
cyclic humidity changes is to the left. The resolution is ca. 127 µm. The profiles are superimposed on a digital
picture along the axis of the cylinder of the wood sample.
Wood drying
Measurements were made from fresh condition and down to end use moisture content with Scots pine.
The measurements in the surface layer showed the very early development of a dry zone close to the
surface interface, Figure 3. In that zone or shell the moisture content was below the fibre saturation
point, FSP, even though the bulk moisture content was far above the FSP.
150
150
275
250 800
125
125
225 700
200
Moisture content (kg/m )
3
100
100 600
Moisture content %
Moisture content %
175
500
150
75
75
Moisture
125 400
50
50 100 300
75
200
25
25 50
100
25
0 0 0
00 52
52 104
104 156
156 208
208 2 60
260
-100 0 100 200 300 400 500
D istance
Dis tanc e (µ m)
m) Distance (µm)
a) b)
Figure 3. Moisture content profiles for Scots pine during drying:
a) o  0.5 h, x  1 h, ∆ 2 h, ∗ 4 h, • 7 h, ♦ 19 h, +  24 h,
b) The actual surface is at 0 µm on the distance axis, ♦ 0 min, ■  5 min, ▲  9 min, *  16 min, •
41 min, +  56 min, ∆ 1254 min.
Figure 3 b) show that after 9 minutes of drying the moisture profile is almost flat. The moisture
profiles remain approximately flat down to a moisture content slightly above 30% at 41 minutes.
At this point, a gradient starts to develop at the surface. The profiles recorded after 41 minutes all show
a gradient from the surface interface towards the bulk. This behaviour showing almost flat profiles
above a moisture content of approximately 30% and a gradient developing from the surface at lower
moisture contents has been reported before, Tremblay et al. (2000); Wiberg (2001); Rosenkilde and
Glover (2002) and Salin (2002). Wiberg (2001) reported about what he called the “dry shell
formation” which means a surface layer where the moisture content is below the fibre saturation point,
30
of approximately 30%. This indicates that present measurements of moisture content are obtained
within that dry shell. Furthermore it is possible to see the very early development of that shell with
a much higher spatial resolution than reported in Tremblay et al. (2000) and Wiberg (2001).
Wood coating
Figure 4 shows the spatial water distribution in at the coating/wood interface. The water proton density
profiles (green start profile before water ingress and red profile after 120 hours) has been
superimposed on an ESEM image of coating (thickness approx. 350 µm) applied to wood. The water
proton density is proportional to the moisture content. It is clearly seen that accumulation of moisture
occurs at the interface between coating and wood, at “distance 750”. This behaviour presented with
MRI technique can be of grate importance for durability of this coating and wood.
Figure 4. Proton density profiles through an acrylic coating on wood during water ingress
Curing of the glue line

In figure 5 a, glue has been placed on a glass cover slip without any upper substrate. The glue signal is
reduced due to chemical hardening and drying through evaporation. The evaporation velocity, defined
as the change of glue line thickness with time, was proportional towards time with 1.6 µm/minute
during the first 70 minutes. In figure 5 b, glue has been placed between a wood substrate and a glass
cover slip. The diffusion velocity into the wood, defined as the change of test piece thickness with
time, was proportional towards time with 0.6 µm/minute during the first 85 minutes.
700000 250000
glue glue
600000
200000
17 minutes
500000 34
Intensity (a.u.)
Intensity (a.u.)
51 17 minutes
68 150000 51
400000 glass
85 119
119 170
300000 255
170 100000
glass 255 850
elastomer
200000 425
air
850 50000
100000 wood
0 0
0 500 1000 1500 2000 0 500 1000 1500 2000
Position (microns) Position (microns)
a) b)
Figure 5 a) Glue curing profiles for glue spread on a glass substrate. The signal to the left is a marker tape. b) Glue
curing profiles for glue spread between a wood and a glass substrate. The elastomer facilitates a precise
determination of the thickness of the glue line.
Combining these separate measurements yields the glue-spreading situation, where the water
evacuation should be around 2.2 µm/minute. The second measurement indicates that during the
assembly, the water evacuation into each wood substrate would be 0.6 µm/minute.
31
CONCLUSIONS
The MRI method presented in this article based on the 1H MRI technique and a standard SPI sequence
is suitable to probe moisture profiles in wood materials at a low moisture content with high resolution.
It was confirmed that the moisture exchange between the indoor environment and wooden material is
confined to a few millimetres below the wood surface during typical daily moisture fluctuations and a
Moisture Buffer Value can be calculated from the NMR profiles with good accuracy. The advantage of
the method is however the spatial information obtained, something that the gravimetric method cannot
provide.
Moisture content profiles with high spatial pixel resolution, 21 µm, have been measured in Scots pine
heartwood while drying from wet conditions to near equilibrium in the surface layer, (0 – 300 µm).
It was shown that when the mean moisture content in the surface layer decreases to the fibre saturation
level, gradients starts to build up from the surface towards the bulk of the sample, although the bulk
moisture content is far above the fibre saturation. Furthermore, the results presented in this study imply
that there is a dry shell forming in the surface layer shortly after drying has begun.
The MRI technique applied to wood coating and wood adhesive research show possibility to analyze
water transport phenomena through the interface which can explain performance of different products
and can be useful for development of new products.
REFERENCES
1. Araujo, C. D., MacKay, A. L., Hailey, J. R. T., Whithall, K. P., and H. Le. (1992). Proton
magnetic resonance techniques for characterization of water in wood: Application to white
spruce. Wood Science and Technology, 26(2), 1001–113.
2. Berglind H., Ekstedt J., Rosenkilde A., Salin J-G., McDonald P., Bennett G., Keddie J.L.,
Brands G. and P. Jokinen (2003) Magnetic Resonance Imaging of wood at its interface with
glue coating and air. Final report to EC MARWINGCA. Trätek P0305014. ISSN 1102-1071.
3. Hameury, S. (2006). The hygrothermal inertia of Massive Timber Constructions. Doctoral
Thesis. KTH Architecture and the Built Environment, Stockholm Sweden. ISSN 0349-5752.
4. Hameury, S. and M. Sterley. (2006): Magnetic resonance imaging of moisture distribution in
Pinus sylvestris L. exposed to daily indoor relative humidity fluctuations. Wood Material
Science and Engineering, 2006,1, pp:116-126.
5. Lindgren, O. (1992). Medical CT-scanners for non-destructive wood density and moisture
content measurement. Doctoral Dissertation, Tech. Rep. No. 1992:111, Luleå University of
Technology.
6. Osanyintola, O. F. and C. J. Simonson. (2006). Moisture buffering capacity of hygroscopic
building materials: Experimental facilities and energy impact. Energy and Buildings (In Press).
7. Rosenkilde, A., Gorce J-P. and A. Barry. (2004) Measurement of moisture content profiles
during drying of Scots pine using magnetic resonance imaging. Holzforschung, 58, 138-142.
8. Rosenkilde, A., and P. Glover. (2002). High resolution measurement of the surface layer
moisture content during drying of wood using a novel magnetic resonance imaging technique.
Holzforschung, 56, 312–317.
9. Rosenkilde, A. and J. Arfvidsson. 1997. Measurement and evaluation of moisture transport
coefficients during drying of wood. Holzforschung 51, 372-380.
10. Salin, J-G. 2002. Theoretical analysis of mass transfer from wooden surfaces. 13th International
Drying Symposium, Aug. 27-30, Beijing, China
11. Tremblay, C., A. Cloutier and Y. Fortin. 2000. Experimental determination of the convective
heat and mass transfer coefficients for wood drying. Wood Sci. Technol. 34:253-276.
12. Wiberg, P. (2001). X-ray CT-scanning of wood during drying. Doctoral Dissertation, Tech.
Rep. No. 2001:10, Luleå University of Technology.
32
Assessment of Early Stage Fungal Decay of Wood
by FT-NIR-spectroscopy
Karin Fackler1,2, Barbara Hinterstoisser3, Manfred Schwanninger4, Cornelia Gradinger1,2, Ewald

Srebotnik1,2 and Kurt Messner2
1
Competence Centre for Wood Composites and Wood Chemistry, St.-Peter-Straße 25, A-4021 Linz, Austria
2
Institute of Chemical Engineering, University of Technology Vienna, Getreidemarkt 9, A-1060 Wien, Austria
3
Department of Material Sciences and Process Engineering, BOKU – University of Natural Resources and Applied Life
Sciences, Peter Jordan Straße 82, A-1190 Vienna, Austria
4
Department of Chemistry, BOKU – University of Natural Resources and Applied Life Sciences, Muthgasse 29, A-1190
Vienna, Austria
Keywords: beech wood, chemometrics, decay fungi, FT-NIR, spruce wood
ABSTRACT
The most efficient wood-rotting fungi are basidiomycetes and ascomycetes. In general, microbial decay processes go along
with a loss of wood quality. For example, brown rot decay leads to a rapid decrease in wood strength already in early
stages of growth. On the other hand, a projected biodegradation may enhance the quality when sound wood or surface
properties of wood are modified by means of biotechnology. The well-known standard methods to assess changes in
infected wood are time-consuming, and therefore unacceptable for efficient process-control and quality-assurance.
We could show, that fungal infestation of hardwood (Fagus sylvatica L.) and softwood (Picea abies L. Karst) can be
assessed by means of FT-NIR spectroscopy in combination with uni- and multivariate data analysis. The rapid method is
shown to be suitable for these purposes, presumed the regression models are well chosen. The results of degradation
experiments of wood blocks, veneers, and wood shavings demonstrated that different types and stages of decay could be
distinguished and lignin content after decay could be estimated from FT-NIR spectra collected from the degraded wood
surfaces and milled wood. Spectral data were subjected to principal component analysis (PCA) and uni- and multivariate
regression models were calculated. Mass loss after degradation and lignin content could be estimated from degraded wood
surfaces and it was possible to estimate the selectivity of white rot fungi for the preferential degradation of lignin.
INTRODUCTION
Wood is colonized and degraded by a variety of micro-organisms. Wood quality suffers from
microbial decay processes. Conventional methods are not suitable to assess early infestation of wood
on the one hand and to evaluate the potential of basidiomycetes for lignocellulose biotechnology
processes such as bio-pulping (Akhtar et al. 1997) on the other hand. Selective white rot fungi remove
more lignin than polysaccharides in early decay stages which makes them interesting as
a biotechnological tool for lignocellulose processing industries. The pretreatment of lignocellulosic
materials with white rot fungi to facilitate the disintegration of wood in the refining processes during
thermomechanical pulp production or to increase the accessibility of wood polysaccharides to cell wall
degrading enzymes to yield monomeric sugars for bioethanol production are two examples of many
that have been investigated (Akhtar et al. 1997; Amirta et al. 2006). Alternatives to rapidly assess
fungal action on lignocellulose in early stages are needed. Fourier transform near-infrared (FT-NIR)
reflectance spectroscopy has been proven a powerful tool to estimate both chemical and physical
parameters of wood (Tsuchikawa 2007). NIR absorption bands arise from overtones and combination
bands caused by vibrations of C-O, O-H, C-H, and N-H groups, which have their fundamental
molecular vibrations in the mid-IR (MIR) region, but also information about wood density (Thygesen
1994) is reflected in FT-NIR spectra. Kelley et al. (2002) were able to predict mass loss of brown
rotted softwood from reflectance spectra of wood meals.
This paper focuses on qualitative and quantitative analysis of FT-NIR spectra from wood surfaces and
wood meals to assess the changes in the composition of wood caused by the action of the fungi (brown
33
rot and white rot). For that purpose our recent work is reviewed, but also new examples of the analysis
of the spectral data are presented.

Fungi
The origin of the cultures is the collection of the Institute of Chemical Engineering, TU Vienna. White
rot (WR) fungi: Bjerkandera adusta, Ceriporiopsis subvermispora CBS 347.63 and FPL 105.752,
Hypoxylon fragiforme, Oxyporus latemarginatus, Phanerochaete chrysosporium, Phlebia brevispora,
Phlebia radiata, Phlebia tremellosa, Trametes cervina, Trametes versicolor ZIM L017 and CTB 863
A, Tyromyces chioneus DSM 5242. Brown rot (BR) fungi: Coniophora puteana CBS 237.91,
Gloeophyllum trabeum, Poria placenta MAD 698. The cultures were maintained on malt extract agar
(MEA, Fluka) slants, and pre-cultivated for two weeks on MEA plates before use.
Incubation of veneers, blocks, and shavings
Veneers (30 mm x 50 mm x 1.6 mm) were steam sterilised for 15 min and soaked in 2% (w/v) corn
steep liquor (CSL, Agrana, Austria) containing suspended fungal mycelium (one malt extract agar
plate overgrown by the fungus was mixed with 150 ml 2% (w/v) sterile CSL in a Waring blender for
30 s at full speed). Then they were put in agar dishes (9 cm Ø) with 25 ml water agar (1.5%) and beech
tooth picks as spacers to allow growth of the fungi throughout the entire wood surface in a moisture-
saturated atmosphere. The Petri dishes were incubated at 28°C for up to ten weeks. Steam sterilised
spruce wood blocks (1 cm x 1 cm x 3 cm) were inoculated following the procedure of EN 113 and
incubated for up to 20 weeks at 28°C. Fresh spruce wood shavings (3 g, particle size 0.4 – 2 mm, 29.0
% total lignin content, 100 % initial moisture content) were steam sterilised at 121 °C for 15 min and
inoculated with 3 ml of mycelium suspension (1 MEA plate / 70 ml 1.5 % (w/v) corn steep liquor,
200 % moisture content). Samples were incubated for up to 14 days at 28°C.
Sample preparation for FT-NIR analyses
After incubation, the fungal mycelia were removed from the surface of the decayed veneers and the
samples were dried for one week at 50°C. Then, they were milled (Retsch Ultra Centrifugal Mill ZM
1000, fixed ring sieve, 80 µm holes) and oven-dried at 50°C for another week. Finally, the milled
wood samples were extracted according to Schwanninger and Hinterstoisser (2002), except
cyclohexane was used instead of benzene (Fengel and Przyklenk 1983), and dried for another week at
50°C. Control samples (non-treated and steam-sterilised) were dried, milled, and extracted in the same
way.
FT-NIR spectroscopy and data analysis
FT-NIR reflectance spectra were recorded at ambient temperature using a fibre-probe connected to a
Bruker FT-IR spectrometer (Equinox 55) (Schwanninger et al. 2004). Ten to twenty spectra from areas
(9-10 mm²) of the solid samples at random positions on the front and backside of the veneers and four
replicate spectra of milled samples and extracted milled samples were recorded (Fackler et al. 2007b).
Spectra were processed by means of 17-points smoothing filter and a second order polynom to obtain
2nd derivatives (Savitzky and Golay 1964) with OPUS software (version 5.5, www.brukeroptics.de).
Partial least squares regression (PLSR) models of average spectra were calculated and optimised with
the software OPUS Quant 2. Principal component analyses (PCA) of average spectra were carried out
after mean-centering the data with the software The Unscrambler” (www.camo.com).

NIR Spectra
Figures 1A, B show 2nd derivatives of FT-NIR reflectance spectra from spruce and beech wood
surfaces before and after degradation by white and brown rot fungi. Calculation of the 2nd derivatives
helped accentuating differences between the spectra and served as data pre-treatment for PCA.
Changing lignin contents of spruce wood caused by degrading fungi can be estimated within the
wavenumber region from 6080 to 5800 cm-1 (Fig. 1A). Bands deriving from the aromatic ring (C-H)
and C-H vibrations of methyl and methylene groups are visible in this region (Shenk et al. 2001).
34
Lower lignin contents – e.g. caused by selectively delignifying fungi – lead to lower amplitude values
of the 2nd derivative spectra. Particularly the amplitude minimum near 5980 cm-1 correlates with the
lignin content of milled spruce wood (Schwanninger et al. 2004) and spruce wood surfaces (Fackler et
al. 2007c). In beech wood spectra (Fig. 1B) C-H deriving bands of xylan superimpose this region.
They show minima near 6000, 5955, 5880, and 5800 cm-1. Thus, in hardwoods, a local maximum from
lignin occurs between two minima (6000 and 5955 cm-1). Changes of lignin contents can be estimated
from this spectral region. Fungal wood decay leads to changes of the polysaccharide contents which
are reflected between 6750 and 6230 cm-1. The bands have mainly been assigned to the 1st overtone
of the fundamental OH-stretching vibrations of cellulose (Tsuchikawa and Siesler 2003a,b).
1.2E-05 1.0E-05
(A) 9.0E-06
(B)
L 5.0E-06
6.0E-06
NIR 2nd derivative

3.0E-06 0.0E+00
0.0E+00 PS
PS PS
-5.0E-06
PS PS PS
PS -3.0E-06
PS L -6.0E-06 -1.0E-05
-9.0E-06 native beech

L+PS
native spruce C. subvermispo ra (sel. white ro t) -1.5E-05
-1.2E-05 T. versico lo r (sim. white ro t)
T. versicolo r (white ro t) L
C. puteana (bro wn ro t) G. trabeum (bro wn ro t) L+PS
-1.5E-05 -2.0E-05
6750 6600 6450 6300 6150 6000 5850 5700 5550 5400 6750 6600 6450 6300 6150 6000 5850 5700 5550 5400
-1
Wavenumber [cm-1] Wavenumber [cm ]
nd
Figure 1. 2 derivatives of characteristic FT-NIR reflectance spectra of wood surfaces after prolonged fungal decay.
The region between 6750 and 5400 cm-1 that was used for PCA is shown. Lignin and polysaccharides assigned bands
are marked as (L) and (PS), respectively: (A) Spruce wood after 20 weeks degradation by T. versicolor and C.
puteana. (B) Beech wood after 10 weeks degradation by C. subvermispora, T. versicolor, and G. trabeum.
Multivariate data analysis: PCA

Principal component analysis allows the projection of multi-dimensional data onto a few orthogonal
features, called principal components (PCs), constructed as a linear combination of the original
variables – e.g. the data points of FT-NIR spectra – to maximise the description of the data variance.
The aim is to reveal the sharpest low-dimensional projection (i.e. the most informative projection) to
find clusters of similar samples. The 2nd derivatives of veneer surface spectra between 6750 and
5400 cm-1 have been subjected to PCA (92 samples, 350 data-points per sample, Fig. 2). This region
includes bands assigned to the 1st overtones of molecular structures of carbohydrates (semi-crystalline
cellulose: 6718 cm-1, crystalline cellulose: 6450 cm-1 and 6287 cm-1 (Tsuchikawa and Siesler 2003
a,b). Lignin dominates the region between 6100 and 5750 cm-1. Beech and spruce wood samples are
separated along the PC1 axis (68% of the variance). PC2 (17% of the variance) separates the samples
in white and brown rot degraded ones. The PC1-PC2 (Fig. 2A) plot reveals two main clusters (beech
and spruce wood) which are divided into four and three sub-clusters: selective white rot, simultaneous
white rot (found only in beech), brown rot, and native samples. The PC1 loading spectrum (Fig. 2C)
shows the highest positive loadings at bands assigned to xylan 5940 and 5805 cm-1 and negative
loadings at 5980 (lignin) and 5850 cm-1 (cellulose). Thus, beech wood samples with lower lignin
contents (lower than the average) and higher xylan contents score negatively on PC1, spruce samples
score positively. Degraded beech wood samples score less negatively on PC1, indicating that xylan,
which dominates the PC1 loadings, was degraded by each of the investigated decay fungi. The PC2
loading spectrum is dominated by lignin derived bands: positive loadings at 5975, 5900; and 5775 cm-
1
; and negative loadings at 6040, 5930, 5844, and 5750 cm-1 reflect more or less the shape of an
inversed second derivative lignin spectrum. Higher lignin contents are reflected in lower amplitude
35
values near 5980 cm-1 (Fig. 1A). Thus, wood samples subjected to brown rot with higher lignin
contents score negatively and selective white rotted samples score positively on PC2.
PC3 (5% of the data variance) separates the samples according to the weight loss of wood caused by
the decay fungi. Polysaccharides dominated bands prevail the PC3 loadings spectrum (6750 cm-1:
amorphous cellulose, hemicelluloses, 6450 cm-1 semi-crystalline cellulose, 6287 cm-1 crystalline
cellulose). Non-degraded samples with high polysaccharides contents score positively on PC3,
samples with lower polysaccharides contents score negatively. Obviously, the degradation of wood
polysaccharides contributes to a higher extent to weight loss than the degradation of lignin. The cluster
with non-degraded samples overlaps with that of brown rotted samples, indicating that early brown rot
cannot be reliably detected. Within the white rot cluster, wood samples separate between those that
had been subjected to selective white rot fungi that score positively on PC2 and PC3, and
simultaneously white rotted samples that score positively on PC2 and negatively on PC3. The PC2-
PC3 scores plot further demonstrates that white rot fungi in general are able to degrade lignin of spruce
wood more selectively than that of beech wood. These findings were confirmed by quantitative
analysis of the same NIR spectra (Fackler et al. 2007a).
beech wood spruce wood high lignin low lignin

2.E-04 2.E-04
.
low weight loss
.
(A) (B)
non-degraded
low lignin
PC 2 (17 % of variance)
selective w hite rot

PC 3 (5 % of variance)
selective
w hite rot
w hite rot early brow n rot

0.E+00 0.E+00
simultaneous.
.
. .
high weight loss
w hite rot native spruce white rot
high lignin
brown rot
native beech brow n rot
brow n rot simultaneous w hite rot

prolonged brow n rot
-2.E-04 -2.E-04
-3.E-04 0.E+00 3.E-04 -2.E-04 0.E+00 2.E-04
PC 1 (68 % of variance) PC 2 (17 % of variance)
0
L L+PS
(C)
PS
PCA-Loadings
L L+PS
PS
PS L
PS
PC1 L L
PC2
PC3
0
6600 6400 6200 6000 5800 5600 5400
-1
Wavenumber [cm ]
Figure 2. PCA of spruce wood (circles) and beech wood (diamonds) surface spectra (second derivative) calculated
between 6750 and 5400 cm-1. (A) Scores plot PC1-PC2; (B) Scores plot PC2-PC3; (C) Loading spectra PC1, PC2,
PC3 [loading bands assigned to polysaccharides (PS), and lignin (L)].
Quantitative analysis - regression models to predict lignin content and weight loss
FT-NIR data were used to calculate regression models in order to predict the lignin content and weight
loss of fungal degraded spruce and beech wood. To obtain reference data, lignin was determined
according to (Schwanninger and Hinterstoisser 2002), weight loss was determined gravimetrically.
The regression models are summarised in Table 1: Linear models were developed to estimate the
lignin content of milled spruce wood and spruce wood surfaces using the amplitude minima of the 2nd
derivative spectra near 5980 cm-1. PLS regression models in the spectral region characteristic for lignin
(6080 – 5800 cm-1) were calculated for beech wood. The regression models to predict the weight loss
are mainly based on two characteristics reflected in FT-NIR spectra: 1.) changes of the wood
composition that have been discussed in the above chapter PCA, and 2.) reduced wood density
as a consequence of fungal action that leads to less intense FT-NIR reflectance spectra.
36
Table 1. Regression models to predict weight loss and lignin content of degraded spruce and beech wood calculated
from FT-NIR spectra.
Component Sample Form R² RMSECV1 [%] SD2 [%] Reference

Lignin Milled spruce wood 0.95 -- 0.3 Schwanninger et al. 2004
Surface of spruce wood 0.95 -- 0.3 Fackler et al. 2007c
Milled beech wood 0.81 1.0 -- Fackler et al. 2007b
Extractives free milled beech wood 0.96 0.6 -- Fackler et al. 2007b
Surface of beech wood 0.81 1.0 -- Fackler et al. 2007b
Weight loss Surface of spruce wood 0.93 4.4 -- Fackler et al. 2007c
Surface of beech wood 0.81 5.1 -- Fackler et al. 2007b,c
1
root mean square error of cross validation
2
standard deviation
Early fungal degradation of wood

Delignification of spruce wood shavings during solid state fermentations with white rot fungi were
determined with the FT-NIR method: differences in lignin contents between consecutive fermentation
days could be elucidated (Fackler et al. 2006). Fig. 3A shows the lignin contents of the shavings after
treatment with four strains of selective white rot fungi and one brown rot fungus (P. placenta) after 5,
10 and 14 days. Whereas most white rot fungi were able to delignify the wood shavings significantly
within 5 days, it took 14 days until wood shavings treated with the brown rot fungus P. placenta
showed the expected increase of the lignin content caused by selective degradation of polysaccharides.
Selective delignification of beech shavings by C. subvermispora can be reliably detected within
10 days, indicated by an increase of the local maximum near 5980 cm-1 reflecting a reduction of the
lignin content by 1 % point (Fackler et al. 2007a).
33 7.0E-06
(A) 0 days 5 days (B)
31 10 days 14 days
NIR 2nd derivative .

29 0.0E+00
Lignin [%]
27
25
-7.0E-06
23
C. subvermispora
21 sterilised beech
-1.4E-05
a
ta
ra
o ra
i at
os
en
po
6150 6070 5990 5910 5830 5750

r ad
i sp
ell
lac
is
Wavenumber [cm -1]

m
rev
er m
p
P.
t re
P.
b
bv
P.
P.
su
C.
Figure 2. (A) Lignin content of fungal treated spruce wood shavings determined with the FT-NIR method. (B) 2nd
derivative spectra of milled beech wood shavings that had been treated for 10 days with a selective white rot fungus.
CONCLUSIONS
FT-NIR spectroscopy serves as a quick and reliable analytical tool to assess fungal degradation of
wood caused by basidiomycetes. The method makes use of chemical and physical changes that occur
in spruce and beech wood, while white and brown rot fungi grow on it. Particularly, the changing
lignin content caused by fungal action turned out to be useful for this purpose. Several regression
models to estimate its content were developed. On the one hand, FT-NIR spectroscopy can be used to
monitor the kinetics of fungal action during short-term modification of wood shavings, and to evaluate
the delignification selectivity of novel white rot strains for industrial purposes. On the other hand,
early white rot degradation of wood can be detected within several days. Thus FT-NIR spectroscopy
is appropriate as a quality control tool to assess fungal wood decay in wood processing industries and
has great potential for process control in lignocellulose biotechnology.
37
ACKNOWLEDGEMENTS
The Competence Centre for Wood Composites and Wood Chemistry acknowledge the funding by the
Austrian government and by the federal governments of Upper Austria, Lower Austria and Carinthia.
REFERENCES
1. Akhtar, M., Blanchette, R. A., Kirk, T. K. (1997). Fungal delignification and biomechanical pulping
of wood. Advances in Biochemical Engineering/ Biotechnology, 57. T. Scheper (ed.). Springer, Berlin
Heidelberg, 160-195.
2. Amirta, R., Tanabe, T., Watanabe, T., Honda, Y., Kuwahara, M., Watanabe, T. (2006): Methane
fermentation of Japanese cedar wood pretreated with a white rot fungus, Ceriporiopsis subvermispora.
J. Biotechnol. 123, 71-77.
3. Fackler, K., Gradinger, C., Hinterstoisser, B., Messner, K., Schwanninger, M. (2006): Lignin
degradation by white rot fungi on spruce wood shavings during short-time solid-state fermentations
monitored by near infrared spectroscopy. Enzyme Microb. Technol. 39, 1476-1483.
4. Fackler, K., Schmutzer, M., Manoch, L., Schwanninger, M., Hinterstoisser, B., Ters, T., Messner,
K., Gradinger, C. (2007a): Evaluation of the selectivity of white rot isolates using near infrared
spectroscopic techniques. Enzyme Microb. Technol. 41, 881-887.
5. Fackler, K., Schwanninger, M., Gradinger, C., Hinterstoisser, B., Messner, K. (2007b): Qualitative
and quantitative changes of beech wood degraded by wood rotting basidiomycetes monitored by
Fourier transform infrared spectroscopic methods and multivariate data analysis. FEMS Microbiol.
Lett. 271, 162-169.
6. Fackler, K., Schwanninger, M., Gradinger, C., Srebotnik, E., Hinterstoisser, B., Messner, K.
(2007c): Fungal decay of spruce and beech wood assessed by near infrared spectroscopy in
combination with uni- and multivariate data analysis. Holzforschung DOI 10.1515/HF.2007.098 - in
press.
7. Fengel, D., Przyklenk, M. (1983): Vergleichende Extraktbestimmungen zum Ersatz von Benzol
durch Cyclohexan. Holz Roh Werkstoff 41, 193-194.
8. Kelley, S. S., Jellison, J., Goodell, B. (2002): Use of NIR and pyrolysis-MBMS coupled with
multivariate analysis for detecting the chemical changes associated with brown-rot biodegradation of
spruce wood. FEMS Microbiol. Lett. 209, 107-111.
9. Savitzky, A., Golay, M. J. E. (1964): Smoothing and differentiation of data by simplified least
squares procedures. Anal. Chem. 36, 1627 - 1639.
10. Schwanninger, M., Hinterstoisser, B. (2002): Klason lignin: modifications to improve the precision
of the standardized determination. Holzforschung 56, 161-166.
11. Schwanninger, M., Hinterstoisser, B., Gradinger, C., Messner, K., Fackler, K. (2004): Examination
of spruce wood biodegraded by Ceriporiopsis subvermispora using near and mid infrared
spectroscopy. J. Near Infrared Spectrosc. 12, 397-409.
12. Shenk, J. S., Workman, J. J., Westerhaus, M. O. (2001). Application of NIR Spectroscopy to
Agricultural Products. Handbook of Near-Infrared Analysis. D. A. Burns and E. W. Ciurczak (eds.).
Dekker, New York., 419-474.
13. Thygesen, L. G. (1994): Determination of dry matter content and basic density of Norway spruce
by near infrared reflectance and transmittance spectroscopy. J. Near Infrared Spectrosc. 2, 127-135.
14. Tsuchikawa, S. (2007): A review of recent near infrared research for wood and paper. Appl.
Spectrosc. Rev. 42, 43-71.
15. Tsuchikawa, S., Siesler, H. W. (2003a): Near-infrared spectroscopic monitoring of the diffusion
process of deuterium-labeled molecules in wood. Part I: Softwood. Appl. Spectrosc. 57, 667-674.
16. Tsuchikawa, S., Siesler, H. W. (2003b): Near-infrared spectroscopic monitoring of the diffusion
process of deuterium-labeled molecules in wood. Part II: Hardwood. Appl. Spectrosc. 57, 675-681.
38
Fingerprint traceability of sawn products using x-ray log scanning

and sawn timber surface scanning
Jens Flodin1, 2, Johan Oja3, Anders Grönlund4

1
SP Technical Research Institute of Sweden, Wood Technology, Skeria 2, 931 77 Skellefteå, Sweden, jens.flodin@sp.se
2
Luleå University of technology, Department of wood technology, Skellefteå Campus, Skeria 3, 931 87 Skellefteå,
Sweden
3
SP Technical Research Institute of Sweden, Wood Technology, Skeria 2, 931 77 Skellefteå, Sweden, johan.oja@sp.se
4
Luleå University of technology, Department of wood technology, Skellefteå Campus, Skeria 3, 931 87 Skellefteå,
Sweden, anders.gronlund@ltu.se
Keywords: Fingerprint, Traceability, X-ray log scanner

ABSTRACT
Traceability in the sawmilling industry is a concept that could be used to more effectively control the production process
and the utilization of the raw material. The fingerprint approach is a traceability concept that rests on the idea that every
piece of wood is a unique individual with unique properties and hence can be identified and separated if a sufficient
number of these properties are measured accurately enough. This study was hosted by a sawmill in northern Sweden and
was aimed at making the fingerprint connection between logs and the center yield sawn from those logs using length and
knot information. The 140 logs involved in the study were of Scots pine with top diameters spanning the range from 153 to
213 millimeters. The center yield sawn from these logs was of two dimensions. The smaller logs (153–187 mm) were sawn
with a 2 ex pattern to 50 by 100 mm, and the larger logs (174–213 mm) were sawn to 50 by 125 mm with a 2 ex pattern.
The data from the logs were collected at the log sorting station by an industrial one-directional x-ray log scanner in
combination with a 3-D optical scanner. The data from the sawn center yield were collected by an industrial cross-fed
surface scanning system situated in the sawmill’s green sorting station. Both systems are used in the sawmill's normal
continuous production. The results show that over 90% of all planks could be matched to the right log, which bespeaks
a great potential for further development and realization of fingerprint tracing as a tool for process control and process
improvement.
INTRODUCTION
Modern forestry and sawmilling companies often have sophisticated measurement equipment that
generates large quantities of data at an individual level. These data are collected at certain points along
the production chain, but are unfortunately almost exclusively used as a means to control the
production process close to the measurement point. Most of the generated data for a specific piece of
wood is therefore discarded as soon as the piece has moved past the measurement point. If the data for
each specific piece were to be collected and stored in a database, the final product could then “be
considered as an information intensive product” (Uusijärvi 2003). The challenge is therefore not to
generate data, but to connect the generated data to each individual piece of wood. The reconnected
data would make it possible to investigate and analyze both large and small sections of the production
chain. A good example is the connection between the diameter classes for logs in the log sorting
station and the volume recovery of sawn planks and boards. Without reconnection of data, one is
reduced to comparing physical properties for a larger group of logs with the physical properties of their
planks and boards. With traceability, i.e., reconnection of data, one is given the opportunity not only to
analyze and find the individual logs in the group that yield high recovery, but perhaps even more
importantly, to find the logs in the group that yield low recovery for a specific sawing pattern. Being
able to make this distinction then makes it possible to adjust process parameters such as log class
limits or sawing patterns for an overall higher recovery.
Since sawmills have a diverging flow, and modern sawmills have high production speed, the
tracing and storing of data is not well suited for manual labor. A better alternative for handling the
tracing and tracking is some form of automated identification (McFarlane and Sheffi 2002). One way
of identifying individual pieces of wood is to use the already existing measurement data and make
39
identification by means of the fingerprint approach (Chiorescu 2003). The fingerprint approach rests
on the foundation that each piece of wood has unique individual features. These can be both outer and
inner features. If these individual features could be measured accurately enough, it would then be
possible to identify individual pieces in the production chain in the same way that human beings can
be identified by the use of their fingerprints. The purpose of this study is to investigate whether the
important individual connection between log and sawn center yield can be made by using the
fingerprint approach based on length and x-ray information from the log sorting station and on length
and surface scanning information from the green sorting station.

The sawmill that hosted this study is a large-size mill situated in northern Sweden with an annual
production of approximately 400,000 m3 of sawn timber. The sawmill handles only Scots pine (Pinus
sylvestris), which was also the only species included in this study. The 140 logs that were involved in
the study were randomly chosen and had top diameters spanning the range from 153 to 213
millimeters. These logs were all sawn with a 2 ex pattern into two different center yield dimensions.
The smaller logs (153–187 mm) were sawn to 50 by 100 mm, and the larger logs (174–213 mm) were
sawn to 50 by 125 mm, making a total amount of 280 sawn center yield pieces. The sideboards
produced were not included in the study.
The data used in this study were gathered at two points in the production chain from systems
that are used in the sawmill’s daily production. The first point was the sawmill’s log sorting station
where data from the logs were gathered with a one-directional x-ray log scanner from Rema Control
AB in combination with a 3-D optical scanner from MPM Engineering Ltd. The data extracted from
these systems were the log’s total length according to the 3-D scanner and the position and length of
the log’s knot whorls according to the x-ray log scanner. The second point of data gathering was a
Finscan Boardmaster surface scanning system situated at the sawmill’s green sorting station. The total
length and the positions of surface knots were recorded for each of the sawn planks. The order in
which the logs and planks passed the measurement systems was written down manually from the end
surfaces, which had been stamped with identification information (Skog and Oja 2007).
The analysis of the gathered data was performed using MatLab 7.3 (The MathWorks 2007).
The first step in the analysis was to investigate the correlation between the total length measurements
from the log sorting and green sorting stations. This was done by calculating the mean and standard
deviation values for all the logs’ lengths minus their corresponding planks’ lengths. Once the length
correlation was known, an algorithm was constructed to perform fingerprint matching between logs
and planks. The algorithm was designed to work in a three-step sequence. The first and second steps in
the sequence read the data into two matrices, first from the logs and then from the planks. Each row in
the log matrix contained the identification, the total length and the starting position and length of all
knot whorls found in the specific log. The information in the plank matrix was setup in the same way,
with the difference that it contained the lengthwise starting point and length of all surface knots found
on the planks. Due to edge effects in the filter, the x-ray log scanner needs a short distance before it
starts registering information. Therefore, knots that were situated within 200 mm of the top and butt
ends of the planks were disregarded.
The third and final step of the sequence was the actual matching procedure. The algorithm
worked iteratively by taking one plank at a time and comparing its surface knot positions with the
positions of knot whorls in all logs that had passed a length filtering. The length filter only allowed
logs with a total length within a span based on the length correlation mean and standard deviation. The
final matching was then made between the length-filtered log and the actual plank that showed the
highest agreement in knot positioning. When all planks had been iterated, the total percentage of
correct matches was calculated.
In order to find well-working settings for the algorithm, different values were tested for the log
length filter as well as for the distance over which knots were disregarded at the plank ends. These
40
values were between 3 and 10 cm for the length filter and between 100 and 400 mm for plank end
disregarding.
RESULTS
Figure 1 and figure 2 show how the agreement in the planks’ surface knots and the logs’ knot whorls
can be used to pair together a certain plank with a certain log.
Figure 1. A correct match shows good agreement between the plank’s surface knot positions (light gray) and the
log’s knot whorl positions (dark gray).
Figure 2. An incorrect match shows poor agreement between the plank’s surface knot positions (light gray) and the
log’s knot whorl positions (dark gray).
41
The results from the fingerprint matching are positive, with a total correct matching percentage of
approximately 90%. The length correlation between log sorting and green sorting station gave a mean
value of -1.2 cm and a standard deviation of 1.6 cm. This means that the planks are in general
measured slightly more than one centimeter longer than their corresponding logs and that the correct
log, with very high certainty, is to be found within +/- 5 cm from the planks length minus 1.2 cm. The
testing of different values for the log length filter and plank end disregarding didn’t give any drastic
results on the total percentage of correct matches. The percentage hovered slightly over and under 90%
with the different values.
When basing the settings for the log length filter on the length correlation to +/- 5 cm (with mean
correction) and using the original setting of 200 mm for plank end disregarding, a total of 92.5 % of
the 280 planks were matched to the correct log.
DISCUSSION
With promising results like this, one can look forward to what might lie ahead for this method of
tracing. One interesting spinoff is the ability to follow up if changes in process parameters such as, for
example, log class limits have had the desired impact on the sawn product. Another idea is to use the
fingerprint connected data to develop sorting models for the log sorting station, i.e., finding the outer
and/or inner characteristics of the logs that yield a specific quality and/or volume recovery.
This study was conducted on Scots pine only. It’s therefore hard to say how the fingerprint
tracing approach would work on Norway spruce (Picea abies) which is the other main species of wood
sawn in Sweden. Initially, it is thought that it will probably be more difficult, since Norway spruce
doesn’t have as clearly defined knot whorls as Scots pine because branches also grow in between the
main knot whorls in the living tree. It would, however, be very interesting to investigate the
possibilities of tracing Norway spruce with this method.
CONCLUSIONS
The results show a high potential for further development and realization of fingerprint tracing
between log sorting and green sorting into a practical application for process control and process
improvement.
ACKNOWLEDGEMENTS
The work was financially supported by the SkeWood program through TräCentrum Norr and through
the Swedish Agency for Innovation Systems (VINNOVA).
REFERENCES
1. Chiorescu, S. (2003): The forestry-wood chain, Simulation technique-Measurement accuracy-
Traceability concept. Doctoral thesis 2003:03. Luleå University of Technology, Skellefteå,
Sweden.
2. The MathWorks (2007): The MathWorks, Inc., 3 Apple Hill Drive, Natick, MA 01760-2098,
United States. <www.mathworks.com>.
3. McFarlane, D. and Sheffi, Y. (2002): The impact of automatic identification on supply chain
operations. International Journal of Logistics Management. 14(1):1–17.
4. Skog, J. and Oja J. (2007): Improved log sorting combining X-ray and 3D scanning –a
preliminary study. Proceedings of the COST E53 Conference on Quality Control for Wood and
Wood Products, Warsaw, 15–17 Oct. 2007.
5. Uusijärvi, R. (2003): Linking raw material characteristics with Industrial Needs for
Environmentally Sustainable and Efficient Transformation processes (LINESET). QLRT-1999-
01476 Final Report. SP Technical Research Institute of Sweden, Wood Technology,
(SP Trätek). Research Report no. P 0309034.
42
Quality Control for Improving Competitiveness of Wood Industries
COST E53 Conference, Warsaw, 15-17 October 2007
Influence of drying technique and process conditions on the drying

quality of beech wood (Fagus silvatica L.)
Željko Gorišek*1, Aleš Straže*
*University of Ljubljana, Biotechnical Faculty, Rozna dolina, C. VIII/34, SI 1000 Ljubljana, Slovenia
1
Contact email : zeljko.gorisek@bf.uni-lj.si
ABSTRACT
Complex structure and specific feature of European beech wood (Fagus silvatica L.) have great influence on its specific
drying characteristics, which are often negatively expressed as various drying defects. Therefore, four drying techniques,
i.e. controlled air drying, normal temperature kiln drying, vacuum drying and drying with high frequency (HF) were
compared to achieve the best technique. Quite the same quality, 32 mm thick beech wood samples were dried from green
condition to the final moisture content of 10%. At successive time intervals during drying processes wood quality was
evaluated regarding drying rate, time of drying, moisture content (MC) gradient, casehardening and occurrence of drying
defects. Drying with the HF was the fastest, followed by vacuum drying and kiln drying. In air drying wood kept good
quality, with the lowest MC gradient, however too high MC of wood was achieved. Low MC gradients were confirmed at
vacuum and HF drying, where the latter was applicable only at MC below fiber saturation point (FSP). Good drying
quality, concern casehardening and visual appearance was achieved at vacuum drying process, quite comparable to air
drying. In this research the vacuum drying was assessed as the most preferential techniques for drying beech wood.
Key words: air drying, kiln drying, vacuum drying, HF drying, drying rate, drying quality
INTRODUCTION
Beech wood (Fagus silvatica L.) represents in Slovenian woodworking industry very important part, in
a sense of quantity and quality as well as in a sense of value. Complex structure and specific feature
of European beech wood have great influence on its specific drying characteristics, which are often
negatively expressed as various drying defects, mostly depending on process conditions.
In dying practice a combination of different techniques has been applied to take advantage from them
or air pre-drying to minimize the energy consumption, despite unpleasant oscillating climate
conditions, has been used. For using more different drying techniques, especially for drying beech
wood, we can find some reasons: c.f. annually we have usually two main cutting seasons with
concentration of great amount of fresh logs and sawn wood; the drying capacity in the industry is
limited; with exploitation of natural condition we get better energy effectiveness of drying. Generally,
the aim of the drying process is to achieve the optimal quality in shortest drying time and with
minimum energy consumption. According to achieve sufficient drying quality of dried wood for wood
products with the best quality and with great added value we have to choose convenient drying
techniques for each drying period.
In our examination we investigated drying kinetics and quality of wood dried with different
techniques. With analyzing drying curve and rate of drying we would like to predict favorable drying
techniques for wood for different end use. With the model it would be able to choose the best
combination of drying techniques and to calculate the optimum MC at which change of drying
techniques would bring the best results.
To attain the main goal of examination we analyzed and compared four drying techniques: air drying,
normal temperature kiln drying, vacuum drying and drying with high frequency (HF). Drying
processes were evaluated regarding drying rate, time of drying, moisture content, moisture content
gradient, casehardening and occurrence of drying defects.
43

For the experiment we used fresh beech wood boards with 78±12% average initial moisture content.
High quality 32 mm thick boards were cut into four 60 cm long pieces with quite the same quality.
All boards were also cut to the same with of 80 mm. We tried to get board with tangential orientation.
We assured comparable material for further examination. In each drying process we controlled kinetics
of drying and quality of wood on 13 samples with dimension 32 mm x 80 mm x 600 mm. Beech wood
samples were dried from green condition to the final moisture content of 10%.
The accurate initial moisture content and moisture gradient were determinate by gravimetrical method
(EN 13 183-1). During drying processes we followed the MC changes with eclectic resistance
moisture meter (EN 13 183-2). At successive time intervals all samples were weighted and MC was
calculated. At the same time smaller controlled samples were taking out to define précis MC, moisture
content gradient and casehardening (Welling, 1993).
At the end of drying wood quality was evaluated regarding drying rate, time of drying, moisture
content, moisture content gradient, casehardening and occurrence of drying defects.
Due to very variable climate condition we followed the air drying during spring and winter period.
At the beginning of experiment the evaluation of MC, MC gradient and casehardening was done every
3 days and after two weeks of drying every 7 days. Air temperature and relative humidity were noted
at the same time intervals.
For »conventional« drying experimental kiln dryer was used (V = 1 m3). Drying schedule was adapted
due to experience and same specific characteristics of the kiln. Drying conditions in the kiln was
controlled with dry and wet bulb temperature through regulation system Vea. Every two hours dry and
wet bulb temperature, mass of wood, energy consumption, as well as MC on 6 places were registered.
Vacuum drying was simulated in small vacuum-pressure chamber. Climate condition in the chamber
was regulated with pressure and temperature. The industrial drying schedule was adjusted
to the possibilities of laboratory chamber. During drying process every 15 minutes the wood and
chamber temperature, moisture content and pressure were collected with date acquisition system
We used the HF only for drying wood below FSP; before fresh wood was carefully dried in the
controlled climate chamber to moisture content around FSP. HF generator had 4 stage regulations
of power. The wood samples were evaluated at successive time intervals (25 min.) with measuring:
mass, moisture content and temperature on the surface and in the centre of the samples.
RESULTS
Air drying
During air drying of beech wood we perceived and confirmed significant influence of season. The
average drop of moisture content per day in spring time was 6,1 %, meanwhile in winter period the
drying rate decreased to 2,2 %/day. So favorable drying was achieved in spring but the constant drying
rate was relatively short (about 1 week). First period with constant rate was finished at 46 % average
moisture content (Fig. 2A). But at less favorable winter period constant drying rate period lasted
at least 3 weeks and finished at lower average moisture content (at about 36 %). In both seasons more
than half of water was driven from wood during constant drying rate. As expected the drying rate was
continued with exponential decreasing and was finished with achieving the equilibrium moisture
content of environment.
The rate of air drying in spring as well as in winter time was good correlated with exponential
mathematical model (Fig. 1).
∆u a
= ( − k ( u − u c ))
∆t 1 + e
44
7,0
6,5
Spring:
6,0 2
R = 0.94
5,5 a = 6,1 ± 0,161
uc = 46,526 ± 1,132
5,0 k = 0,103 ± 0,007
4,5
∆u / ∆t [% / day]
Winter:
4,0 2
R = 0.89
a = 2,219 ± 0,062
3,5
uc = 30,303 ± 0,765
3,0 k = 0,199 ± 0,02
2,5
2,0
1,5
1,0
0,5
0,0
0 10 20 30 40 50 60 70 80 90
u [%]
Fig. 1 Drying rate during air drying in spring and winter.
More favorable drying during spring months induced in wood grater moisture gradient. Relatively fast
drying of surface caused that MC of surface fell below FSP and interrupted the effective flow
of capillary water from the core of the wood. Slower drying induced smaller moisture gradient so there
was very little danger for appearance of diffusion barrier on the wood surface.
80 0,3 80 2
70 70 1,75
60 60 1,5
Moisture content [%]
0,2
Drying rate [%/h]
Drying rate [%/h]

50 50 1,25
40 40 1
30 30 0,75
0,1
20 20 0,5
10 10 0,25
0 0 0 0
0 200 400 600 800 1000 1200 1400 0 50 100 150 200
Drying time [h] Dring time [h]
A B
80 2 80 16
70 1,75 70 14
60 1,5 60 12

Drying rate [%/h]
Drying rate [%/h]
50 1,25 50 10
40 1 40 8
30 0,75 30 6
20 0,5 20 4
10 0,25 10 2
0 0 0 0
0 10 20 30 40 50 60 0 0,5 1 1,5 2 2,5
Drying time [h] Drying time [h]
C D
Fig. 2 Drying curves and drying rates for A/ air, B/ kiln, C/ vacuum and D/ HF drying.
Kiln drying
45
Overall drying time in the kiln drying was 178 hours (heating 8 h, drying 143 h, conditioning 27 h).
At the beginning the drying was very intensive, but drying rate dropped very quickly and decreased
exponentially to the final moisture content (Fig. 2B).
The beginning of drying induced very sharp moisture gradient (Fig. 3). Consequences were seen
in generating drying stresses which could cause surface checking or casehardening at the end of drying
(Fig. 4). Therefore, kiln drying hat to be fished with conditioning.
Vacuum drying
The complete time of continuing vacuum drying was very favorable (58 h). The drying rate during first
period was not so high but last till MC dropped well below FSP (Fig. 2C). We concluded that quite all
free water was eliminated through capillary flow. Also the decrease of diffusion flow was not
significant as in other drying techniques (c.f. Ressel 1999).
The drying rate in vacuum was comparable with that achieved in kiln drying, but favorable conditions
lasted in vacuum drying much longer with smaller moisture gradient (Fig. 3). Despite of small
moisture gradient or potential the drying was relatively fast until 20 % average moisture content was
reached. Because of small differences between MC on the surface and MC in the core, also drying
stresses was not so distinctive (Fig. 4). The risk for split or wrap of wood was therefore very small.
HF drying
Before drying the wood in the dielectrics field of high frequency we hat to dry it to MC just about FSP,
so we have no date about effectiveness of HF drying of green wood. HF drying below FSP was very
effective and the rate of drying was 10 times higher than in others drying techniques (Fig. 2D). Before
drying with HF the MC in the core was a little higher than at the surface. The temperature at the core
of the wood increased faster than in the surface, therefore the condition for further drying was very
favorable. As in other techniques also in HF, drying rates decreased exponentially with decreasing
MC. Drying gradient was small but some casehardening occurred (Fig. 3 and Fig. 4) (c.f. Kobayashi et
al, 1999).
80 32
30
70
AIR 28 AIR
60
KILN KILN
26
VACUUM VACUUM
50
HF 24 HF
40
22
30 20
Board thickness Board thickness
A B
20
15 AIR
KILN
VACUUM
10 HF
5
Board thickness
C
Fig. 3 Moisture content gradients in three stages of drying 1 – initial state, 2 – around FSP and 3 – final
state for air, kiln, vacuum and HF drying.
46
2,5
1,5
Casehardening
1 AIR
KILN
VACUUM
0,5 HF
0
INIT. FSP FIN.
-0,5
-1
Fig. 4 State of casehardening in three stages of drying 1 – initial state, 2 – around FSP and 3 – final state for
air, kiln, vacuum and HF drying.
Constant drying rate period was significant only for air drying (Fig. 5 and Fig. 6). In other drying
techniques MC of surface dropt quickly below FSP and interrupted effective capillary flow of free
water from the center of wood. In this period the air drying is also competitive if the climate conditions
is favorable, for example in spring. Air drying below FSP was very slow and wasteful. In hygroscopic
range much longer, than in other techniques lasted favorable drying conditions in vacuum drying.
This is also the reason for smaller moisture gradient and very good quality of dried wood. Because of
small capacity HF drying was not competitive with other drying technique.
80
70
60
50 AIR
KILN
40
VACUUM
HF
30
20
10
0
0 1 10 100 1000 10000
Drying time [h]
Fig. 5 Comparison of drying curves for air, kiln, vacuum and HF drying
We can conclude that in some cases we get the best results with combination of different drying
techniques. For this reason we set up a model with which we are able to predict the best combination
of drying techniques and to calculate the optimum MC at which change of drying techniques brings the
best results. Because of great variability of wood some more date for verification is needed.
47
100
10
Drying rate [%/h]

AIR
KILN
0,1
VACUUM
HF
0,01
0,001
0,0001
0 10 20 30 40 50 60 70
Moisture content [u]
Fig. 6 Drying rate and their regression
In further investigation we will include in the model besides drying kinetics and drying quality also
energy consumption and costs of drying, which is by the way very specific for each firm.
CONCLUDING REMARKS
Drying with the HF was the fastest, followed by vacuum drying and kiln drying, whereas expected
duration of air drying was the longest.
In air drying wood kept good quality, with the lowest MC gradient, however drying to lower MC is too
long.
Low MC gradients were confirmed at vacuum and HF drying, where the latter was applicable only
at MC below fiber saturation point. Good drying quality, concern casehardening and visual appearance
was achieved at vacuum drying process, quite comparable to air drying. In this research the vacuum
drying was assessed as the most preferential techniques for drying beech wood.
With a mathematical model we are able to predict best combination of drying techniques
and to calculate the optimum MC at which change of drying techniques brings the best results.
REFERENCES
Kobayashi Y. et al. 1999. High performance drying using combination of HF and hot air under
atmospheric pressure. V: Wood Drying Research & Technology for sustainable Forestry beyond 2000.
6th International Wood Drying Conference, Stellenbosch, 25 th-28th January. 1999. Stellenbosch,
University in Stellenbosch, Faculty of Forestry, Department of Wood Science: 18-21
Neumann R. et al.. 1992. Comparison of conventional and convective vacuum drying of beech.
V: Understanding the Wood Drying Process: A synthesis of theory and practice. 3rd IUFRO
international wood drying conference. Vienna: 222-226.
Ressel, J.B. State of the art for the vacuum drying in wood working industry. Edinburgh 1999.
COST E 15.
Straže, A., Gorišek, Ž. 2007. CAE analysis and optimization of wood drying energy consumption
with use of air pre-drying Les. 56. 142-147
Welling J. 1993. Spezifikation und Üuberprüufung der Trocknungsqualität vor Schnittholz.
Holzbearbeitung 40 11, 56-62.
48
Mass customized production and design processes

for the parquet flooring industry
Hans Häuslmayer1, Manfred Gronalt1, Alfred Teischinger 2

1
BOKU - University of Natural Resources and Applied Life Sciences Vienna, Department of Economics and Social
Sciences, Institute of Production and Logistics, Feistmantelstr.4, A-1180 Vienna
2
BOKU - University of Natural Resources and Applied Life Sciences Vienna, Institute of Wood Science and Technology
Keywords: mass customization, parquet industry, process redesign, supply chain reconfiguration
ABSTRACT
We propose new ideas of applying mass customization concepts in the production of wooden floorings, especially for the
parquet industry. Due to market requirements this industry is obliged to observe special quality standards during
production. The current grading system of hardwood does not meet the specific demands for the appearance of the wood in
the final products. The standards are aimed at reducing the number of characteristics which define the various grading
classes regarding the appearance of the boards. Mass customizing parquet flooring could open new business opportunities
for the industry. In order to offer customers the possibility of individually composed floors, several types of customized
modules instead of the currently produced floor boards can be manufactured. The transfer of natural patterns of wood into
the final product by involving the customer in the design process was one goal of our research. Natural patterns and
features, which would not be used in a standard driven production process, could be allowed. The target and advantage of
the system is the higher added value of a low grade material and probably fewer complaints about defects of the final
product by the customer.
Our research explores these new production concepts in a case study approach with an Austrian manufacturer of parquet
flooring, using simulation as a method. We analyze several possible product-variety/customer-order mixes and their
distinctive production outcomes when producing different customized modules of floorboards. We also present new ideas
how an individualized supply chain process in the production of parquet flooring could look like.
INTRODUCTION
The term Mass Customization has been coined in the late 1980ies and has become subject to research
concerning operations management since Pine (1993). McCarthy (2004) defines Mass Customization
as the capability of companies to produce a relatively high volume of product options for a relatively
large market, which is demanding customized products without tradeoffs in cost, delivery and quality.
Implications of mass customization on the supply chain, concerning information and material flows
and the connection between product types and the decoupling point, have been researched for example
by Yang and Burns (2003). Mass Customisation strategies, such as postponement, can also have
effects on customer satisfaction. Reiner (2005) shows for a supplier of the telecommunication industry,
that postponement strategies such as labeling imply a substantial potential for customer-oriented
improvements.
Sigg & Jonas (2003) report on applications of mass customization and use of postponement strategies
as well as customized product configuration in furniture production. According to a Delphi-study of
the wood-working-industry undertook by the Fraunhofer Institute for Manufacturing Engineering and
Automation over 25% of furniture shall be produced in lot size 1 on industrial scale (ProWood 2004).
Working solutions for customized furniture design, production and ordering processes, like the famous
IKEA kitchen configurator, are well known solution within the wood-working-industry.
What concerns production processes of parquet flooring, in the last 30 years innovations have mainly
taken place on the technical equipment side. State-of-the-art descriptions of production processes from
the 1970ies (Kisseloff 1974) do not differ very much from analyses in the 1990ies (Hamberger 1995)
or our observations of production processes in a case study in 2005.
Mass customizing parquet flooring could open new business opportunities for this industry. In order to
offer customers the possibility of individually composed floors, several types of customized modules
49
instead of the currently produced floor boards can be manufactured. During the production process
several possibilities to modularize floorboards arise. We present and evaluate these possibilities on
their implications on the current production process. These results are then compared to the status quo
of parquet manufacturing at a case study company. We show how reconfigured production processes
affect storage levels of semi finished products, service rates and thus customer satisfaction.
CURRENT PRODUCTION PROCESSES

Analyzing the relevant business and production processes we identify the parameters currently faced in
the manufacturing of parquet flooring. The case study company produces floorboards in 5 different
qualities, 19 different wood sorts and allows for several different finishing possibilities, like lacquering
or oiling, from which the customers can choose freely. The production of floorboards is split in two
stages: First, the (visible) upper layer is produced in a make-to-stock environment, followed by
a buffer in which conditioning processes of the floorboards (according to the wood sort) take place.
This buffer is also functioning as Customer Order Decoupling Point (CODP). In the second step of the
production, a make-to-order environment can be stated. Floorboards are retrieved from the
conditioning chamber and then finished according to customer orders. Customers can order their
desired boards only in the distinctive qualities. Due to its character as a natural resource, wood is
subject to several anomalies and different quality patterns. The company is faced with uncertainties
and cannot be sure to produce the quality desired by the customer with the raw material it uses as input
to the process. Thus a lot of work-in-progress of different product variants has to be kept in stock at the
conditioning chamber, which is one of the major bottlenecks of the production process. A multi layer
3-strip parquet production process in general can be divided into four sub-sections (Kruse et al. 2003):
• Manufacturing of the upper layer (the one visible to the customer) and storing in a conditioning
chamber,
• Manufacturing of the middle layer or support layer and undercoating (a process not considered
and described here, as this is a standardized semi finished product),
• Press sizing of upper layer and support layer to a floorboard and finishing operations of the
surface of the floorboard, and
• Finishing Processes like milling of tongue and groove, quality end control and packaging.
After the storing of the upper layers in a conditioning chamber for a defined time, they are retrieved
according to customer orders and as a next step, press-sized with the middle layer. The retrieval
process marks the decoupling point of the process. It is here where each floorboard is assigned to
a specific customer requirement. One of the main difficulties in developing mass customization
concepts for the parquet industry lies within the discrepancy of the point in time, when a product
becomes unique and distinguishable from other products, and the point, where an order is assigned
to it. The former is called the Product Differentiation Point (PPD), the latter is the Customer Order
Decoupling Point. Figure 1 shows the current production processes of parquet and where the two
points mentioned above differ from each other.
SIMULATION STUDIES
In order to conduct simulation experiments for testing modularization strategies, we implemented real-
life production data from more than 13 months using the simulation software Arena®. The real-life
data are available for a production line which produces floorboards in 19 different types of wood and
up to five different qualities. Customer orders for the same time frame were implemented in the
simulation as well. Floorboards are produced according to our specifications and customers are able to
order also according to these specifications. The results of the simulation runs of the current
production processes were used as references to compare the performance of two new modularization
concepts. Table 1 shows the input-percentages of the different wood types, Table 2 shows the matrix
50
of the originally known quality outputs per wood type. It gives also an idea of the difficulties caused
by inconstant quality of the raw material and its different quality output per wood type.
Figure 1. Production Processes including PPD and CODP
Table 1. Wood Types Table 1.Quality Output per Wood Type

Wood Input Wood
qA qB qC qD qE
Type Percent Type
6 19,84% 1 4,47% 61,56% 24,61% 8,15% 1,21%
9 19,35% 2 7,72% 52,36% 30,50% 2,39% 7,04%
1 14,21% 3 0,00% 100,00% 0,00% 0,00% 0,00%
7 11,04% 4 0,06% 93,21% 0,06% 6,66% 0,01%
5 6,97% 5 0,00% 64,62% 0,06% 35,29% 0,03%
2 6,02% 6 6,83% 67,25% 10,77% 14,97% 0,19%
16 4,31% 7 5,15% 57,88% 27,07% 4,63% 5,26%
13 3,03% 8 0,00% 96,84% 0,00% 3,16% 0,00%
12 2,64% 9 5,70% 54,79% 25,16% 14,22% 0,13%
4 2,64% 10 0,00% 100,00% 0,00% 0,00% 0,00%
11 2,38% 11 9,72% 72,38% 16,06% 1,83% 0,00%
8 2,36% 12 0,00% 79,76% 0,24% 20,00% 0,00%
15 1,81% 13 0,00% 56,07% 0,00% 43,93% 0,00%
14 1,40% 14 0,00% 80,26% 0,00% 19,74% 0,00%
17 0,70% 15 0,00% 98,70% 0,00% 1,30% 0,00%
10 0,54% 16 0,00% 90,66% 0,00% 9,34% 0,00%
3 0,44% 17 0,00% 100,00% 0,00% 0,00% 0,00%
19 0,25% 18 0,00% 100,00% 0,00% 0,00% 0,00%
18 0,07% 19 0,00% 100,00% 0,00% 0,00% 0,00%
Three different scenarios were evaluated further, where scenario Status Quo is the core model of all
tested production environments and used as a reference to compare the new modularization concepts.
One product line of floorboards is produced in the mentioned 19 different wood types in 5 different
qualities. Customers can place their orders in the same distinctive categories. The CODP is located at
the retrieving process of the upper layers from the conditioning chamber. 18 lamellae of the same
quality and wood type are glued together to form one floorboard, the mixing of qualities is forbidden,
thus five different so-called M-boards per wood sort can be produced. Customer orders in this scenario
correspond exactly to the observed real-life data, the demand for M-Boards d(M) corresponds exactly
to the demand for the five qualities qA, qB, qC, qD and qE for every single wood sort.
51
Scenario Concept 1.0 does not allow for a mix of different qualities on a single M-Board, thus an M-
Board MA consists only of lamellae of the quality qA while MB can either be composed of quality qB
or quality qC and MC can either be composed of quality qD or quality qE. Customer orders are
modeled analogously, thus for example demand for floorboards MB equals the sum of the current part
demands for qB and qC.
Scenario Concept 1.1 differs again in the combination possibilities of M-Boards. Now a mix of
different qualities of lamellae on single floorboards is allowed. This does not cause any difference in
composing MA, but for MB, which now can be composed of any combination of the qualities qB and
qC, as long as the restriction of 18 lamellae per M-Boards is satisfied. The same is true for the
composition of MC. Customer orders are analogous to the former model. Of course these
modularization metrics can and should be adapted over time and from every producer, in order
to reflect his actual input quality distribution.
DISCUSSION
All relevant simulation scenarios resulted in different, but robust outputs, both in produced variants
and parameters such as throughput time or customer satisfaction. This allows for a comparison
between the models. Robustness was tested in several simulation runs of the same model subject to a
by-chance variation of time and quantity in both raw material input and customer demand. To compare
the output of each model, its order fulfillment results were set into relation to the distinctive customer
orders. Also benchmarks of the storing process at the CODP have been measured. Table 3 gives an
overview of the results. All benchmarks are presented in both the median and the mean, as the results
of the (up to 95 different) product variants differ on a broad scale.
Table 3. Output of the tested simulation environments
mean Status Quo Concept 1.0 Concept 1.1
Order Fulfilment Rate 80,02% 84,11% 84,42%
Rate of Number of
85,53% 89,62% 89,67%
Fulfilled Orders
CODP Stock Level

138,46% 168,31% 156,22%
Orders
CODP Stock Output

81,42% 82,63% 80,32%
CODP Stock Input
CODP Stock End-Level

CODP Stock Start-Level 34,57% 25,63% 26,72%
median Status Quo Concept 1.0 Concept 1.1
Order Fulfilment Rate 93,72% 92,45% 94,69%
Rate of Number of
98,01% 97,42% 98,22%
Fulfilled Orders
CODP Stock Level

98,55% 94,86% 98,07%
Orders
CODP Stock Output

92,94% 96,29% 93,16%
CODP Stock Input
CODP Stock End-Level

CODP Stock Start-Level 11,85% 7,30% 10,08%
52
RECONFIGURING SUPPLY CHAINS

The above described possibilities of individualization via the use of product-configurators also imply a
variety of new to be developed or restructured supply chains for the parquet industry. Before
describing new ways of ordering for customers and supply processes for the companies we have to
describe shortly the current last-mile-processes in the parquet industry. Figure 2 shows current and
possible future supply chains in the parquet industry.
Today, customers can buy parquet floorings either as a finished product in Do-It-Yourself-markets or
they can visit a specialised retailer who sells wooden floorings using conventional samples. The floor
boards are then delivered to the customer who can either lay the floor himself or has a professional do
the work. Another “high end” possibility would be inlays developed by designers and produced in
piece production.
The modularization concepts combined with a product configurator allow for more flexible supply
chains and order fulfilment processes in the parquet industry. Borders and competences valid in the
supply chains of today are dissolving, helping customers in becoming co-designers or co-producers.
Toffler coined the term of the so-called prosumer (cf. Toffler 1970, cited in Piller and Moeslein 2002).
The producer of today, who supplies wholesale and retail, can become a direct seller to customers via
internet configuration processes. A shift of market power and new cooperation concepts and other
pricing strategies will be the result of this process. For supply chains in the parquet industry one can
develop a number of scenarios.
Figure 2. Current and Future Supply Chains in the Parquet Industry
The first scenario completely leaves out the retail or wholesale processes. The customer directly orders
his individually created floor from the producer via a configurator and the internet. The producer
53
delivers directly to the customer, including a detailed laying plan for his individual floor. Another
possibility for the producer would be to offer the service of laying the floor at the customers home
himself. Another possibility for the customer would be to continue ordering via a specialised retailer.
With the help of the trained professional, the customer can then create and order his individual flooring
online. This concept has the advantage that customers who are not technically interested can also
create and buy an individualized wooden floor.
The next possible supply chain scenario implies the use of configuration terminals in DIY-markets.
Customers who don’t have the possibility of using internet applications at home or at work could then
also order online. Another possibility of using a configuration terminal in a DIY-Market or at a retailer
could be a function of the software, which creates automatically proposals of ready-to-take-home
floors. This concept would need a direct communication between the configurator and the inventory
management system. The customer can state his preferences and the configurator then presents
solutions and examples of individual floorings which can be created with the products the DIY-market
has on stock.
This concept could be extended to a vendor managed inventory application, which allows for a direct
filling of the DIY-markets inventory by the producer. Thus the software could offer the customer valid
due dates and more examples of individual floorings as a function of waiting times. It is possible that
customers face more possibilities when they wait for the next production batches finished. Creating
communication links between the production management system of the producer, the inventory
management system of the retailer or DIY-market and the product configurator could result in
committed and held due dates and higher customer satisfaction, combined with better use of raw
material output and lower stock levels.
REFERENCES
1. Hamberger, P. (1995): Die Parkettindustrie in Westeuropa. Eine Branchenanalyse unter

besonderer Berücksichtigung des österreichischen und deutschen Marktes. Dissertation,
Universität für Bodenkultur, Wien
2. Kisseloff, P. (1974): Die moderne Parkettherstellung – Stand der Technik. Holz als Roh- und
Werkstoff, 32. 127-134.
3. Kruse, K. / Teischinger, A. / Emmler, R. / Tobisch, S. (2003): Wood based Floor Coverings, in:
Johansson et al. Wood adhesion and glued products, State of the art Report, Cost Action E13.
4. McCarthy, I.P. (2004): Special issue editorial: the what, why and how of mass customization.
Production Planning & Control, 15, 4. 347-351.
5. Piller, Frank T., Moeslein, Kathrin (2002): From economies of scale towards economies of
customer integration. Value creation in mass customization based electronic commerce.
Arbeitsbericht Nr. 31 des Lehrstuhls für Allgemeine und Industrielle Betriebswirtschaftslehre
der Technischen Universität München.
6. Pine B., J. II. (1993): Making Mass Customization Happen: Strategies for the New Competitive
Realities. Planning Review, 21, 5. 23-24.
7. ProWood Stiftung. (2004): Presseinformation „Holzbearbeitungstechnologien 2015“. Retrieved
March 15, 2005 from http://www.prowood.org/PDF/PDF166.pdf
8. Reiner, G. (2005): Customer-oriented improvement and evaluation of supply chain processes
supported by simulation models. International Journal of Production Economics 96, 381-395.
9. Sigg, A. / Jonas, U. (2003): Der standardisierte Sonderwunsch. möbelfertigung, 2. 131.
10. Yang, B. / Burns, N. (2003): Implications of postponement for the supply chain. International
Journal of Production Research, 41, 9. 2075-2090.
54
Tensile Proof Loading to assure Quality of Finger-Jointed Structural Timber

Georg Jeitler1, Gerhard Schickhofer2
1
holz.bau forschungs gmbh, Inffeldgasse 24, 8010 Graz, georg.jeitler@tugraz.at
2
Insitut für Holzbau und Holztechnologie, Inffeldgasse 24, 8010 Graz, gerhard.schickhofer@tugraz.at
Keywords: proof loading, tensile proof loading, safety factor, grading
ABSTRACT
The paper describes how a quality assuring measure can be applied for finger jointed structural timber by integrating a
tensile proof loading device in the production process. Thereby every produced beam is clamped on both ends with profiled
steel plates and subjected to a defined tensile loading in terms of duration and stress level. With this method, depending on
the set proof level, greatly strength reducing timber features such as the global and local grain deviation, faulty finger
joints, compression failures or reaction-wood are recognized by failure and can be rejected. With the currently common
grading processes such defects are only ascertainable with difficulty and often not economically.
This paper shows the possibility to improve the safety factor of timber within the scope of design purposes by using this
proof loading procedure depending on the proof level and the coefficient of variation of the tested material.
The results of experimental research work on a high number of specimens (series A: 4,886 #, series B: 565 #) show clearly,
that there is no appreciable damage to surviving timber due to tensile proof loading at low load levels. Within a double
and a triple proof loading procedure about 99.96 % of all specimens could sustain higher stresses than at the first time,
indicating not being damaged.
The advantage of the developed procedure is that all specimens are loaded in tension which is rather more sensitive
in respect of failure recognition than bending, because the whole cross section and the whole length is stressed uniformly.
Overall, the assessed tensile proof loading procedure enhances confidence in the wood as a building material.
A timber product with more reliable minimum strength can be made available to the construction industry.)
INTRODUCTION / PROBLEM / MOTIVATION
Timber as a natural growing raw material displays large variations in its mechanical characteristics like
strength and stiffness in comparison to other materials such as e.g. steel. These variations can be
considerable precise with the beam-shaped product structural timber, characterized by lack of
homogenisation over the cross-section through gluing of individual components. A statistical ‘system
effect’ which can be considered for glulam or bi- or trilam is not present for single sections. Although
grading criterions are defined in DIN 4074-1 with the currently common grading processes strength
reducing defects such as the global and local grain deviation, compression failures, reaction wood, pre-
broken timber or damage of tree-top are only with difficulty and often not economically ascertainable.
Rogues in the lowest quantile area of strength cannot be excluded for sure. The grading process within
the production of structural timber is therefore still a challenge.
Even so, performance and minimum production requirements for finger joints of structural timber are
regulated in EN 385, a similar difficulty comes up with the joining. This is because for internal and
external quality control only the bending strength and mode of failure of few randomly taken finger
joint samples are determined in destructive tests. This also results in the fact that structural timber with
features responsible for poor finger joint strength can reach the customers.
STATE OF THE ART
Grading of Structural Timber
In the German speaking area the most common grading method for structural timber is done by visual
inspection according to DIN 4074. The structural timber is mostly graded to class S10 which is
assigned strength grade C24 pursuant to EN 338. The grading criterions most commonly used are
knottiness, cracks, deformations, annual ring width, wane and discolorations. For the product KVH®
(Konstruktionsvollholz) additional stricter requirements like the moisture content (um = 15 ± 3%),
sawing pattern (pith separated), the dimensional stability (± 1 mm for cross-sections ≤ 100 mm, ± 1.5
mm for cross-sections > 100 mm) and visual appearance (seasoning cracks, knottiness, discolorations,
55
warping, surface quality, wane) are to be obtained. The requirements for KVH® differentiate between
applications in visible and non-visible areas.
Stress grading of structural timber by means of bending machines, which determine average Modulus
of Elasticity (MOE) over short lengths, is because of the limited operating range restricted to the
grading of glulam laminations and scaffold boards with a maximum thickness of 75 mm. By use of
X-ray radiation and vibration measurements (eigenfrequency) joists and beams up to 100 mm can be
graded [4]. The preferred cross sections (width up to 140 mm, thickness up to 240 mm) used
for KVH® production go beyond the capabilities of approved grading machines.
Proof Loading
Commonly spoken, ‘proof loading’ as a testing method is defined by specimens which are subjected to
a defined and generally brief mechanical loading. All samples not reaching a set proof level due to
premature material failure can be separated from those with greater strength. Proof loading is a
recognized quality control technique to improve the characteristics of the lower tail of strength
distribution. Numerous scientific research works on the topic and especially in respect of possibly
damaging the material have been published since the late sixties of the last century. Proof of any
possible damage is generally considered to be very difficult to impossible. Strickler et al. (1969) for
example investigated in [7] proof loaded finger joints and concluded that a bending proof load
up to 90 % of the expected ultimate strength did not significantly reduce the strength and by
comparison, a tensile proof load was considered feasible, without qualification.
Woeste et al. (1987) conducted in [8] experiments on 1,200 pieces of lumber with single and reverse
bending loads and detected no damage due to proof loading. Heatwole et al (1991) stated the following
in their literary research on damage [9]: ‘Based on published research it is valid to assume there is no
appreciable damage to surviving lumber due to proof loading in tension or bending at these low load
levels’. Lam et al. (2003) pointed out in [10] that one of the difficulties is the need of rather large
sample sizes for an experimental-based study to develop statistical solutions to quantify the
effectiveness on the use of proof loading in relation to the proof level, the potential damage on the
members and the improvement of performance in the context of reliability based design methods.
Proof loading is therefore not a new development in the timber construction sector; it is rather already
familiar for many years, primarily from North America and Australia and also embedded in manuals
[5] and standards [6]. Whereas in Europe apart from those stress grading machines mentioned in 2.1,
which in principal do a proof loading of the material in bending, the authors know of no approved
industrial proof loading application for structural timber. Test methods working with tensile loads are
still uncommon obviously due to the difficulties of applying the loads.
TENSILE PROOF LOADING / FUNCTION OF THE TESTING DEVICE
The target definition of the research project ‘qm_online’ was the development of a quality control
method to assure high product performance of finger jointed structural timber especially in respect
to strength characteristics. In particular the following aspects should be fulfilled:
• Every produced piece and therefore the whole volume of the material should be tested.
• No dents should remain on the surface and damage is allowed.
• Testing has to be integrated in production and must not reduce production output.
• Information valuable for design purposes - like strength and stiffness - should be available.
The approach to comply with all defined aspects was tensile proof loading at a high production level.
Because loading in bending is difficult to achieve with cross sections typical for structural timber and
has disadvantages when stress reducing failures have to be detected in the compression zone, loading
in tension was selected. Further this implies, in contrast to bending, a constant stress distribution over
the entire timber volume within the free test length.
For development of an appropriate device for industrial applications research was carried out in respect
of determination of the most significant mechanical parameters of the testing device (maximum tensile
56
force, clamping plate geometry and surface structure, capacity of the machine and measuring
technique) and to determine the time dependant strength of the adhesive used for the finger joints [11].
The 8 to 18 m long rods are individually loaded into the transverse conveyor of the proof loading
device at least two hours after the finger jointing process. A curing time of 120 minutes for the PU
adhesive Purbond® HB 530 is considered as sufficient for the application of a proof level up to
10 N/mm² without damage of the finger joint. Tensile tests on not fully cured finger joints were
worked out to for confirmation. The PU adhesive Purbond® HB 530 showed that already after
90 minutes a strength level is achieved which is within the scattering of the end strength determined on
fully cured joints.
Figure 1. System sketch of the tensile proof loading device for industrial application
Within the proof loading process a centering device puts each beam into a defined test position.
As shown in figure 1 the beam ends are then clamped with profiled steel plates over a length
of 400 mm and the corresponding section width. In this way the structural timber is subjected to
a defined stress in terms of duration and load factor. During the stress test the proof load and
associated deformations are continuously recorded, whereby the mean MOE over the full length up to
18 m can be determined. Rejection parameters of the control program can be sudden drops of the
tensile force, too great deflections or when the set proof level is not reached in a certain time or held
constant over the defined period. Only those rods running through the test without fraction or error in
the control program are fed to the following profiling process. Adjustment to the length
of the structural timbers to be tested is provided by the lengthways continuously movable clamping
unit.
CYCLING TENSILE PROOF LOADING / EXPERIENCE
Especially to clarify the risk of eventually damaging the material within a tensile proof loading process
a cyclic stressing of finger jointed structural timber was analyzed with a high number of specimens.
Therefore, within special observation periods timber with various cross sections and lengths were
produced. All rods were mechanically tested (proof loaded) using the device as described in figure 1 in
industrial environment. In principle two different series (A and B) were investigated in respect of type
of loading. Within series A the specimens were tensile stressed to approximately 7 N/mm² and after
a short release to a level of approximately 8 N/mm². In contrast the specimens of series B were
stressed three times to 12.8 N/mm² but had the same level of release as in series A. Within all cycles
the time of effectively loading the material at constant stress was in the range of 1.3 to 2.0 seconds.
The speed of loading was, depending on the actual cross section, in the range of 20 to 30 kN/s.
Within the total time of loading data (time, tensile force, and extension of end grain) was recorded
automatically at a rate of 4 Hz. In table 1 the proof loading programmes are illustrated exemplarily.
57
Table 1. Test series to clarify risk of damage when tensile proof loading
Double stress test (Series A) Triple stress test (Series B)

10
14
Applied tensile stress [N/mm²]
Applied tensile stress [N/mm²]

8
12
10
6
8
4 6
4
2
2
0
0 2 4 6 8 10 12 14 0
Time of loading [sec] 0 2 4 6 8 10 12 14 16 18
Time of loading [sec]
proof level_1 = 7 N/mm², level_2 = 8 N/mm² proof level_1 = level_2 = level_3 = 12.8 N/mm²
Various cross sections: Cross section:
65 mm < width < 125 mm, 63 x 145 mm²
105 mm < height < 285 mm
Tested volume: ~ 1.100 m³ Tested volume: 63 m³
number of tested rods: 4,886 # (39,000#) number of tested rods: 575 # (5,480#)
number of tested finger joints: ~ 30,000 # number of tested finger joints: ~ 3,000 #
The tested volume equals approximately one thousand one hundred (!) cubic meters and 63 m³
of structural timber in series A and series B, respectively. The mean free testing length for both series
was approximately 12.0 m. Transposed onto the referred test piece length, acc. to EN 408 of nine times
the larger cross-sectional dimension, this would roughly equal 39,000 (!) and 5,480 (!) tests with
a free span of approximately 1.6 m und 1.3 m for series A and series B, respectively.
The timber (spruce and pine) for both series was graded acc. DIN 4074 to class S10 by means of visual
inspection and an X-ray scanner. There was no separation of higher class material. The finger joints,
fulfilling the requirements acc. to EN 385, were characterized by a finger length of 20 mm
and a distance between fingers of 5 mm.
The dominant cause of failure within the lower tail of strength distribution was, as shown in table 2,
failure of the wood at 70.9 % and 88,9 % for A and B, respectively. The failure analysis further shows
that the local grain deviation often associated with the surrounding area of knots, knot clusters or a
broken tree-top is thereby the main cause of failure of this material. It has to be noted that many of the
failure causing features could only be detectable with difficulty and apparently not with the applied
grading procedure. This was confirmed by close examination of the broken pieces in respect to the
grading criterions. Hence the associated grading represents the limiting factor for structural timber
from this production.
The following figures show some typical examples of severe timber defects which could be detected
by means of tensile proof loading. Figure 2 shows an extreme local grain deviation caused by a broken
tree top responsible for low tensile strength. It was also observed that not only one defect causes
the failure, rather as illustrated in figure 3, it is a combination - often with reaction wood.
58
Table 2. Results of cyclic stress tests of finger jointed structural timber
Double stress test (Series A) Triple stress test (Series B)

Number of fractures: Number of fractures:
within 1st loading: 37 # within 1st loading: 79 #
within 2nd loading: 28 # within 2nd loading: 3 #
within 2nd loading and below level_1: 2 # within 3rd loading: 1 #
Magnitude of damage expressed Magnitude of damage expressed
as loss of strength = 9 % resp. 12 % as loss of strength = 0 %
32 32
30 30
28 28
26 26
24 24
Absolute frequency [#]
Absolute frequency [#]

22 22
20 20
18 18
16 16
14 14
12 12
10 10
8 8
6 6
4 4
2 2
0 0
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 7 8 9 10 11 12 13
Tensile strength f t [N/mm²] Tensile strength f t [N/mm²]
60% 60%
55% 55% 52.4%
50% 50%
44.0%
45% 45%
Causes of failure [% ]
Causes of failure [%]
40% 40%
35% 35%
29.1% 29.0%
30% 30%
25% 25%
20% 17.4% 20%
15% 15% 11.1%

9.5%
10% 10% 7.5%
5% 5%
0% 0%
local grain failure of reaction-w ood global grain local grain failure of reaction-wood global grain
deviation finger joint deviation deviation finger joint deviation
Figure 2. Low tensile strength due to local grain deviation caused by a broken tree-top
59
Figure 3. Tensile strength about 5.8 N/mm² due to reaction-wood and global grain deviation
Compression damages are due to deformations of the wood fibres resulting from excessive
compression shakes (impacts) along the grain. They may develop in standing trees due to high loads
from storm or snow. They also may result from stresses imposed by lumbering or inadequate handling.
As shown in figure 4 (left) they are very difficult to detect on planed surfaces. Because the distorted
fibres lead to brittle fracture in processed timber already at relatively low stresses, compression
damages can be detected with a tensile proof loading procedure.
Figure 4. Tensile strength about 6.4 N/mm² due to compression damage of the timber
REDUCTION OF THE PARTIAL-SAFETY-COEFFICIENT

Generally, a timber product with a more reliable minimum strength should be made available to the
construction industry by the presented tensile proof loading method as every piece in the lower area of
the strength distribution is rejected. The increased reliability for proof loaded finger jointed structural
timber could also be reflected in a more favorable partial-safety-coefficient. The corresponding
quantification in dependency of the proof level and coefficient of variation of the base material was
part of an investigation in cooperation with G.I. Schuëller (Institute of Engineering Mechanics,
Leopold-Franzens University, Innsbruck, Austria).
60
σproof A ≡ Aproof
fR(x)
Decrease of the partial safety coefficient
fS(x)
depending on the change of
Rproof the proof level and the
Coefficient of variation (COV)
A >> Aproof
S
1,30
partial safety coefficient γ m

mS mR mR,proof
R 1,25
1,20
x COV = 20%
1,15
probability of failure A COV = 25%
before proof loading 1,10 COV = 30%
COV = 35%
1,05
COV = 40%
1,00
probability of failure Aproof
50% 55% 60% 65% 70% 75% 80% 85%
after proof loading 'proof level' in [%] of f t,0,k
Figure 5. procedure for the decrease or the partial safety coefficient
DISCUSSION AND CONCLUSION

The completed cyclic stress tests, as described in this paper, confirm that a low tensile stress not
leading to failure, only minimally affects the strength of structural timber. The evidence that the
material is not significantly damaged is herewith clearly adduced. The number of tested specimens
(4,886 #) or rather 39.000 # with the referred test length of 1.6 m of series A in relation to the number
of faults with slightly reduced strength characteristics (2 #) after the first stressing seems to be
sufficient to confirm that statement. The triple stress test of series B confirm further, that a tensile load
that could be sustained once (not leading to failure) can be sustained in 99,47 % of the cases again and
in 99,29 % of the cases a third time, indicating not being damaged. The results of experimental
research work as presented on that high number of specimens show clearly that there is NO
appreciable damage to surviving timber due to tensile proof loading at low load levels. Or in other
words: The timber is not significantly damaged within tensile proof loading as described in this paper.
The conclusion therefore clearly is that it is better to have tensile proof loaded timber in structural
applications than the risk of ‘rogues’ with poor strength characteristics. Further there should not be any
doubt of stressing timber up to the level of design strength which is specified for grade C24 with ft,0,d =
ft,0,k / γm* kmod = 14 / 1.3 * 1.1 = 11.8 N/mm², assuming an instantaneous load duration.
A further area of application of the test method presented here and implanted on an industrial level
exists for other sawn timber products in the branch. Glulam production is particularly considered here.
It is conceivable to also implement the presented proof loading method in an adapted form for the
online quality assurance of finger jointed single lamellas. Furthermore application of the method
for testing finger jointed flange sections of I-profiles and nail plate binders is considered sound.
REFERENCES
[1] DIN 4074-1:2003. Strength grading of wood - Part 1: Coniferous sawn timber.
[2] EN 385:2001. Finger jointed structural timber - Performance requirements and minimum
production requirements.
[3] EN 338:2003. Structural timber - strength classes.
61
[4] EN 14081-4:2005. Timber structures - Strength graded structural timber with rectangular cross
section - Part 1: General requirements, Part 3: Machine grading; additional requirements
for factory production control, Part 4: Machine Grading - Grading machine settings
for machine controlled systems.
[5] AITC 200-2004. Manufacturing quality control systems manual for structural glued laminated
timber. American Institute of Timber construction, pp. 55-57.
[6] AS 3519-1993. Timber - Machine proof-grading. Australian Standard.
[7] Strickler, M.D., Pellerin, R.F., Talbott, J.W., 1970. Experiments in Proof Loading Structural
End-Jointed Lumber. Forest Products Journal Vol. 20, No.2., pp. 29-35.
[8] Woeste, F.E., Green, D.W., Tarbell, K.A., Marin, L.A., 1987. Proof loading to assure lumber
strength. Wood and Fibre Science, pp. 283-297.
[9] Heatwole, E., Woeste, F.E., Green, D.W., 1991. Allowable bending strength enhancement of 2
by 4 lumber by tension & compression proof loading. Wood and Fibre Science 23(1), pp.1-14.
[10] Lam, F., Abayakoon, S., Svensson, S., Gyamfi, C. 2001. Influence of Proof Loading
on the Reliability of Members. Holz als Roh- und Werkstoff 61, 432-438.
[11] Katzengruber, R., Jeitler, G., Schickhofer, G., 2005. Nondestructive quality assurance of finger
jointed structural timber based on a tensile proof loading procedure. 14th International
symposium on non-destructive testing of wood, pp. 225-234.
[12] EN 408:1995. Timber structures - Structural timber and glued laminated timber –
Determination of some physical and mechanical properties.
[13] Weibull,W., 1939. A statistical strength theory of the strength of materials.
Ingeniörsveteskapsakademiens Handlingar Nr. 151, Generalstabens Litografiska Anstalts
Förlag, Stockholm.
[14] prEN 15497:2006. Finger jointed structural timber - Performance requirements and minimum
production requirements.
62
Effects of heat treatment of small clearwood samples on equilibrium moisture

content and deformation
Björn Källander1, Pierre Landel2
1
Källander Development, Box 111, SE-78041 Gagnef, Sweden, bjorn.kallander@kallander.eu
2
ENSAM, Centre d'Enseignement et de Recherche de Cluny, 71250 CLUNY, France
(Presently Fristad Bygg AB, Trumslagar Blomgrens väg 5, SE- 51333 Fristad, Sweden, pierre.landel@fristadbygg.se)
Keywords: clear wood, deformation, drying, equilibrium moisture content, heat treatment
ABSTRACT
High Temperature drying has by numerous studies been shown to affect strength properties as well as moisture properties
of the wood treated. The present study aims at determining the effects of different heat treatments of small clearwood
samples on material properties. Samples were treated in dry and moist air, saturated steam and water. Temperatures
varied from 20 °C to 150 °C. Treatment periods varied from 6 h to 96 h. The results of the study indicate that the effects of
treatment are dependent not only on the method of treatment and time, but also on the pre treatment of the samples such as
ingoing moisture content.
INTRODUCTION
Strength properties of Norway spruce planks after industrial drying in elevated temperatures have been
studied in a series of projects at SP Technical Research Institute of Sweden (Betzold, 1999), (Bengts-
son and Källander, 2001), (Imbaud, 2001), (Källander et al. 2001), and (Dahlberg, 2002). Laboratory
studies on clearwood samples were then initiated to investigate degrade mechanisms at elevated
temperatures. As expected the laboratory studies on clearwood samples showed different results as
compared to the full size planks (Källander and Bengtsson, 2003) and (Landel, 2004). It is evident that
the result of strength tests on small clearwood samples will rely on different material properties than
corresponding tests on planks, were the stresses around local defects will determine the result.
However, it can also be expected that the kiln drying climate will lead to different internal wood
climate in a plank than in a small sample, and thus influence the wood material properties differently.
The combined effect of the factors above is shown in Figure 1 where the results from tensile tests at SP
of full size planks after High Temperature (HT) drying are compared with tests on small clearwood
samples as reported by Stamm, 1953. The "loss of tensile strength" is 100 - 1000 times faster in the
planks than in the clearwood samples.
Figure 1.Heating time versus temperature to reach various degrees of wood degradation. Solid line corresponds to
10 % loss of modulus of rupture on dry clearwood samples. The square marks the time and temperature range
resulting in similar strength loss in planks in the SP studies (Källander and Bengtsson, 2003).
63
Chemical processes causing degrade of wood material
Fengel and Wegener, 1989 describes three different chemical processes that will lead to degrade of
wood at elevated temperatures: 1) hydrolysis, 2) dehydration, and 3) oxidation.
1) Hydrolysis is generally accepted as the main cause of degrade during HT treatment of wood. During
hydrolysis the polysaccharide is broken as water is bound to the molecule. It is obvious that hydrolysis
requires water, and that the total mass of the polymer (parts) will increase as the water molecule is
added. Hydrolysis is accelerated in acidic environment (Fengel and Wegener, 1989), and hence
accelerated by organic acids like acetic acid and formic acid formed during the hydrolysis. The organic
acids will act as catalysts for the hydrolysis, but since acetic acid and formic acid are highly soluble in
water, part of the acids can evaporate together with the moisture.
The main part of the resulting degrade compounds are shorter saccharides. These saccharides are not
volatile and will not evaporate from a solid piece of wood during drying. However, if the hydrolysis is
carried out on sawdust in an acidic solution, a large portion of the reaction products can be expected to
be extracted with reduced total mass as a result. During a drying process of sawn planks water will
always be present, either as bound or free water or as water vapour or steam. On the other hand, small
samples quickly dries out in high temperatures, resulting in a dry treatment process. It is clear that the
size of the sample treated and the treatment agent will influence the result.
2) Dehydration implies that the polysaccharide is broken as water molecules are released (Fengel and
Wegener, 1989). Dehydration does not require presence of water since the water is produced in the
process. When dehydration takes place at temperatures above 100 °C, the water molecules produced
will evaporate from the wood in the form of water vapour or steam, leading to a reduction of the mass
of the sample treated.
3) Oxidation reactions occur in parallel with hydrolytic degradation during for instance acidic pulping
(Fengel and Wegener, 1989). Oxidation leads by definition to an increase in molecule weight. The
sample size and treatment atmosphere can be expected to have great influence. If oxidation has a
significant effect on wood properties after drying is not known to the authors.
MATERIAL AND METHODS
A total of 1108 clear wood samples from Norway Spruce and Douglas-fir for heat treatment were
prepared from the sapwood of taper sawn 60 mm Norway spruce planks (576 samples) and a 0,55 m
diameter Douglas fir log (532 samples). Sample dimension 40 mm x 40 mm x 40 mm. All samples
were conditioned in standard climate 20 °C / 65 % RH to approximately 12 % MC and density of the
samples determined.
The samples were divided into sets of 10 samples with corresponding average density. Two sets from
each wood species were used for each thermal treatment, or 20 samples per species. The samples were
exposed to elevated temperatures in moist air, in steam, and in hot water for times ranging from 6 h to
96 h. Temperatures ranged from 95 °C to 125 °C. Treatments were done with samples impregnated
with water (MC at start 60 % - 150 %) as well as conditioned samples (MC at start 12 %).
Four varieties of treatments were used: 1) moist air treatment, 2) saturated steam treatment,
3) hot water treatment with free samples and samples wrapped in aluminium foil to reduce extraction
of reaction products from the samples, and 4) hot water treatment followed by moist air treatment.
The moist air treatment was done to simulate the wood climate during the diffusion phase of the drying
process and the wood surface will be exposed to water vapour. The treatments were done in closed
chambers were steam from boiling water was continuously fed into the chamber. Saturated steam
treatments were done to simulate the capillary phase. The treatments were done in a pressurized vessel
with samples placed above boiling water and excess steam evacuated by a pressure valve. Hot water
treatments were made to simulate the climate inside the evaporation front of the wood. The treatments
were made in 20 litre stainless steel cauldrons.
64
Kiln climate and internal wood temperature were recorded during treatments. After treatments samples
were inspected, weighed and reconditioned in 20 °C / 65 % RH before shear testing and energy of
rupture testing. The wood samples were stored more than 2 years in 20 °C / 65 % RH before the final
MC was determined by dry weight measurement.
RESULTS ON EMC
The results show a drop in average EMC of the samples during treatment from 11,8 % to 10,6 % for
Douglas and from 12,5 % to 11,2 % for Spruce. The drop in EMC seems to happen early in the
treatment process. Figure 2 shows EMC after treatment in humid air at 105 °C for different periods.
14
12
Final average MC (%)
10
8
6
Spruce
4
Douglas
2
0
0 24 48 72 96
Treatment time (h)
Figure 2. Final average MC after treatment in 105 °C humid air.
The results show a faster drop in EMC during treatment in humid air as compared to saturated steam,
see Figure 3. The total EMC drop in humid air seems not to be influenced by the treatment temperature
at 105 °C and above, the final MC is approximately 9 % for Douglas samples and 10 % for spruce.
Treatment in saturated steam shows a slower effect on the EMC, and a increase with temperature.
Water treatment in temperatures up to 100 °C show no significant effect on average EMC in this study.
14
12
Final average MC (%)
10
8 Spruce steam
6 Douglas steam
4 Spruce humid air
2 Douglas humid air
0
0 25 50 75 100 125 150
Treatment temperature (°C)
Figure 3. Final average MC after treatment in different temperatures for 24 hours.
65
The temperature measurements during treatment show that the internal wood temperature rapidly
increases to the level of the treatment climate, indicating that all or most of the initial moisture in the
samples has been removed. Figure 4 shows temperature readings from treatment of water impregnated
samples in 125 °C.
140
120
Temperature (°C)
100
80
60
Dry bulb (°C)
40 Spruce (°C)
Douglas (°C)
Wet bulb (°C)
20
0
0 1 2 3 4 5 6 7
Process time (h)
Figure 4. Temperature development during treatment in 125 °C humid air. Internal wood
temperature shows that samples have dried out after approximately 1h 30 min.
Results on deformations and internal checking
The initially cubical samples were generally more or less deformed into rhombic shape during
treatment due to the anisotropy of the wood. Several of the samples that first were boiled and then
immediately humid air treated showed radial checking and deformations as if there was collapse in the
interior parts. These samples also showed internal radial checking when the samples later were split.
Figure 5 shows radial checking in the earlywood of Douglas samples boiled for 12 h followed by
drying 12 h in 125 °C.
Figure 5. Radial checking in earlywood of Douglas samples boiled 12 h and dried 12 h in 125 °C.
66
DISCUSSION
The study has resulted in a significant reduction of EMC after heat treatment. This confirms that the
methods of heat treatment used have influenced the chemical composition of the wood.
The results seem to show a rapid reduction in EMC of the samples during the early stages of the
temperature treatment, after which the rate of EMC reduction is greatly reduced or halted. At the same
time the temperature readings show that the internal wood temperature rapidly increases above
temperatures where free water can be present, and within 2 hours reaches the surrounding temperature.
As the wood temperature is at the same level as the surrounding atmosphere, little or no evaporation
can occur. The wood is dry. As the moisture of the wood is removed, the rate of EMC reduction is
greatly reduced.
The lower reduction rate in EMC during treatment in saturated steam is somewhat unexpected. Earlier
studies have indicated a higher effect on wood properties from treatment in steam as compared to air
(MacLean, 1953 and Stamm, 1956). The results in this study can possibly be linked to the water
impregnation of the samples prior to treatment, by which the air treated wood has been supplied with a
reservoir of water for quick hydrolysis. A different pattern could have shown if the treatment periods
with steam had been longer. Then it is possible that the EMC would have continued to be further
reduced. The treatment periods and temperatures in this study have been limited to drying schedules
used in European wood industry.
This study has not shown any stronger effects on the EMC of the wood material after the initial
moisture of the wood has been removed. This indicates that the wood has not been seriously affected
by oxidation or dehydration.
The results indicates that the difference between the effects of a specific treatment climate on planks as
compared to small clearwood samples is linked to the water content of the wood. The small samples
quickly dries out at elevated temperatures, after which hydrolysis is decelerated or stopped. This would
in turn indicate that tests on small clearwood samples need to be performed in such a way that the
internal wood climate corresponds to the internal climate of planks during drying in order to result in
corresponding results.
CONCLUSIONS
- The study has lead to significant reduction sin EMC of small clearwood samples after heat treatment.
- The hydrolysis during thermal treatment seems to be dependent on the water in the wood rather than
the humidity of the atmosphere.
- As the small samples dry out, hydrolysis is halted.
- The procedure during test on small clearwood samples need to be adapted to the internal wood
climate in order to provide results that correspond to industrial drying of planks.
67
REFERENCES
1. Bengtsson, C. and B. Källander (2001). Högtemperaturtorkat virke - varför minskar
hållfastheten? Borås, Swedish National Testing and Research Institute.
2. Betzold, D. (1999). Maschinelle Festigkeitssortierung Einfluss der Hochtemperaturtrocknung
auf die elastomechanischen Eigenschaften des Schnittholzes. Fachbereich Holztechnik.
Eberswalde, Fachhochschule Eberswalde: 116.
3. Dahlberg, J. (2002). Hållfasthetsförändringar i högtemperaturtorkat granvirke. Skellefteå,
Luleå Technical University: 115.
4. Fengel, D. and G. Wegener (1983). Wood: chemistry, ultrastructures, reactions. Berlin, Walter
de Gruyter.
5. Imbaud, O. (2001). Bois séché à hautes températures: influence sur les propriétés méchaniques.
Lausanne, Ecole Polytechnique Federale de Lausanne.
6. Källander, B., C. Bengtsson, et al. (2001). Influence of drying in temperatures between 70 °C
and 125 °C on selected wood properties of Norway spruce. 7th international IUFRO wood
drying conference. K. Hayashi. Tsukuba: 306-311.
7. Källander, B., C. Bengtsson, et al. (2001). Reduction of strength but not stiffness of Norway
spruce planks dried at 125 °C as compared to 70 °C. Proceedings. Cost Action E15 3rd
Workshop on softwood drying to specific end-uses. A. Ranta-Maunus. Helsinki, VTT Building
and transport.
8. Källander, B. and Bengtsson, C. (2003) High Temperature drying of Norway Spruce: effects of
elevated temperature on wood properties. Proceedings. Cost Action E15. Final conference
Advances on the drying of wood. Athens, October 2nd-3rd, 2003.
9. Landel, Pierre. (2004) Influence of high temperature drying on degrade of clear wood.
Ingénieur de l´Ecole Nationale Supérieure d´Arts et Métiers (ENSAM) Cluny.
10. MacLean, J. D. (1953). Effect of steaming on the strength of wood. Proceedings American
Wood- Preservers' Association 49, American Wood- Preservers' Association: 88-112.
11. Stamm, A. J. (1956). "Thermal degradation of wood and cellulose." Industrial and Engineering
Chemistry 48(3): 413417.
12. Stamm, A. J. (1964). Wood and Cellulose Science. New York, The Ronald press company.
68
Modeling the properties of strength graded timber material
Jochen Köhler, Markus K. Sandomeer
ETHZ, Swiss Federal Institute of Technology Zurich, CH-8093 Zurich, Switzerland

jochen.koehler@ibk.baug.ethz.ch sandomeer@ibk.baug.ethz.ch
Keywords: Timber strength grading, timber material properties, structural reliability
ABSTRACT
The subordinate goal of the COST Action E53 is to increase the market share of wood based products in the construction-,
furniture- and other related sectors by means of advanced methods for quality control. A prerequisite for the development
of strategies for quality control is that the requirements of the end users are considered carefully. In the present
presentation user requirements for wood and wood products are first discussed in general and then the specific
requirements for timber as a structural material are discussed.
BACKGROUND
According to the presentation of Gerald Reiner at the COST E53 meeting in Sopron (2006) [3] there
are two different strategies of quality control enhancement – progressive and reactive (Table 1).
Table 1: Strategies of quality control enhancement – progressive and reactive [3].
STRATEGIES FOR ENHANCING QUALITY CONTROL

PROGRESSIVE REACTIVE
More possibilities by better Better quality of existing products.

control/knowledge of properties.
Increasing market share by means of Increasing market share by means of
exploring new products and markets. increased customer contentedness.
It is important that the focus of this COST Action is not only directed on reactive aspects of quality
control enhancement. The quality control of the whole timber production chain has to be considered
closely starting from harvesting logs and ending at e.g. the development of new solid or engineered
timber products.
After sawmilling the timber is graded to the product ‘graded material’ which is utilized to produce
timber structures. For the end-product ‘timber structures’ there are different user requirements as
indicated in Figure 1. Structures which fulfil these requirements are competitive and will stimulate the
demand for forest products in the construction sector. The efficient quality control with relation to the
requirements as well as to the end product (timber structure) is also illustrated in the graphic and
performs a close and interacting relationship with triangular shape.
69
timber structure
designer execution / maintenance
framer requirements quality control production chain
end user grading rules / building codes
Figure 1: Interrelationship between the end product (timber structure), the product requirements and quality control.
The end users of the structure will state their preferences in regard to the building material to be used
in the contracting phase, more or less detailed. Of course they will support their decision on the
expectation that their preferences (requirements) will be fulfilled by the chosen material. The architect
or engineer who performs the design and detailing will have some freedom to choose his preferred
material or will influence the stated preferences of the end user with adequate consulting. The timber
framer will not decide which building material will be used, but he will look for proper timber material
at the market.
Table 2: Requirements for structural timber according to different pressure groups.
designer framer end user
explicitly: - correct target - reliability

sufficient strength, moisture content - serviceability
stiffness and density - appropriate form - durability
- sawing accuracy - good appearance
implicitly: - sufficient durability - value for money
cp. framer next - good appearance - …
column - price value
Table 2 shows the particular requirements of the affected pressure groups. For designers mechanical
properties like strength, stiffness and density of the material are of special importance for the design of
the load bearing constructions. Rather physical properties related to moisture content, form and
appearance are assessed with regard to the general market situation. However, the latter properties are
also the crucial quality requirements for practitioners / timber framers.
To ensure the mentioned requirements to be fulfilled, the timber research and code authorities provide
the timber building sector rules and guidelines to produce the end product “timber structures”. This
includes also codified grading rules to insure properties as sufficient strength, stiffness and density.
Therefore grading of structural timber becomes an important core element in the field of quality
control.
Traditionally, codes are very much based on experience. However, recently codes are to a large extend
also drawn upon a rational support, where reliability and serviceability are considered explicitly.
Modern building materials are defined on basis of required properties which are assessed
probabilistically. The probabilistic modeling of timber material properties can be performed by
directly analyzing graded timber material. This includes the consideration of all possible combinations
of different grading schemes and raw material, since these two conditions are crucial for the statistical
characteristics of graded sub-populations. It is therefore of utmost importance that existing grading
provisions allow for the formulation of consistent probabilistic models of the grading procedure itself –
not only for properties of the material which is already graded.
70
GRADING STRUCTURAL TIMBER

Reliability analysis of structures for the purpose of code calibration in general or for the reliability
verification of specific structures requires that the relevant failure modes are represented in terms of
limit state functions. The limit state functions define the realizations of resistance parameters, i.e. the
material properties and the load variables resulting in structural failure. In reliability analysis of timber
structures the probabilistic modeling of the material properties is an issue of special interest due to the
particular way this material is “produced”.
Considering timber as a natural very inhomogeneous building material the material properties are a
product of e.g. the specific wood species and the geographical location where the wood has been
grown. Given species and geographical location the material properties depend on factors such as the
age, the diameter of the timber logs and the number of knots together with the moisture contents and
the duration of loading. In comparison to other building materials such as steel and concrete, the
properties of timber materials are not designed or produced by means of some recipe but may be
ensured to fulfil given requirements only by quality control procedures – hereafter referred to as
grading. Quality control and selection schemes are implemented in the production line, typically
already at sawmills where the construction timber is produced from the timber logs. Various schemes
for grading have been developed using different principles, however, the basic idea behind them all is
that the material properties of interest such as, e.g. the ultimate compression stress, are assessed
indirectly by means of other properties such as e.g. the density or the modulus of elasticity (see for
example Madsen [4], Walker et al. [5] and Green and Kretschmann [6]).
The Strength Class System
As a result of grading, timber is provided to the market as a graded material. The grades imply that the
material properties lie within desirable and predictable limits. However, the material properties
of timber grades have to be considered as random variables and the properties of timber grades
are characterised (and communicated) through specific fractile values of the assumed probability
distribution functions of the material properties of interest.
In general, structural timber is assigned to a specific strength class. Several strength class systems exist
on an international scale, e.g. in Europe it is the EN 338 which constitutes the classification of timber
based on the prescription of characteristic values for the material properties; i.e. for every timber
strength class a characteristic value for every relevant material property is given. Timber that is
assigned to a certain strength class is also referred to as a timber grade. In EN 338 the characteristic
values for the strength properties and the density are defined as the 5th-percentile values of the
underlying distribution functions. The modulus of elasticity (MOE) and the shear modulus are
specified by mean values.
Grading Strategies
In general two different strategies of timber grading exist: visual grading and machine grading. Visual
grading is based on visual inspection of timber structural elements. Visible defects, such as knots,
fissures and cross grain are assessed and according to the appearance of timber structural elements in
regard to these defects they are sorted to a certain grade. Visual grading in general is standardized; the
first formal visual grading rules, the USA ASTM Standard D245 were published in 1927 (Madsen [4]).
Since the 1930s formalized rules for visual grading were introduced in the European countries (Glos in
[7]). Visual grading has proven as an efficient tool to reduce the variability of timber material
properties, however, the grading effect strongly depends on the person who is performing the visual
grading. The statistical characteristics of the material properties of visual graded timber are therefore
difficult to assess explicitly based on information about the applied visual grading rules.
The above mentioned disadvantages of visual strength grading may be overcome by machine grading,
where a more formal assessment of the grading process can be performed. In contrast to visual
grading, machine grading is in general based on indicative characteristics of a timber structural
element which can be measured non-destructively by some device. The indicative characteristics have
to be related to the basic material properties of interest. Typical indicative characteristics are:
71
• Directly related to the MOE: flat wise bending stiffness, ultrasonic pulse measurement,
frequency response measurement.
• Directly related to the density: measurements of weight and dimensions, X-ray detection.
• Directly related to visible defects: microwave response, optical detection and subsequent image
processing.
A good overview about the different measuring schemes can be found in Thelandersson and Larsen
[8].
Control of Grading Machine Settings
Commonly, grading machines are operating either machine controlled or output controlled. The output
controlled grading system was developed in North America. Control is based on frequent destructive
strength testing or proof loading of control samples of the machine graded timber. This system is
relatively costly but it permits a modification of the machine settings in order to optimise the yield, i.e.
the predictability of the properties of the graded timber material. This method requires large quantities
of timber of similar dimension and origin, so that it can be assumed that the characteristics of the
timber are stationary. These conditions rarely exist in Europe, where a variety of sizes, species and
grades in small quantities are typical. For these conditions the machine controlled systems are
developed. Machine control means that the settings are derived within a substantial assessment
procedure prior to the operation phase of the machine. The settings are optimised to a representative
un-graded timber population which might be typical for the daily use of the grading machine.
In general these assessments are done for entire geographical regions, e.g. assessments for the gross
supply in France or Scandinavia suggest common settings for certain grading machines used in these
countries or regions.
Modeling the properties of strength graded timber material
As already discussed in the first part of this chapter it is the main requirement for the use of timber in
load bearing structures that their strength, stiffness and density related material properties can be
modeled with sufficient accuracy. This includes a sound representation of all uncertainties involved
into these material models, i.e. uncertainties due to the natural variability of the natural grown timber,
but also uncertainties due to limited information in form of data, and uncertainties due to mechanical
and physical simplifications in the model. As discussed in several publications before [10], [11], [12] ,
a large part of the uncertainty can be controlled by proper schemes for quality control and grading.
Existing schemes for timber grading, however, are not able to represent these uncertainties explicitly.
These traditional schemes mainly focus on the qualification of timber in certain sub-populations,
whereas the sub-populations are characterized in terms of minimum values of predefined characteristic
values for some material properties of interest. The information which can be gained out of these
procedures is not sufficient to derive efficient probabilistic models for the material properties of graded
timber. As mentioned before, such an efficient probabilistic model for the strength, stiffness and
density related material properties is the prerequisite for the safe and efficient use of building materials
in load bearing constructions.
It can be seen as a major challenge for the timber producing industry (and the associated research
community) to develop grading procedures that allow for an explicit utilization of information during
the grading process and use this information for the probabilistic modeling for timber material
properties. The method described in the following can be seen as a first step into that direction.
Possibilities for its implementation in practice should be discussed in the near future. The COST
Action E53 might be a proper platform for these discussions.
72
PROBABILISTIC METHOD
In the following the very basic principle idea behind the probabilistic modeling of graded timber
material properties is outlined. The kernel of the method is a linear regression model of the following
form:
yi = β1 + β 2 xi 2 + ε i
(1)
That is, simultaneous observations of the timber material property of interest xi and the grading
machines indicating property yi are related to each other by Equation (1), where β1 and β 2 are the
regression parameters and ε is the error term.
The regression model can be used to predict the probability density function of graded timber material
properties conditional on the settings of the grading machine; given by two limiting values of the
indicative properties. The regression model is formulated that kind that new information can enter into
the model continuously, i.e. information gained during the grading process can be used to improve
the model and to refine its predictions.
Figure 2: Alternative approach for the control of grading machine settings, based on the regression analysis.
Based on the regression analysis and the probability density function of the indicating property, first an
conditional probability function and out of this an predictive distribution function of the property of interest
can be assessed.
CONCLUSION
The probabilistic modeling of timber material properties has been considered with special emphasis to
the strength grading procedure of structural timber. The suggested probabilistic approach not only
forms a very strong tool for the statistical quantification of the material characteristics of timber but
furthermore provides a consistent basis for quantifying the efficiency of different quality control and
grading procedures.
It is of utmost importance that the statistical characteristics of timber material properties are assessed
and treated in consistency with the implemented quality control and grading procedures. Only then a
consistent basis may be established for the quantification of the reliability of timber structures - the
basis for codification of design and assessment.
73
REFERENCES
1. Joint Committee of Structural Safety (JCSS, 2001). Probabilistic Model Code, Internet Publication:
www.jcss.ethz.ch.
2. COST Action E24, Reliability of timber structures. Several meetings and publications, internet publication:
http://www.km.fgg.uni-lj.si/coste24/coste24.htm , 2005.
3. COST Action E53, Quality control for wood and wood products. International workshop on state of the art in
quality control. University of West Hungary, Sopron. November 8-10, 2006. Internet publication:
www.coste53.net .
4. Madsen B. (1992). Structural Behaviour of Timber. Timber Engineering Ltd., Vancouver, Canada.
5. Walker J. C. F. (1993). Primary Wood Processing. Chapman & Hall.
6. Green, D.W. and Kretschmann, D.E. (1997). “Properties and Grading of Southern Pine Timbers”, Forest Products
Journal, Vol.47, No.9.
7. Blass H.J. (editor) (1995) STEP 1: Timber Engineering, Basis of Design, Material Properties, Structural
Components and Joints. Centrum Hout, The Netherlands.
8. Thelandersson S. and Larsen H. J. (2003). Timber Engineering. John Wiley & Sons Ltd. Chichester, UK.
9. Köhler, J. and Steiger R. (2006) A discussion on the control of grading machine settings – current approach,
potential and outlook. Proceedings, 39th Meeting, International Council for Research and Innovation in Building
and Construction, Working Commission W18 – Timber Structures, CIB-W18, Paper No. 39-5-1, Florence, Italy,
August 28-31, 2006.
10. Köhler J., Faber M.H. (2003) A Probabilistic Approach to Cost Optimal Timber Grading.
Proceedings of the 36th Meeting, International Council for Research and Innovation in Building and Construction,
Working Commission W18 – Timber Structures, CIB-W18, Paper No. 36-5-2, Colorado, USA, 2003.
11. Faber M.H., Köhler J. and Sorensen, J.D. (2004) Probabilistic Modeling of Graded Timber Material Properties
Journal of Structural Safety, Volume 26, Issue 3, Pages 295-309, July 2004.
12. Sandomeer M. K., Köhler J., Linsenmann P. (2007). The efficient control of grading machine settings.
Proceedings of the 40th Meeting, International Council for Research and Innovation in Building and Construction,
Working Commission W18 – Timber Structures, CIB-W18, Paper No. 40-5-2, Bled, Slovenia, 2007.
13. EN 14081 parts 1-4: Timber Structures – Strength Graded Timber with rectangular Cross Section. Comité
Européen de Normalisation, Brussels, Belgium, 2005.
14. EN 338: Structural Timber – Strength Classes. Comité Européen de Normalisation, Brussels, Belgium, 2003.
74
Wood and Wood Products Quality Evaluation in Slovakia

Rastislav Lagaňa1, Stanislav Kurjatko1, Alena Rohanová1, Ján Dubovský1, Anna Šuriková1
1
Technical University in Zvolen, T. G. Masaryka 24, 96053 Zvolen, Slovakia, lagana@vsld.tuzvo.sk
Keywords: grading standards, roundwood, structural timber, timber, wood quality
ABSTRACT
For optimal utilization of wood raw material, the wood and wood products quality evaluation is one of the key points
in wood processing. Authors pointed on recent state of wood quality evaluation in Slovakia based on direct or indirect
methods for quality evaluation in relation to quality of final products, such as structural timber etc. Advantages as well as
disadvantages of recent wood quality evaluation state in Slovakia are addressed.
INTRODUCTION
Compatibility of Slovak and European standards and usability of individual standards in recent
Slovakian wood products industry are the key problems for the best utilizing of wood raw material.
The aim of the article is to address standard-setting state of acceptance and implementation
of European standards in Slovakia. Authors focused mainly on quality grading of roundwood, timber
and timber for building constructions.
Quality grading of wood

Quality grading of roundwood, timber and timber for structural utilization is described in specific
national standards STN as well as EN standards, which were accepted in Slovakia.
Quality grading of roundwood

Recently, we have two valid STN standards for quality grading of roundwood, STN 48 0055
for softwood and STN 48 0056 for hardwood species. Both standards are based on visual appearance
grading of roundwood. Beside visual determined characteristics, the standards define minimal
dimensional characteristics (length and diameter) for each individual grading class. Both softwood
and hardwood species are assorted into six main quality classes: A1, B1, C1, C2, C3 a D1.
Roundwood of the highest quality A1 is used as veneer logs for sliced veneer or as a special raw wood
for musical instruments. High quality class B1 is used for making peeled veneers, sporting goods
or wooden barrels.
Average quality roundwood is categorized in class C1. This is used as saw logs (C1.1, C1.2, C1.3),
logs used for constructions (C1.4) and logs for aggregate processing (C 1.5). Quality class C2 is used
in mining industry as pit wood (C2.1) or thin poles (C2.2).
Pulpwood for chemical or mechanical processing and chips for particle boards are made from logs
of C3 class.
Wood logs not sorted in previous classes belong to quality class D. Firewood assigns into
this category.
Volume of individual assortment classes produced in 2005 in Slovakia is in Figure 1. As one can see
from the graph, there are two majority volume classes, C1 saw logs and C3 pulpwood for chemical
and mechanical pulp, respectively. Small volume in class A1 and B1 is caused by selling of good
quality wood resources as stumpage as well as raw stems. Moreover, there is discussion about statistic
reliability of Green report 2006.
Sorting criteria for appearance evaluation are based on determination of size and quantity of individual
logs characteristics such as discoloration, fungal and insect attack, spiral grains, type and size of knots,
checks and splits, shakes, taper and sweep of logs, eccentricity, reaction wood, double pith,
false heartwood, etc.
75
No roundwood
No roundwood
No roundwood
Mm , in 2005
C1.1, 1.2, 1.4, 1.5

3
C3.1, 3.2
3
C1.3
C2.1
C2.2
1
A1
B1
D1
0
Veneer logs II. class
Veneer logs I. class
Thin poles
Pit wood
Fuel wood
Stumpage
Forest chips
Saw logs
Raw stems
Pulpwood
Poles
Figure 1. Volume of individual assortment classes produced in 2005 (Green report 2006).
During the last decade, several European standards for grading roundwood have been accepted.
According to Slovak classification, Standards STN EN 1316-1, 2, 3 and STN P ENV 1927-1, 2, 3
are assigned into group called “Qualitative assortment of roundwood without specified utilization
purpose”. The newest edition of the standards STN 40 0055 and STN 0056 from 2004 is assigned
into group called “Qualitative assortment of roundwood with specified utilization purpose”.
Despite the progress in Slovak normalization system, most of users are using old STN standards
from 1984. The Slovak law 264/1999 allows using any standards in business relation including expired
standards or national standards of other countries. In reality, only 5 % of users work with new
STN standards. Implemented European standards STN EN 1316 and STN P ENV 1927 are barely used
(Gejdoš et al. 2006).
Grading of sawn timber

Validity of different Slovak and European standards related to grading of sawn timber is intersecting
recently. Both types of standards are based on similar appearance grading principles. Wood species
of timber distinguishes usability of a European or Slovak standard due to continuous progression
in handling of European standards. Slovak standards are used when EU standards does not cover
grading wood specie. After implementation of any EU standard, a related Slovak standard
is automatically expired. Sawn timber is usually not subject to mechanical or technological tests and
is evaluated according to appearance. Required measured dimensions and appearance of timber
characteristics are compared with prescriptive conditions of given sawn timber quality class according
to a specific grading standard. Harmonization of measured matter with prescriptive conditions means
determination of corresponding sawn timber class.
Objective grading respects needs of final utilization as well as differently defined requirements
of specific quality timber classes. According to utilization, the grading standards can be divided
into the following groups (Šúriková 2002 and 2003):
General grading of commercial timber (without specified of a utilization purpose)
STN 49 1011: Unprocessed timber. Softwood timber. Technical requirements.
STN 49 1012: Broad leaved sawn timber. Technical requirements.
STN EN 975-1+A1: Sawn timber. Appearance grading of hardwoods. Part 1: Oak and beech.
STN EN 975- 2: Sawn timber. Appearance grading of hardwoods. Part 2: Poplars.
STN EN 1611-1: Sawn timber. Appearance grading of softwoods. Part 1: European spruces, firs, pines
and douglas firs.
76
Grading of structural timber

STN 49 1531: Structural timber. Part 1: Visual strength grading.
STN EN 338: Structural timber. Strength classes.
STN EN 384: Structural timber. Determination of characteristic values of mechanical properties
and density.
STN EN 518: Structural timber. Grading. Requirement for visual strength grading standards
STN EN 519: Structural timber. Grading. Requirement for machine strength grading standards
and grading machine
STN EN: 1912: Structural timber. Strength classes. Assignment of visual grades and species.
Timber used in pallets and packaging
STN EN 12246: Quality classification of timber used in pallets and packaging.
STN EN 12248: Sawn timber used in industrial packaging. Permitted deviations and preferential sizes.
STN EN 12249: Sawn timber used in pallets. Permitted deviations and guidelines for dimensions.
Exported timber is divided into the following groups: a) pilopad, b) export structure, c) sawn
and partially sawn timber, d) Imballo a Charpente.
The last two timber export groups are graded according to export provisions of commercial timber
given by STN 49 1113: Unprocessed timber. Export softwood timber. Technical requirements. Criteria
from oversees standards.
Structural timber grading

The most significant regulation dealing with quality of wood for building structures in Slovak republic
are the following: STN 49 1531, STN EN 338 and STN P ENV 1995-1-1. The standards
are intertwining together and have position on national and/or European level. The basic is the
standard STN EN 338, which determine strength classes for individual wood species and their
characteristic values (poplar and softwood – classes C14 to C50, hardwood – classes D30 to D70).
In the area of wood quality for building constructions, the standard aims to connect together
the following standards:
1. STN 49 1531 (national) – specify four quality classes of timber (for softwood S0, SI a SII,
and hardwood SD).
2. STN P ENV 1995-1-1 – specify design of wooden construction elements in terms of carrying
capacity, serviceability, and durability.
The following methods are used for the determination of timber quality:
- Visual method – timber quality classes are determined on the basis of the defects occurrence.
This requires a good knowledge of wood structure. Density of wood is not considered.
The standard STN 49 1531 determines limiting values of the permitted timber defects
for individual classes.
- Mechanical method – utilizes elasticity characteristics of wood in grading. The easiest
and the most accessible method is based on the principle of bending. The standards
STN EN 408, STN EN 384 and STN EN 338 include methods as well as utilization of the
results.
From perspectives of timber quality evaluation, knots have dominant position. Natural occurrence
of knots is visible mainly in sawn and edged timber. It has been recognized that effect of knots
on timber strength depends mainly on proportional share of the cross section. Other characteristics
of visual timber grading are spiral grains, size of checks and round face, annual ring width, biological
properties and so on.
Size, quantity and position of knots together with other parameters such as density, moisture content,
strength, elasticity are the substantial characteristics for mechanical timber grading according
to strength.
77
Visual evaluation of knots

The standard STN 49 1531 determines for individual classes the limiting values of the classification
parameters which influence the strength and elasticity characteristics. From the classification point
of view, knots have the greatest influence. There are two methods for judging the knots:
1. according to proportional knot dimensions (ratio of the sum of knot sizes across the surface
and the edge to the double width of the sawn timber)
S
2. according to cross-sectional proportion of the knots ν = knots ( ratio of cross-sectional surfaces
S total
of all knots S knots to cross-section of the total board surface S total ).
Figure 2 describes measuring of knots and calculation of proportional size of knots for timber grading.
Figure 2. Measuring and calculation of proportional knot size of sawn timber. a) measurement method, b) one knot,
Side knot
Edge knot
c) group of knots.
The second method determines a total cross-sectional proportion of knots as a ratio of cross-sectional
surfaces of all knots on the considered cross-section to a total cross-sectional surface of the element. It
also considers two marginal zones (h/4) creating ¼ of the side surfaces of the total surface (Figure 3).
Figure 3. Way of the estimation of the cross-sectional knot proportion.
78
Figure 4 depicts the way of the knot measurement and calculation of the total cross-sectional knot
proportion according to which the grade of sawn and squared timber is determined.
Figure 4. Cross-section of the board with the knot occurrence.
Effect of knots using bending method

When using the bending method, the influence of knots is manifested directly in the values of strength
and elasticity characteristics which is one of the criteria for the determination of strength classes by
means of the bending method (STN EN 408). When designing timber structures, the strength classes
are considered according to the standards STN P ENV 1995-1-1 – SI and SII. There is currently an
endeavor to include the strength class S0 which corresponds to a class C 30 according to STN EN 338.
Effect of knot occurrence was experimentally observed according to visual and bending method. From
the viewpoint of the knot occurrence there was a relative correspondence in comparison with the
bending method. Visual method overestimate parts on a board without knots. Therefore, bending
(mechanical) method is more objective and reliable.
According to the law of Slovak republic, a construction company is responsible for structural timber
grading. In reality, only Slovak standard STN 49 1531 is only used for this purpose. Mechanical
grading according to STN EN 51 is rarely used. On the other side, Slovak construction companies
working for European market used required national standards of EU states (DIN, Eurocode 5).
The normalization is a living process. Even EU standard implemented in Slovakia are not without
problems. After tragic disasters of wooden construction last year, the Slovak Institute of Technical
Standardization issued a recommendation of not using the standard STN P ENV 1995-1-1 known as
Eurokód 5 (SÚTN 2007). The reason is underestimation of joints according to this standard.
CONCLUSIONS
- Common European market leads in majority EU countries towards diverse supply of roundwood
or timber. This supply increase number of species used by the wood processing industry, enlarge
quality composition as well as expands different characteristic properties of wood supply.
Normalization at the European level unifies and simplifies examination of selected material
feasibility for further utilization. In the area of timber quality grading, many wood species are not
cover; there are elaborated only the most used European wood species.
- Process of European standard implementation is not static matter. There is a need for revising of
normalization process, which was confirmed by negative results of implemented of an EU
standard on wood building constructions.
- Slovak republic, as a member of European Committee for Standardization, can not only
implement others standards, but also can contribute on editing or creation new ones.
79
ACKNOWLEDGEMENT
This work was supported by Slovak Research and Development Agency under the contract No.
APVV-0282-06.
REFERENCES
1. Gejdoš, M. Teischinger, A., Suchomel, J., (2006): Comparison of roundwood sorting according to
European standards and the national grading rules in Slovakia and Austria. In: Proceedings of the
5th International Symposium Wood Structure and Properties ’06, 3-6 September, 2006, Sliač-
Sielnica, Slovakia. pp. 455-460.
2. Green report, (2006): Ministry of Agriculture of the Slovak Republic Bratislava, National Forest
Centre – Forest Research Institute Zvolen. 149p.
3. Šúriková, A., (2002): Timber grading [in Slovak]. Stolársky magazín. (3) 6:10-11.
4. Šúriková, A., (2003): Timber grading [in Slovak]. Stolársky magazín. (4) 11:14-15.
5. STN 48 0055 Qualitative classification of softwood round timber.
6. STN 48 0056 Qualitative classification of hardwood round timber.
7. STN 48 1531 Structural timber. Part 1: Visual strength grading.
8. STN EN 338 Structural timber. Strength classes.
9. STN EN 384 Structural timber. Determination of characteristic values of mechanical properties and
density.
10. STN EN 408 Timber structures. Structural timber and glued laminated timber. Determination of
some physical and mechanical properties.
11. STN EN 1316-1 Hardwood round timber. Qualitative classification - Part 1: Oak and beech.
12. STN EN 1316-2 Hardwood round timber. Qualitative classification - Part 2: Poplar.
13. STN EN 1316-3 Hardwood round timber. Qualitative classification - Part 3: Ash and maples and
sycamore.
14. STN P ENV 1927-1 Qualitative classification of softwood round timber - Part 1: Spruces and firs.
15. STN P ENV 1927-2 Qualitative classification of softwood round timber - Part 2: Pines.
16. STN P ENV 1927-3 Qualitative classification of softwood round timber - Part 3: Larches and
douglas fir.
17. STN P ENV 1995-1-1 Eurocode 5: Design of timber structures. Part 1-1: General rules and rules
for buildings.
18. SÚTN (2007). Announcement. http://www.sutn.org/?menu=details&id=227.
The law of the Slovak Republic No. 264/1999 Z.z. on technical requirements of goods, on conformity
and on changes and supplements of some laws.
80

Evaluation of the French oak for structural use

Lanvin Jean Denis1, Reuling Didier1, Costrel Y2 , Ducerf J2
1.FCBA Timber Engineering Unit Bordeaux, France
2. APECF, 6 rue Francois 1er, 75000 Paris, France
keywords : Harwood, mechanical properties, visual grading
Abstract
This project was born from the wish of French professional groups (sawmills, Institutions, wood suppliers, …) clustered
within the association APECF to promote French oak species. Their aim is to obtain technical elements necessary to the
diversification of their products, and to develop as well as possible, particularly in housing, the main French forest
resource. With more than four million hectares of Petraea and Robur oak, France holds the greatest reserve of Europe.
In order to position oak like a performing building raw material, professionals realized that some technical and
commercial information were missing. Indeed, until now oak was used for structure in large dimension (traditional frames)
without design. Thus optimization questions of dimensions compared to the supported loads did not arise. Now, structural
design has become current for architects and builders. However the raw material knowledge must be important and
resistance rules clearly laid down to ensure the guarantees on the constructions viability.
420 tested trees in 2 species (Petraea and Robur oak) with 3 Dbh class (25 cm up to more 65 cm) were taken
from 70 stands exploited in 6 French areas. 3500 beams (50*150*3 000 mm) were used in non-destructive and destructive
tests. A sub sample (230 beams) was done in squared timber.
Three main criteria (density, modulus of elasticity and modulus of rupture) establishing the strength classes according to
EN 338 are high, conferring on oak very good mechanical properties. The optimal grading gives 71% in D40 and 13% in
D30 at the end of bending tests. Therefore, correlations between knots and mechanical properties
are low, visual grading cannot grade efficiency beams (57 % in D24 and 27% in D18). Investigations are ongoing
in order to find oak in upper grade by a previous log sorting.
For the one hand, CTBA proposed new elements of visual grading in order to use oak in building into the French
construction standard NF B 52-001, and on the other hand established the grade D24-D18 into the EN 338.
Introduction
With more than four million hectares of Petraea and Robur oak (40% of France Hardwood forest),
France holds the greatest reserve of Europe. In order to position oak like a performing building raw
material, professionals realized that some technical and commercial information were missing. Indeed,
until now oak was used for structures in squared timber (traditional frames) without design. Now,
structural design has become current for architects and builders.
However the raw material knowledge must be important and grading rules clearly laid down to ensure
the guarantees on the constructions viability. In one hand, since the visual grading for French oak is
not established, saw miller can not sell product with EC mark and in the other hand, the yields in high
mechanical class is not as high as expected.
Enhancing the value of a species for use in structural timber requires an accurate knowledge of its
characteristics, especially the mechanical ones. To suppress the lack of information, FCBA engaged in
2002 a study of the 2 Oak species (Petraea and Robur oak) at the national level. 6 French areas were
retained for sampling and one of the questions is the possibility to improve the output of products
usable in housing by using graded timber.
Materials and methods

70 stands were selected and 420 trees sampled in six-forest subdivisions defined by the French trees
Inventories (IFN). In order to be representative with the oak resource, the stands and the trees have
been selected according to:
1. The geographical distribution: taking away within six forest subdivision,
2. Two sylviculture treatments (high forest stand and coppice with standard)
3. The characteristics of the site (acid-neutral sites and wet site),
4. The DBH: from 25 up to 80 cm,
81

Table 1: Anatomical and morphological properties of the sample

Averag
420 trees CV %
e
Age (year) 119 24%
DBH (cm) 48 26%
Texture (%) 74 8%
Annual growth ring
2,4 27%
(mm)
The adult resource of oak has low annual growth ring width related to the species and the sylviculture.
Its texture is strong compared to resinous such as Spruce and European Silver Fir, which confers a
high modulus of elasticity (particularly in bending strength).
Each tree has been processed into test pieces (50*150 mm²) by 3000 mm long. The different pieces
(more 2500 pieces tested in edgewise four points bending) have been identified to know their origin
(forest, stand, log, etc, …). The following table give some overview on properties.
Table 2: Mechanical average values for French oak:
Coefficien
Average t of
variation
Density to 12% (kg/m3) 722 8%
Dimension of knots
projection in timber face 29 62%
(%)
Dimension of knots
projection in timber edge 40 72%
(%)
Knot Area Ratio : KAR
22 47%
(%)
Annual growth ring (mm) 2.4 68%
Modulus of Elasticity
13.1 27%
at 12 % MOE (GPa)
Modulus of Rupture
51.9 42%
MOR (MPa)
The mean mechanical properties of the Oak are very interesting, only the modulus of rupture can be a
criterion discriminating in the choice of the timber pieces according to the various uses in the
construction.
Mechanical grading rules
After destructive testing, the mechanical potential of each piece is known and the oak timber pieces
can be classified according to the mechanical criteria defined in the standard EN 338. Therefore, two
classes D24 and D18 did not exist yet but they will be included into the next draft of the EN 1912. The
outputs would be the following for optimal grading and visual grading.
82

Table 3: Optimal grading for squared edged timber for two kind of sorting
MOR MOEL MV Number
Grade (5th (average (5th of Output
s perc) ) perc) specimen s
MPa GPa kg/m³ s
D40 40 14.4 635 1516 71%
D30 30 11 622 270 13%
D24 24.1 10.1 612 160 7%
D18 * * * * *
Reject 8.8 8.1 620 204 9%
Total 2150 100%
The aesthetic of the oak timber pieces is strongly knotty what implies a low correlation between the knots
and the mechanical properties. Visual grading (with knots criteria) is not efficient as illustrated below.
Table 4: Visual grading for squared edged timber for two kind of sorting
MOR MOEL MV Number
MPa GPa kg/m³ s
D40 * * * * *
D30 * * * * *
D24 24.7 14 625 1197 56%
D18 18.4 12.6 633 563 26%
Reject 13 10.8 637 390 18%
Total 2150 100%
MOR MOEL MV Number

MPa GPa kg/m³ s
D35 38.1 16.3 744 58 3%
D27 27 14.9 704 679 32%
D18 18.6 12.5 625 1118 52%
Reject 12.3 10.4 584 295 14%
Total 2150 100%
Knot effects in large squared timber

The sub sample in large squared timber (230 beams) was done and we have observed that the
relationship between knottiness and mechanical properties were better.
Table 5: Visual grading for squared timber
MOR MOEL MV Number
MPa GPa kg/m³ s
D40 * * * * *
D30 37.6 12.6 709 118 49%
D24 28.1 10.4 732 123 51%
D18 * * * * *
Reject * * * * *
Total 242 100%
83

In case of squared timber, the correspondence between aesthetic and mechanical behaviour is well
established as shown in the following curve.
Figure 1: distribution in box plot of MOR for squared timber per aesthetic grading.
Figure 2: Overview of FCBA bending device for squared timber.
The analysis of results of visual criteria and mechanical properties leads to grade French oak in a new
draft of NF B 52-001 included all cross sections. The results of sorting are summarized as follows:
Table 6: Visual grading for French oak for all cross sections
MOR MOEL MV Number
MPa GPa kg/m³ s
D40 * * * * *
D30 31.1 14.2 632 611 26%
D24 24.5 13.5 632 539 23%
D18 * * * * *
Reject 16 12 634 1242 52%
Total 2392 100%
84

The gap between table 4 and table 6 is not enough relevant (reject is too high) but the D30 strength
class is now present. However 50% of the material cannot be used in construction but could be
affected to D16. A new sampling in medium section should be done in order to check, if we can merge
all cross section in a same interesting grading rules or if we are obliged to establish two grading rules
functions of the cross section.
Logs sorting according their diameter (squared edged timber)
Investigations are ongoing in order to find oak in upper grade by a previous log sorting. The
mechanical behaviour decreases with the annual growth rings. When the trees are young, better are the
strength at the opposite of resinous trees. That’s the reason we have established visual grading for
three classes of logs diameter as follows.
Table 7: Visual grading for French oak for DBH [25 – 40 cm] (average age of the tree : 100 years
old)
MOR MOEL MV Number
MPa GPa kg/m³ s
D40 41.7 16.8 639 91 25%
D30 30.1 15.1 623 145 40%
D24 26 14.5 646 73 20%
D18 18.9 13.4 645 50 14%
Reject * * * * *
Total 359 100%
Table 8: Visual grading for French oak for DBH [40 – 60 cm] (average age of the tree : 125 years
old)
MOR MOEL MV Number
MPa GPa kg/m³ s
D40 * * * * *
D30 30.6 14.5 635 377 28%
D24 24.9 13.9 632 248 19%
D18 18.2 12.3 641 410 31%
Reject 14.5 11.2 641 289 22%
Total 1324 100%
Table 9: Visual grading for French oak for DBH [> 60 cm] (average age of the tree : 140 years
old)
MOR MOEL MV Number
s perc) ) perc)specimen s
MPa GPa kg/m³ s
D40 * * * * *
D30 * * * * *
D24 * * * * *
D18 18.2 12.3 606 346 74%
Reject 10.8 9.9 624 120 26%
Total 466 100%
Today, any standard establish requirement to take into account log information. However, the analysis
of those tables show that “young” oak give better lumber for housing than the “old” ones.
85

At the end, FCBA simulates the proportion of specimen [Reuling D., 2007], which could be graded
according to EN 338 mechanical classes without any visual grading, only with the provenance of wood
in the tree.
Table 10: Grade modelled per forest parameters (NS not significant).
40-
25-40cm >60cm
60cm
H1 (3.3m) D30 D24 D18
H2 (6.6 m) D27 D18 D14
H3 (9.9 m) D20 D16 Reject
H4 (13.2 m) NS Reject NS
Optimum grading
88% 70% 58%
D40
This model did not take into account the squared timber, the forest information were missing.
Conclusions
This project was born from the wish of French professional groups (sawmills, Institutions, wood
suppliers, …) clustered within the association APECF to promote French oak species. Their aim is to
obtain technical elements necessary to the diversification of their products, and to develop as well as
possible, particularly in housing, the main French forest resource.
420 tested trees in 2 species (Petraea and Robur oak) with 3 Dbh class (25 cm up to more 80 cm) were
taken from 70 stands exploited in 6 French areas. More than 2500 beams were used in non-destructive
and destructive tests. A sub sample (230 beams) was done in squared timber.
Three main criteria (density, modulus of elasticity and modulus of rupture) establishing the strength
classes according to EN 338 are high, conferring on oak very good mechanical properties.
The optimal grading gives 71% in D40 and 13% in D30 at the end of bending tests. Therefore,
correlations between knots and mechanical properties are low, visual grading cannot grade efficiently
beams (56 % in D24 and 26% in D18). If the effect of knottiness in squared edged timber is taken into
account with squared timber, the yield becomes 26% in D30 and 23% in D24.
Investigations find oak in upper grade by a previous log sorting. A simple model has been developed
to predict quality of product (mechanical behaviour) with height and diameter of trees.
Therefore, the model did not prevent the final grading for the EC marking (EN 14-081 2006) which be
done by visual or mechanical grading but increase the output in higher mechanical classes.
Saw millers have now basic tool to determine easily the final quality of wooden construction product
with forest parameters according to their production and gain add values on final product.
Now, we proceed with mechanical grading in different technologies (X ray, ultrasonic or vibratory
machines). We observed the same difficulties to grade oak than in visual grading and the relevance of
a previous log sorting to increase yield of the highest grade for Oak (D 40).
Références
EN 338 (2003) Structural timber. Strength classes
EN 1912 A1 (2007) Structural timber. Strength classes. Assignment of visual grades and
species
EN 14081 parts 1 – 4 (2006) : Timber structures - Strength graded structural timber with
rectangular cross section
NF B 52 001 (2007) : Regulations governing the use of timber in structure - Visual
classification for the use of French coniferous and deciduous sawn timber in structures.
Reuling D, Lanvin JD, Chatelu R, Godard J. (2007) “Evaluation of the French oak for structural
use” International scientific conference on hardwood processing 2007 congress, Quebec 24-26
of September 2007
86
Indisputable key – a research and demonstration project

aiming at an improved quality through traceability systems
Anders Lycken, Jarl-Gunnar Salin
SP Technical Research Institute of Sweden
P.O. Box 5609, SE-114 86 Stockholm, Sweden
Anders.Lycken@sp.se , jarlgunnar.salin@sp.se
Keywords: EU-project, process improvement, traceability, RFID transponder.
ABSTRACT
A large EU project – Indisputable key –started in October 2006 and is focused on traceability in the production chain
from the forest to the end user. The objective of this project is to develop methodologies and advanced technologies that
can improve the use of wood and optimize the forestry-wood production line through the chain of transformations while
minimizing environmental impact. The magnitude of the problem, based on an estimation of the total production of sawn
wood in Europe 2002, is that approximately 25 million m3 of wood raw material is going to waste, worth billions of €.
A major reason is that important information regarding the raw material, the process or the final use is not available
throughout the process.
Making information available on-line, for process steering as well as for process evaluation, at different stages along
the forestry-wood production chain requires automatic traceability systems. The developed systems will be based on the
Individual Associated Data (IAD) concept. The logs, boards and packages are individually marked with unique numbers
and data added to a database, reachable through the whole production chain. Each time a measurement is made or
a decision is taken the data is added to the database. The log is marked in the harvesting process and read when
forwarding, when the logs are measured and when sawn. The boards are marked after sawing and read in green sorting as
well as after final sorting and at the end customer. For log marking especially the use of RFID transponders
is investigated, including applicator and reader system development. Readers are installed at selected points in the sawmill
production chain so that the piece can be detected and additional information obtained in the process connected to it in the
database. Decision-making at different points (sawing pattern, selection of processing parameters and product)
is based on all data gathered up to that point.
This paper describes the project in more detail, the benefits expected and some preliminary results obtained.
INTRODUCTION
Today wood is treated as a bulk material, while in reality wood is an inhomogeneous biological
material having individual quality and property attributes, depending on geographical origin, growing
conditions, age etc. Also the wood material itself is inhomogeneous; heartwood and sapwood, butt logs
and top logs, density, moisture content, all these factors will have an influence on the final product
appearance, durability, strength etc. As a consequence of the lack of information about individual
property aspects, end products have to be graded in the final stages of the process to be able to satisfy
customer demands. However, accurate data such as origin of logs, location in stem, treatments
at different stages of the production chain, are collected in different process steps and are available
during each step, but most of the data are lost later in the process as illustrated in Figure 1.
The core problem is to, more or less automatically, acquire the required data, i.e. product and
production parameters for individual items, and to keep them throughout the production chain.
The basis for having the necessary parameters at hand when needed is to have data associated
with each specific item – log, board or package – data that can be followed through all steps
of the production chain. The ultimate target is an unbroken traceability chain from the tree in the forest
to the end user of the wooden product.
The main objective of the EU-project ‘Indisputable key’ is to develop methodology and technology
needed for the implementation of traceability in the forestry-wood chain. The project aims also
to demonstrate that the technology developed has the potential to improve material utilization, to select
the optimal raw material for each product and to introduce a more holistic approach to environmental
and economic issues. Not to forget the possibility to use all data collected for an item for the local
optimisation of the procedure in each unit operation based on a global knowledge.
87
ACCUMULATED
INFORMATION
LOST
LOG BOARD
FOREST SORTING SAWING SORTING DRYING GRADING
-
FORESTRY WOOD CHAIN PRODUCTION
PROCESSES
Figure 1. In traditional production, valuable data are produced, used and then lost. Information is created or
recreated where it is needed in different locations: in the forest, in the sawmill’s different production processes and
at the end manufacturer or user. With traceability and IAD it is possible to retrieve and use all accumulated data
instead of only the data produced at each stage.
Table 1. Participants in Indisputable key.
Participant name Country

SP Technical Research Institute of Sweden SE There are 29 partners from five countries –
Association Forêt Cellulose FR
Estonia, Finland; France. Norway and Sweden
– involved in the project. These are listed in
Royal Institute of Technology SE
Table 1.
Centre Technique du Bois et de l’Ameublement FR
CIRIS Engineering FR
The partner list includes both universities,
research institutes, equipment developers,
IVL Swedish Environmental Research Institute SE
software suppliers and companies in the forest-
Rottne Industri AB SE
wood-chain.
Technical Research Centre of Finland FI
SP Technical Research Institute of Sweden is
TietoEnator Forest & Energy Oy FI
the co-ordinating partner for the whole project.
Confidex Ltd. FI
There are several wood related production units
Idesco Oy FI
represented in the list, where demonstrations of
Tampere University of Technology FI certain parts of traceability chains will be
Lappeenranta University of Technology FI implemented. The main demonstration site is
Tallinn University of Technology EE the Malå Sawmill (SETRA Group) in the
Oskando OÜ EE northern part of Sweden. A part of their
AS Hekotek EE production is further used for window
Skog-Data AS NO manufacturing at Norsjöfönster AB – also
Norwegian Forest Research Institute NO a partner. Other sawmills participating
Norsk Treteknisk Institutt NO are Raunio sawmill in Finland and ESAS
Forestry Research Institute of Sweden SE sawmill in Norway.
Sveaskog Förvaltnings AB SE Not only softwood sawmills are involved as
Ducerf Scierie FR production units: also an oak sawmill and a
Raunion Saha Oy FI plywood factory in France and a pole factory in
Eidskog-Stangeskovene AS NO
Norway are places for traceability installations.
Scanpole AS NO The project started in October 2006 and is
Etablissements Pierre Mauchamp S.A. FR scheduled for three years. The total budget for
SETRA Group SE
the project is about 12 million euros.
Norsjöfönster AB SE
Rolpin FR
88
OBJECT IDENTIFICATION BY MARKING AND READING

The Individual Associated Data (IAD) concept requires of course that the object to which data
is associated is marked in some way so that a unique coupling between data and the object is created
and maintained throughout the production process. It is equally important that the mark can be read
later on in the chain so that new data can be added to the list of properties for that object
in the database. It is obvious that both the marking and reading procedures should preferably be
as automatic as possible. There is a multitude of marking methods used for very different purposes
and different items – we are all familiar with some of them in every day life; bar codes, printed
numerical identifications, colour markings etc.
A special attention is in this project paid to Radio Frequency Identification (RFID) transponders,
especially for marking of logs, i.e. from the forest to the saw intake. Popularly described gives
an RFID transponder an “echo” when subjected to an electro-magnetic pulse with a predetermined
frequency. This echo is modulated according to the transponder identity, i.e. it gives a unique number
to the object to which the transponder is attached. Transponders are today available in many different
forms – from the size of a short match to thin labels – and the cheapest versions cost 10 – 20 euro cents
when produced in large amounts. The realization depends on the frequency used, which in turn
influences the reading distance.
Development of transponders suitable for softwood logs (Nordic countries) is a very demanding
challenge. The harvester is using about 20 seconds per tree for felling, delimbing and cutting,
and produces generally two to four logs (butt, middle and top logs). In order not to decrease
the productivity, the application of a transponder should be finished in seconds. Further, as a lot of
transponders will end up at a pulp mill – either directly or with chips from the sawmill – no plastics
can be tolerated. In addition the harsh harvester environment – rough handling, dirt, snow and ice –
will cause problems. It will be interesting to see the result from the project in this respect.
For hardwoods the situation is quite different. An oak sawmill in France is one of the industrial
partners in the project and oak trees are normally harvested manually (chainsaw) and forwarded
as a single long log to the mill. Due to the high value of each oak log, a much more individual
procedure is found all the way and application of a transponder should be a relatively easy task.
If the transponder marking concept fails, there are other options such as colour marking etc.
An interesting method is to use nanoparticle inks that exposed to IR light will fluoresce in a way that
can be detected through a reasonably thick layer of ice or dirt.
Another challenge regarding marking is the transfer of the log identity to the boards sawn from the log.
The same problem may occur when boards are divided into smaller pieces in a secondary
manufacturing. In some cases the boards from the saw are brought forward in a linear, unmixed way
and the transfer of identity is easy. In other cases boards from several logs are mixed before marking
is possible and this requires some additional information – as an example matching of log length and
board lengths, maybe complemented by some other parameters like knot structure. In this project only
centre boards are included in the traceability system.
Boards are preferably marked by printing a machine readable digital code or bar code, not by
a transponder as these probably are too expensive. An example is seen in Figure 2.
89
Figure 2. Code marked board ends.
It is of course important that the code can be correctly and automatically read so that the object can be
identified later on in the chain. In some special cases all codes need not to be read; a pile of logs is
identified (truck driver in the forest) by identifying one or a few logs in the pile. A stickered kiln
package can be identified from a few boards even if the package mark is missing. However, it has
to be accepted that some marks will not be read correctly, and the corresponding objects will then
cause some problems in databases etc. On the other hand, it should be mentioned that even if the mark
is not read correctly, there are some methods to reconstruct the information. If a 10 digit numerical
code is used then there are 1010 possible combinations. But consider that only a subset is used where
each combination differs from all other combinations in the subset, in at least three positions.
Then even if there is a reading error in one position, the original combination can be determined with a
high probability, as there is only one combination in the subset that differs in one position from
the combination read (error correcting code principle). The most important thing in this case is that all
numbers put in the database are correct. If a reader is uncertain of a number, that number will not be
stored.
BENEFITS WITH TRACEABILITY FROM THE FOREST AND WITHIN A SAWMILL

The public may conceive traceability as a possibility to check whether a wooden product origins from
a certified forest or not, and other similar issues. A much more important benefit, however,
is the possibility to use traceability for optimisation of production procedures, product quality
improvement, waste minimisation etc.
The traceability makes it possible to calibrate the measurement systems used in the forest to the ones
used at the sawmill. As the logs are cut to length in the forest, and the wanted log lengths are
depending on log diameter and log quality, it is necessary that the harvester’s measurement system
is correct and maintained. With feed-back built on traceability it is a simple task to compare,
log by log, the measurements made in the forest with the ones in the sawmill.
One problem for most sawmills is to measure the log diameter correct, as the measurement and sorting
is performed on logs where the bark is still on (some sawmills debark the logs before sorting).
The interesting measure is the diameter under bark, and that is calculated from measurement above
bark using a bark function, which is empirically established. By knowing the logs’ diameters when
sorting (above bark) and comparing them with the diameter in the saw intake (under bark),
it is possible to improve the bark functions both at the log sorting station and in the harvester which
will make the sorting more efficient and improve the yield.
When marking logs, either in the forest or at the sawmill’s log sorting station, it is possible to compare
log quality with the quality of the final board. A comparison could also be made between the green
sorter and the final grading. This makes it possible to calibrate quality models to help grade and sort
the logs and the boards to better fit the final products.
Some features are more easily detected in green (undried) condition than they are in dry.
That is the case with some knot types where resin complicates the definition of the knot’s type and size
90
after drying. A tracing between green sorting and final grading, where each grading system looks
for features visible best at that place improves the overall grading.
One major board quality problem is twist, especially regarding studs for building purposes. It is well
known that the basic reason for twist is spiral grain, i.e. the fibres in the log are not aligned completely
in the direction of the log but form a helix around the pith. As the shrinkage properties are quite
different in the fibre direction compared to the transverse direction, the grain angle will cause twist
when the board is dried. For many species – like Norway spruce (Picea abies) – the spiral grain angle
is highest close to the pith and decreases with the radial distance from the pith. The boards most prone
to twist are thus those sawn close to the pith.
If the grain angle is measured then a rather good estimate of the twist development is obtained.
In an industrial environment the grain angle can be measured using laser. If a circular laser beam is
pointed at a wooden surface, the light spot will be elongated in the direction of the fibres (tracheid
effect).
Figure 3. The tracheid effect is shown in the pictures, where the laser point is elongated into an ellipse, following the
fibre direction when lit on a wooden surface, to the left of a board and to the right of a log.
The direction of the main axis of this ellipse can be measured with camera technique. This makes it
possible to determine on-line if a log will give prone to twist boards or not. The measurement can be
made on the surface of the (barked) log, which together with the log diameter (log class) gives
an estimate of the twist level. This log can then be used for products that are not sensitive to twist or
even directed to a pulp mill. A better estimate of the twist level for the centre boards is obtained
by measuring the grain angle on the surface of the block from the first saw. These centre boards may
then be characterized as prone to twist if a high grain angle is found and this information is entered
into the traceability system.
Later on in the production chain the boards prone to twist are directed to kiln packages that are placed
as the bottom package in the kiln stack. Boards kept straight during drying will show a final twist level
that is about half of the level found for boards free to move during the drying process. Almost all
centre boards will twist in the same direction. Based on this fact an improvement is obtained if the
prone to twist boards are dried in a pre-twisted position, i.e. twisted in the opposite direction compared
to the inherent direction. Measures of this kind can be implemented only if the appropriate information
is associated with the object in question and retrievable downstream.
The majority of customer complaints are directly or indirectly related to product moisture content and
thus to the drying process. Traceability can be used for investigations regarding the reasons for such
complains. If for instance too high moisture content is found, the history of that consignment can be
analysed; dried in which type of kiln, kiln number, location in kiln, drying schedule used, time
between stickering and loading into kiln, responsible kiln operator, disturbances in boiler operation etc.
This kind of information will help to find and correct an accidental production flaw or to initiate
new or improve existing procedures in the production chain.
91
Within the sawmill it is also possible to pinpoint kilns in need for repair, as if the boards dried
are more cracked or distorted than boards dried in other kilns.
Boards with different features need different drying programs, and should be sorted accordingly;
for instance do inner and outer boards contain different amount of heartwood, and should be dried with
different drying programs. This could be solved with a marker direct after the saw and a reader
in connection with the green sorting.
It is known to saw millers that some kinds of boards are causing problems in the production. There are,
though, some discrepancies about which boards it is. With tracing of boards, from logs, it is easy
to find out which type of boards, and from which type of logs they emanate. When the problem boards
are mapped, it is possible to see to that the logs never enter the sawmill but are sorted out at the log
sorting station. Problems might be caused, for instance, by twisted boards or boards with a large
amount of reaction wood.
Special products, that need special treatment, or are aimed for certain customers, might be traced and
steered to the right process fit for that product. One example is boards with long internode lengths.
These boards are scanned, knots and other defects are cut off, and the remaining knot free pieces
are finger jointed for making window frames. If the lengths between knots are increased, the number
of cuts and joints are reduced and the yield is improved, both for production and material. Logs with
long distance between knots, suitable for this product, might be traced all the way from the forest
and kept track of through the sawmill process to the window maker. It is also possible for the window
factory to know exactly the yield improvement, due to the “new” board quality.
The sawmill can keep thorough track of which forest is most suitable for which product, based on the
final grading, and thereby direct felling crew and set time for felling based on collected empirical data
from earlier volume and quality yield. The payment can be based on actual yield, as it is possible
to know exactly from which supplier each log and board is bought.
Stock management and logistics can be much improved by efficient traceability of logs and packages
within the sawmill’s gates. One important duty on sawmills is searching for “lost” packages.
With an automatic package tracer, based on GPS or similar technology, every package movement
is recorded and each package’s position is known to the system. The fork lifts are directed to the right
storage place to pick up the package, whose number is read automatically to confirm picking up, and
then the fork lift is directed to the right delivery position, be it a kiln, another temporary storage
or another process. By this system the storage time before the dryer can be known and minimised, and
cracks kept to a minimum. It is also possible to know exactly how much timber is in stock, where it is,
at each process stage, as green packages, in the dryers, as dry packages and as packages ready for sale
and shipping.
Each harvester and harvester operator has an operating level based on inner and outer conditions, such
as training, forest type machine maintenance etc. With traceability system it is possible to locate where
training level is low or machine maintenance is needed, as well as it is possible to see if certain forests
are tougher to harvest than other, due to more damage to the logs.
CONCLUSIONS
According to the agreed plans for this EU project –Indisputable key – the strategic objective is
to improve competitiveness and resource efficiency of the European wood industry by initiating and
stimulating an industrial breakthrough of traceability systems for biological raw materials in general
and for wood in particular. The project will specify, develop, introduce, evaluate and disseminate
a powerful distributed and collaborative network-oriented system for improving the use of wood raw
material and production resources. The project results will make it possible to extend the knowledge
about the usability potential of wood for different products, greatly beyond what is realistic today
with traditional production feed-back systems.
A webpage www.indisputablekey.com is established where more information can be found
and progress followed up on how successful the fulfilling of the strategic objectives will be, during
the two remaining years.
92
Variation in yield of strength graded timber due to board origin

Audun Øvrum
Norsk Treteknisk Institutt (NTI), P.O. Box 113 Blindern, NO-0314 Oslo, Norway, audun.ovrum@treteknisk.no
Keywords: forest variation, grade yield, strength grading
ABSTRACT
Normally strength grading is done on batches of timber without any information about the origin of the boards. The yield
from the strength grading machines is hence accepted, and there is rarely any active pre grading of batches or logs to
increase the yield of high grades. Different batches of timber will evidently vary in yield of different grades, but no active
investigation is normally done on this matter due to the large area saw logs are purchased from. Also it is very difficult to
track boards backwards towards origin in an industrial saw mill today. In this investigation a batch of timber where all
data from the stand, the trees position in the stand and the boards position in the tree where collected, and linked to the
strength grade of the boards given by a Dynagrade machine. The stands were sampled after fertility and general forest
quality, with site indices ranging from G11 to G23. The results show that there is a great variation in yield of strength
grades from timber from different stands even in stands in very close proximity to each other. An assessment of the forest
quality before logging showed very good correlation with the yield in high strength classes and indicates a possibility of
pre-sorting in the forest.
INTRODUCTION
The later years work has been done on making models for the strength properties in boards based on
log parameters (Brännström, Oja et al. 2007; Jäppinen 2000; Oja, Grundberg et al. 2001). Jäppinen
(2000) found that through a pre-sorting of logs based on variables from a 3D scanner the yield of
glulam lamellas could be improved significantly. Oja et.al. (2001) found that data from an X-ray
logscanner could give good indications about the stiffness in the centre boards graded by a Cook-
Bolinder machine. Brännström et.al. (2007) found that an X-ray logscanner predicted the strength in
centre boards with the same accuracy as a standard strength grading machine (Goldeneye 702). In
Norway strength grading is done on saw falling batches of timber and no pre-sorting based on 3D
scanners or X-ray scanners are done. In this way there exists no information about the origin of the
boards and the yield from the strength grading machines is hence accepted. Different batches of timber
will evidently vary in yield of different grades, but no active investigation is normally done on this
matter due to the large area saw logs are purchased from in big saw mills. Theories about reasons for
the variation exists of course but it is very difficult to track boards back towards origin in an industrial
saw mill today, and no formal investigations on this matter has been performed. (Chrestin 2000) found
that Norway spruce timber from the mountainous regions in northern Sweden had lower values for
mechanical properties, while timber from central Sweden and some parts of southern Sweden had
higher values for mechanical properties. The mechanical properties was the modulus of elasticity
measured by a Cook-Bolinder machine, the modulus of elasticity tested after EN 408 and the bending
strength tested after EN 408. In this study the aim is to discover the variation in strength grade, i.e. the
modulus of elasticity measured by a Dynagrade-machine, within a specific region, based on a sampling
that covers normal forest stands from the worst to the best case scenario for a typical Norwegian saw
mill.

This study comprises 586 boards which are originating from 108 trees extracted from six different
forest stands. The stands were sampled after fertility and general forest quality, with site indices
ranging from G11 to G23 in the H40 system used in Norway (Tveite 1971). All stands were situated
within the same geographic region within a maximum distance of 16 km between two stands. In each
stand all trees within a specific area were cross measured at breast height and ranked by diameter. The
trees were then grouped in three groups defined as dominant, co-dominant and suppressed. Within
93
each group 6 trees were randomly selected resulting in 18 sample trees from each stand. The trees were
cross-cut in fixed lengths of 4 m or 6 m. The logs were divided according to Nordic practice with a
splitting of the log in the middle with a saw cut –heart splitting. The dimensions ranged from 38 x 100
mm2 to 50 x 225 mm2, and dimensions were given by the sawing pattern giving the maximum sawing
yield from each log. The boards were then dried by yard seasoning for two years giving a moisture
content of about 18 % at the time of grading. Finally they were machine strength graded by a
Dynagrade, which is the most common machine for strength grading in Norway. The Dynagrade
measures the resonance frequencies originating from a strike by a metal hammer in the end of the
board. Together with the length measured by a laser the machine calculates the dynamic modulus of
elasticity (Edyn). This Edyn is correlated to the strength of the boards resulting in an allocation of boards
in to strength classes. The borders for the different strength classes are set after the requirements in EN
519 or EN 14081, but in this study the Edyn is used directly as an expression for board strength.
RESULTS
Table 1 shows the data from the different stands with the yield from the strength grading included.
Table 1. Sample size and strength properties in the different stands

Stand No. of Site index Forest Altitude Edyn Grade yield [%]
boards [n] quality [m] (st.dev) C30 C24 C18 Reject
1 122 G17 Medium 450 5 710 000 20 48 30 2
(1 070 436)
2 104 G20 Good 470 6 496 935 53 43 3 1
(867 464)
3 76 G17 Medium 620 5 948 421 17 70 9 4
(927 472)
4 64 G11 Poor 630 5 250 781 3 48 33 16
(780 893)
5 146 G23 Good 350 6 627 671 59 38 3 0
(603 437)
6 64 G11 Poor 770 5 096 563 32 48 16 4
(1 636 775)
Total 576 5 996 910 32 48 16 4
(1 121 127)
A stepwise regression was performed, and the following variables were tested:
• Status of tree in stand (SUPPRESSED, CO-DOMINANT and DOMINANT)
• Log type (BUTT log or not)
• Type of board (INNER, which is the centre boards closest to the pith and OUTER which is all
other boards)
• Forest quality (divided in GOOD , MEDIUM and POOR by a general assessment before
logging)
• Altitude (meters above sea level)
The variables forest quality, status of the tree in a stand and log type were significant contributors to
the fitted model with a significance level of 5 %. The model gave R2= 25,5 % and RMSE=970 333 and
gave a residual plot with no specific trends indicating that the model describes the variance quite well.
The parameter estimates are shown in Table 2.
94
Table 2. Parameter estimate for the regression model for Edyn

Term Estimate Std Error t Ratio Prob>|t|
Intercept 5 893 971 47 097 125,15 0,0000
Tree status[Co-dominant] 106 545 61 753 1,73 0,0850
Tree status[Dominant] -150 279 56 291 -2,67 0,0078
Forest quality[Good] 704 934 56 227 12,54 <,0001
Forest quality[Medium] -43 402 58 184 -0,75 0,4560
Butt log -33 556 42 365 -0,79 0,4287
DISCUSSION
A general assessment of quality in a stand before logging showed very high correlation with the yield
of strength graded timber. The stand giving most C30 with 59 % of the total of boards graded (stand 5)
was assessed as the stand with the highest quality before logging, while the stand giving least C30 with
3 % of the total of boards graded was assessed to have a poor quality. The assessment of the
“goodness” before logging consisted of several parameters, the most important being frequency of top
ruptures from snow and other influences, and taper. All logs from the trees were used as long as they
satisfied the demand for minimum dimension, which in this case was 13 cm outside bark and did not
have rot or excessive sweep which made them impossible to saw in a standard sawing machine. This
implies that many of the logs from the poor sites will not be classified as saw logs since the log quality
will be poorer than the grading rules for saw logs allows. Large knots and knot groups are possibly
much of the reason for the great differences between stands.
The effect of the status of a tree in a stand was significant, but not very important but showed that at
least dominating trees had a lower Edyn than the co-dominant and suppressed trees. This is probably
due to the larger knots in dominant trees, and the larger annual ring widths which will give lower
density.
A significant lower grade in butt logs is somewhat surprising, and contradicts results from other
studies (Blomqvist and Nylinder 1988a; Blomqvist and Nylinder 1988b; Høibø 1991) who found
better quality in butt logs than in the rest of the stem. The effect was however quite small, and may be
caused by coincidences, but the fact that this study comprises all logs and not only classified saw logs
might be a factor in this case. This study is only a first screening of this data, and further studies will
hopefully give more consistent and significant results.
CONCLUSIONS
Great differences between stands in strength properties were found in this study. This indicates that
one can expect large differences in yield from strength grading in one saw mill even if the raw material
are collected from a close proximity. Pre-sorting of logs by choosing stands which will give higher
yields of high strength classes should therefore be of interest.
95
REFERENCES
Blomqvist H, Nylinder M (1988a) 'Samband mellan granstockars geometri, utbyte och kvalitet: resultat
från provsågningar vid Rockhammars sågverk.' Sveriges lantbruksuniversitet, Institutionen för
virkeslära, 202, Uppsala.
Blomqvist H, Nylinder M (1988b) 'Samband mellan tallstockars geometri, utbyte och kvalitet: Resultat
från provsågningar vid Rockhammars sågverk.' Sveriges lantbruksuniversitet, Institutionen för
virkeslära, 91-576-3573-0, Uppsala.
Brännström M, Oja J, Grönlund A (2007) Predicting board strength by X-ray scanning of logs: The
impact of different measurement concepts. Scandinavian Journal of Forest Research 22:60 - 70.
Chrestin H (2000) Mechanical properties and strength grading of Norway spruce timber of different
origins. In 'Proceedings from the World Conference on Timber Engineering'. Whistler Resort, British
Columbia, Canada
Høibø OA (1991) Virkeskvaliteten til gran (Picea abies (L.) Karst) plantet med forskjellig avstand.
Doctoral thesis, Agricultural University of Norway.
Jäppinen A (2000) Automatic sorting of saw logs by grade. Doctoral thesis, Swedish University of
Agricultural Sciences.
Oja J, Grundberg S, Grönlund A (2001) Predicting the Stiffness of Sawn Products by X-ray Scanning
of Norway Spruce Saw Logs. Scandinavian Journal of Forest Research 16:88 - 96.
Tveite B (1971) 'Site-index curves for Norway Spruce (Picea abies (L.) Karst.).' Meddelelser Norsk
institutt for skogforskning, 33(1), Ås.
96
Aspects of timber quality in the United Kingdom

Tim Reynolds
Building Research Establishment, Garston, Watford, WD25 9XX. UK. reynoldst@bre.co.uk:
Keywords: Timber distortion, grade, knots, spiral grain, trussed rafters.
ABSTRACT
Developments in the use of structural timber particular to the UK are discussed together with end user quality
requirements. Case histories of problems resulting from poor specification, selection and understanding of wood behaviour
are given. Particular reference is made to the qualities of British-grown Sitka spruce.
INTRODUCTION
Timber is a highly capable and versatile construction material, used in applications as diverse as roof
tiles (shingles) to foundation piles. Aside from being man’s only truly sustainable building material,
timber is easy to work with being easy to cut and fix using a wide variety of connection types.
Probably the biggest limiting factor in timber usage is its lack of dimensional stability on drying.
Timber which distorts on drying can cause problems with fixings (eg joist hangers), result in cracks
in plasterboard and creaking or uneven floors; whilst excessive differential movement of timber frame
can lead to problems with windows, doors and services. These building defects can lead to extremely
expensive remedial work, including re-housing of occupants.
For General Structural (ie C16) timber BS 4978 (2007) Visual strength grading of softwood which is
often quoted in the UK by building designers, gives the following tolerances for distortion at 20%
moisture content:
• bow not greater than 20mm over 2m,

• spring not greater than 12mm over 2m,
• twist not greater than 2mm per 25mm over 2m.
These are clearly considerable deviations from the ideal of timber which is straight, at least to the eye.
In service moisture contents are typically around 12% for stud walls, but may be as low as 6%
for intermediate floor joists exposed to effects of high levels of central heating or under floor heating.
Distortion and shrinkage can therefore develop in service.
Timber is not the only material available for use in construction. Notable in the UK are recent
developments in competing, alternative construction methods such large format thin joint masonry,
light steel frame and Structural Insulated Panels. Off-site production methods using closed panel
systems and modular units are now commonplace. Some system builders are known to have switched
from timber frame to non-timber panelized systems specifically because of quality issues with timber.
Good dimensional accuracy is needed to incorporate prefabricated bathroom and kitchen sub-
assemblies. Even floor coverings are pre-cut in some instances to fit rooms. Although the proportion
of timber frame construction has been increasing in recent years, particularly with the advent of multi-
storey timber frame, on occasion there have been problems with differential movement caused by
frame shrinkage. These problems can be caused in part by use of timber at high moisture content.
For both timber frame and conventional construction there has been a significant move away from the
use of solid floor joists to plyweb and metal web beams, specifically because of problems with timber
distortion and shrinkage. These engineered wood products generally also require higher grades
of timber such as TR26 or C24, as defined by BS 5268 (2002) Structural use of timber. Trussed
97
rafters (which were invented in the 1950s by Shirley in Florida, and introduced into the UK in the late
1960’s) also gravitate through design code requirements towards the usage of higher grade and better
quality timber. BS 5268 Part 3 (2006) Code of practice for trussed rafter roofs stipulates much tighter
limits for the distortion of timber used for trussed rafters:
• bow not greater than 10mm over 2m,

• spring not greater than 4mm over 2m,
• twist not greater than 1mm per 25mm over 2m.
The code also has other timber quality requirements for other defects such as fissures, wane and knots
(with particular reference to their effect on connector plates).
QUALITY OF SITKA SPRUCE
The properties of British-grown Sitka spruce were the subject of extensive study by the Forest
Products Research Laboratory (FPRL) during the 1930’s, with the overall conclusion that, with
suitable silvicultural management, it ought eventually to become more comparable with imported
spruce for “utility” purposes. The end uses of timber at that time are in many respects quite different
to that of today, with pit props and trench-lining being examples of uses which no longer exist.
Domestic Sitka spruce and imported European (or Norway) spruce can be compared by reference
to the Ministry of Technology Forest Products Research publication Bulletin 50 (Lavers, 1969).
On the basis of the small clear samples studied, mean density at 12% mc for UK Sitka spruce is given
as 384 kg/m3, whilst the density of European spruce is given at 417 kg/m3. For MOE the values are
8100 N/mm2 and 10200 N/mm2 respectively; and for MOR 67 N/mm2 and 72 N/mm2 respectively.
It is perhaps unreasonable to compare the properties of timbers which have been produced over
different timescales and conditions, yet an established grading and design system is bound to look less
favourably on newer sources timber which are inferior. Nevertheless the possible use of domestic Sitka
spruce for trussed rafters has been the subject of some experimentation (Harrod, 1975). BRE recently
carried out number of structural tests on trussed rafters fabricated from Sitka spruce on behalf of
a commercial client, which passed the requirements of BS 5268 Part 3. British-grown Sitka spruce
generally readily meets the requirements for C16, but where the higher grades such as C24 or TR26
are sought, the necessary strictures of the grading system results in a much higher level of rejects
(Figure 1).
With better tree selection processes and application of scanning technologies, high reject rates could be
avoided when using higher grade combinations. Boards with spiral grain, and high proportions
of juvenile wood or compression wood are fundamentally prone to distortion on drying (Figures 2
and 3). Scanning technologies can also be used to segregate the worst of this timber. Battens with
central pith (termed “box pith”) are prone to twisting even at low levels of spiral grain angle.
A particular problem noted of British grown Sitka spruce is that the majority of boards twist
in the same direction, regardless of their orientation. This can cause problems with glue-laminated
timber. Compression wood can also cause problems with laminated timber particularly when
fabricated at high moisture content (Figure 4). Board scanners can be used to sort timber on the basis
of knot content, and can also be used to cross cut timber so that defects are avoided in critical areas
such as at connections or nailing points.
98
FR1 FR1 FR1

100mm 150mm 200mm
C16 100% C16 97% C16 98%
C16/R C16/R C16/R
R 0% R 3% R 2%
C24 38% C24 40% C24 25%
C24/C16/R C16 32% C24/C16/R C16 27% C24/C16/R C16 25%
R 30% R 33% R 50%
FR2 FR2 FR2

100mm 150mm 200mm
C16 92% C16 93% C16 78%
C16/R C16/R C16/R
R 8% R 7% R 22%
C24 60% C24 40% C24 0%
C24/C16/R C16 18% C24/C16/R C16 25% C24/C16/R C16 11%
R 22% R 35% R 89%
FR3 FR3
100mm 150mm
C16 98% C16 100%
C16/R C16/R
R 2% R 0%
C24 57% C24 73%
C24/C16/R C16 28% C24/C16/R C16 12%
R 15% R 15%
FR4 FR4
100mm 150mm
C16 97% C16 89%
C16/R C16/R
R 3% R 11%
C24 77% C24 50%
C24/C16/R C16 10% C24/C16/R C16 14%
R 13% R 36%
Figure 1. Example of Cook Bolinder strength grader reject and pass rates for stands of Sitka spruce.
Figure 2. Bow of timber dried unrestrained, caused by juvenile wood and compression wood.
99
Figure 3. Twisted timber caused by spiral grain and high proximity to pith.
Figure 4. Bow of 3-ply glue-laminated elements caused by compression wood.
In comparative studies carried out by BRE, home grown Sitka spruce was found to be more prone
to distortion that imported timber destined for the timber frame market, and to contain higher levels
of knots (Figure 5 and 6).
Average values (mm)

Twist Bow Spring Cup Moisture (%H2O)
High m.c. 1.31 1.98 0.7 0.14 22
Imported Medium m.c. 5.49 2.37 1.03 0.32 16.9
Low m.c. 6.24 2.53 1.12 0.35 14
High m.c. 2.76 1.44 0.81 0.02 22.6
Home grown Medium m.c. 4.97 2.78 1.61 0.32 16.8
Low m.c. 7.44 2.99 1.75 0.3 14
Figure 5. Comparison of distortion measured in batches of timber studding during drying.
100
Knot Average no. of Average knot

cover (%) knots per batten size (mm2)
face
Imported 0.52 6.1 185
Home-grown 0.82 8.4 206
Figure 6. Comparison of knot content in batches of timber studding.
Figure 7. Battens used in timber frame panel manufacture. Imported timber (left), domestic (right).
PERFORMANCE PROBLEMS WITH TIMBER
Timber distortion can, on occasion, cause problems in service. In the examples given below, both
imported and home-grown timber were involved.
The uneven gallery walls of a museum were investigated by BRE. The wall linings consisted of square
edge butt jointed MDF panels fixed to timber studwork. In several places the boards were out of line
with differences in the studwork alignment of up to around 8mm over 1m. The Architect’s
specification for un-graded softwood for framing out and non-structural use generally had referred to
regularised timber; whilst the specification for cross section dimensions of structural softwood and
hardwood timber referred to both BS EN 336 Tolerance Classes T1 (for sawn surfaces) and T2
(for further processed surfaces), but also to the use of timber generally at 20% moisture content.
The lack of alignment horizontally along the length of the walls was determined to have been caused
by a combination of shrinkage and distortion of the timber since installation together with poor
carpentry. The builder had chosen to use rough sawn un-graded timber, probably at high moisture
content initially.
Similar problems were reported at a timber frame office development in Ireland, where the 3.6m high
wall panels and long corridor lengths extenuated the effects of timber distortion and movement. Wall
panels were installed at high moisture content following attempts to prefabricate on the same site.
The specification had not been clear on the quality of the timber used, with a non-standard size being
quoted. The use of Canadian Lumber Size (CLS) timber which is machined at around 20% moisture
101
content would have avoided many of the problems caused by the use of timber which was rough sawn
at green.
Distortion in the form of twist and cup, together cross grain shrinkage, were noted to have caused
problems of unevenness and creaking of timber floors in a catering development. The joists were
determined to have been installed at high moisture content and were not well fixed in their hangers.
Heat from the ovens had caused the timber to dry to 6% moisture content. Movement in the floor
had caused a linoleum floor covering to leak cleaning water, resulting in a need to replace the decking
and refit the kitchen.
Similar problems were caused for the extension of a hospital where solid timber joists had been used
for an additional storey. The joists had been fixed without restraint from strutting. High levels
of central heating caused cross grain shrinkage resulting in the floor becoming uneven.
DISCUSSION
Whilst British sawmills can and do provide high quality timber, large amounts of low grade timber are
also sold in structural sizes. Non-discerning builders have been known to select and use low quality
timber at inappropriate moisture contents, resulting in problems developing in service. Architects and
developers, on occasion, have also been noted to specify in an illogical or unrealistic manner.
Properly conditioned and regularized timber is available. Inadequate consideration for the movement
and distortion of timber can cause serious problems in construction. These defects can result
in expensive remedial work.
CONCLUSIONS
Items such as trussed rafters generally require high stiffness material with low knot content and low
levels of distortion. Markets such as timber frame also prefer timber with low knot content. High reject
rates have been noted when British grown Sitka spruce is graded for higher quality structural timber.
In a comparative study, domestic timber was also noted to have a higher level of knots and distortion
than imported. With better tree/log selection processes and application of scanning technologies, high
reject rates could be avoided when using higher grade combinations to produce timber suitable
for markets such as trussed rafters and other engineered wood products. Boards with spiral grain, and
high proportions of juvenile wood or compression wood are fundamentally prone to distortion
on drying. Scanning technologies can be used to segregate the worst of this timber, directing it
to usage such as wood-fuel. Board scanners can be used to sort timber on the basis of knot content, and
can also be used to cross cut timber so that defects are avoided in critical areas such as
at connections or nailing points.
ACKNOWLEDGEMENTS
This work was funded by the Forestry Commission and BRE Trust
REFERENCES
Harrod P. S. K. (1975) Trussed rafters and home grown Sitka spruce PRL Report PD 131/75 Limited
Circulation
Lavers G. M. (1969) The strength properties of timbers Bulletin 50. Ministry of Technology.
Forest Products Research Laboratory. HMSO, London.
For BS and EN standards see British Standards Institution, London.
102
Variables affecting the performance of British grown Sitka spruce

Tim Reynolds
Building Research Establishment, Garston, Watford, WD25 9XX. UK. reynoldst@bre.co.uk:
Keywords: Compression Wood, Distortion, Density, Knots, Slope of Grain, Stiffness.
ABSTRACT
This paper details the findings of a study into the relationships between tree, log and board variables to the mechanical
properties and distortion of British grown Sitka spruce. The influence of knots, compression wood, slope of grain, log
shape, density and other parameters is discussed, together with the effect of using these variables as sorting criteria or the
basis of log or sawn timber scanning equipment. The database amassed by BRE in collaboration with Forest Research can
also be used to develop predictive silvicultural models.
INTRODUCTION
Timber is an immensely useful but naturally variable material. Inherent in its structure, both in log
and sawn form, are features which affect qualities of dimensional stability, stiffness and strength.
These include knots, compression wood and spiral grain. Physical characteristics such as log shape,
density, rate of growth, presence of juvenile wood and microfibril angle also affect these qualities and,
in turn, utilisation. British-grown Sitka spruce trees reach maturity relatively quickly; as a consequence
the timber differs significantly from slower grown softwoods imported from northern America
and northern Europe. British-grown Sitka spruce tends to meet a lower structural grade than imported
softwoods, which can exclude it from certain markets. For example, none at present is used for trussed
rafters. Some UK timber frame manufacturers prefer imported timber because they consider the level
of distortion and knots in British-grown material to be too high. Ply web beams and other glued
laminated elements require timber which is not prone to distortion on drying, as does timber supplied
to other markets such as the DIY trade where boards are often stacked unrestrained in heated
buildings. Sorting of timber in the future is likely to be aided by automatic scanning equipment, but the
effects of segregating the better quality from the lower, particularly on strength grading, need to be
considered. With better information on the material being processed, optimised sorting can be
performed. For example, material likely to distort excessively or be rejected at machine grader stage
can be segregated prior to kiln drying. Relatively simple laser/camera setups can be used to measure
grain angle and knot content on boards. Three dimensional log shape scanners are already used
to optimise volumetric yield, and also have the potential to be used to determine timber quality.
By investigating the relationships between timber variables the potential worth of these techniques
can be indicated.

The overall methodology used was to obtain sample batches of timber and relate the predictor
variables of the logs and boards to the criterion variables of the end product. By additionally studying
the behaviour of small scale samples, free of defects such a knots, the effect of variables such as
density and compression wood content could be evaluated. Practical difficulties in the characterisation
and measurement of timber variables, and their adequacy, were also investigated. The main bulk of
the practical work comprised the testing and assessment of around 500 battens of Sitka spruce obtained
from two localities in Scotland. Ninety logs of 3m length were manually assessed, then scanned by a
3D laser scanner. The timber was “curve sawn” into three nominal sizes 200 x 47mm, 150 x 47mm,
100 x 47mm which are commonly used in construction. Aside from the obviously required site, tree,
log and batten number identification numbers and sawn timber dimensions, the following variables
were determined: Tree height and diameter at breast height (dbh), tree taper (height/dbh), log diameter
103
(top, middle, bottom),log taper (ratio of bottom diameter to top diameter), log maximum and overall
curvature, number of whorls per log, whorl spacing (maximum, minimum and mean spacings), log
spiral grain angle, log ovality, log pith eccentricity, log compression wood content, batten density,
batten slope of grain on outer and inner faces and edges, compression wood content on outer and inner
batten faces, knot content (number and area of knots on batten edges and sides, and maximum
concentration of knots in any 300mm span), knot area ratios (MKAR, TKAR etc as defined
by BS 4978), average ring width, percentage of juvenile wood, and distance from pith to centre
of the batten.
The battens were machine strength graded using a Cook Bollinder grader to obtain detailed Indicating
Parameter (IP) values along the board lengths. Since all of the battens were of the same nominal
thickness a value of IP normalised for batten width could be determined. E (Cook Bollinder) was also
calculated. Since the position of every batten was known, stiffness profiles both axially and radially
within the trees could be established. MOE and MOR from bending tests to EN 408 were also
obtained. Distortion measurements of bow, spring twist and cup where obtained following drying.
Other sets of timber were also used for purposes of validation, comparison and further study included
sets of battens obtained from sawmills which were rejected at machine grading and because
of excessive distortion. An extensive dataset of earlier test work on Sitka spruce was also analysed.
SPIRAL GRAIN AND TWIST

Spiral grain has long been known to influence distortion in the form of twist (Stevens 1961). Balodis
(1972) noted that twist increased with increasing angle of spiral grain and decreased with increasing
distance of the board from the pith. His analysis showed that twist was proportional to the ratio
of grain angle: distance from pith; and that the constant of proportionality is a function of the
tangential shrinkage component of the wood. Figure 1 shows the observed relation between twist
and slope of grain recorded on the inner and outer faces of the boards for a subset of 100 x 47mm
battens.
Twist vs slope of Grain
6
y = 0.0999x + 2.2156
2
R = 0.2097
4
2 Twist vs Inner slope

y = 0.2073x - 1.3298
Slope
2
Twist vs Outer slope
R = 0.3323
Linear (Twist vs Outer slope)
0 Linear (Twist vs Inner slope)
-5 0 5 10 15 20 25
-2
-4
-6
Twist
Figure 1. Twist (mm) plotted against slope of grain (deg)
104
A reasonable relationship between slope of grain measured on the faces and twist is evident. However
the relationships between slope of grain on the outer face and inner face are different, most
significantly on the intercept with the y axis. From the above graph (together with detailed inspection
of the boards), the following could be deduced. Only a few boards with high values of slope of grain
which are prone to twisting can be identified and possibly removed. There are very few boards which
twist in the opposite direction to the bulk. Outer face slope of grain is almost always positive (relative
to the observer, given the provisos of the measurement protocol). Twist appears to increase as inner
face slope of grain ranges from negative values to positive. Large negative values of spiral grain are,
apparently, associated with low twist. Battens with inner face spiral grain angles which are positive
tend not to contain pith. A feature also observable in the dataset, since the orientation of the battens wrt
height in the tree is known, is that the majority of battens have come from trees (or sections of trees)
that have remained left hand spiral grained. It is also clear from the above graph that sorting on the
basis of the magnitude of spiral grain angle alone will be ineffective. For predictions of twist,
measurements of spiral grain angle must always be made on a tangential face. The influence of
a variable relating to proximity to pith was also noted, with boards containing pith being particularly
prone to twisting.
COMPRESSION WOOD
Compression wood is a type of reaction wood that tends to form in conifers that have been partially
blown over, in trees on the windward side of exposed plantations, in the lower part of trees growing
on a slope, and below heavy branches (Desch and Dinwoodie, 1996). The greater longitudinal
shrinkage of compression wood causes bow and spring on drying. Compression wood in logs may be
indicated by their shape and form, as shown in other work on Scots pine and Norway spruce
(Warensjo, 2003). Industrial 3D laser scanners are used in many sawmills to optimise yield, and also
have the potential to be used automatically for log sorting on the basis of propensity to distort.
In this study no clear systematic relations were observed between the log shape (ie ovality, pith
eccentricity, arc), batten compression wood content, and distortion in the form of bow and spring.
Although several logs with abrupt changes in curvature were noted to yield battens which bowed
or sprung considerably (Figure 2), some relatively straight round logs contained high levels of
compression wood, or yielded timber which distorted on drying. The high distortion of timber coming
from these straight logs was attributed to the consistent imbalance of compression wood that could
occur from one face or side of the board along its entire length (Figure 3).
Figure 2. Scanned image of logs with abrupt changes in curvature which yielded distorted timber
105
As expected, severe compression wood was found to be associated with stem correction, particularly
leader loss. Compression wood is denser than normal wood, and it was found that its presence
undermined the quite good relation observed between stiffness and density for normal wood.
Figure 3. Distorted batten, originating from relatively straight log. (compression wood shown marked)
EFFECT OF KNOTS AND DENSITY ON STIFFNESS

A key feature of British grown Sitka spruce often noted is the presence low stiffness material
near to the base of the tree. In this study, since the orientation of the battens was known, and kept
constant during machine grading, stiffness profiles for each tree could be established (Figure 4).
This feature was noted to be highly variable both between trees and stands. The effect is not attributed
to any outwardly measurable feature on boards (such as knots, low density, slope of grain etc),
but to high microfibrill angle. The feature is grade determining for many battens, and this has
a profound effect on the efficacy of any potential sorting system.
10
6
TREE HEIGHT (m)
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
IP (STIFFNESS)
Figure 4. Batten stiffness profiles showing low stiffness at the base of trees.
106
As expected from the batten stiffness profiles, log variables associated with batten height (eg taper)
were found significant. Figure 5 (below) shows that there is a reasonable relationship between density
and stiffness for upper logs, but that this relation is apparently not present at all for the butt logs. For
both small clear samples of normal wood and battens from upper logs, stiffness is was found to be
directly proportional to density. Spiral grain angle was not found to be a significant variable.
12000
Upper logs
Butt logs
10000
Upper logs Linear (Upper logs)
y = 17.037x + 600.49
R2 = 0.2094 Linear (Butt logs)
8000 Linear (All battens)

All Battens
y = 7.9388x + 3475.8
R2 = 0.0421
Ecb (N/mm2)
6000
Butt logs
y = 1.4751x + 5018.4
R2 = 0.0021
4000
2000
0
200 250 300 350 400 450 500 550 600
Density (kg/m3)
Figure 5. Stiffness plotted against density for subgroups of butt logs and upper logs.
Figure 6 below shows the observed non-linear relation between KAR and stiffness for a set of 97 x
45mm timber.
16000
14000
12000
10000
E (N/mm2)
8000
6000
4000
2
y = -9713.9x - 654.41x + 10308
2
R = 0.2215
2000
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
KAR
Figure 6. Relation observed between KAR and stiffness.
107
DISCUSSION
Good knowledge has been gained of the numerical relations between the variables affecting
the performance of British grown Sitka spruce. The presence of low stiffness material near to the base
of the trees was noted to be highly variable both between trees and stands, and the feature grade
determining for many battens. This effect can occlude relations observed between other variables.
Clearly extremes of knot content, slope of grain and log shape can be used as sorting criteria
to improve quality. For segregation of timber into different quality groups, knowledge of the effect
of combined variables is required. The necessary strictures of the grading system can also make certain
forms of sorting un-worthwhile, or counter-productive. By selecting out better quality logs or battens
the average quality of the remainder may fall. The database amassed by BRE in collaboration
with Forest Research can also be used to develop predictive silvicultural models. From the study of
machine grader rejects, severe cross grain associated with either large knot groups (Figure 7) or log
curvature was noted to be highly significant. However determining descriptive variables for this
feature is problematic.
Figure 7. Machine grader rejects with severe cross grain.

CONCLUSIONS
It was found that the presence of low stiffness clear wood at the base of the trees studied, together with
the natural variability between individual trees and stands, has a profound effect on the ability to sort
timber on the basis of industrially measurable variables such as density or knot content. Practical
difficulties with the use of slope of grain measurements to determine batten twist were also noted.
Relatively straight, round logs were observed to contain compression wood and hence yield timber
which distorted on drying.
ACKNOWLEDGEMENTS
This project work was funded by the EU, DTI, Forestry Commission, Scottish Enterprise
and BRE Trust. Extensive assistance was provided by colleagues at BRE and Forest Research. Log
scanning and timber processing was carried out at Adam Wilsons and Sons sawmill, Troone.
REFERENCES
1. Balodis V(1972): Influence of grain angle on twist. Wood Science vol 5 no.1 p44–50
2. Desch H. E. and Dinwoodie J.M. (1996): Timber – structure, properties, conversion and use.
MacMillan Press, London
3. Stevens W.C. (1961): Drying with and without restraint. Forest Products Journal Aug p348-356
4. Warensjo M. (2003): Compression wood in Scots pine and Norway spruce – Distribution in relation
to external geometry and the impact on dimensional stability in sawn wood. Doctoral thesis Swedish
University of Agricultural Sciences. Umea.
108
MC based or time based drying schedules?

Calculation of the influence on final MC variation
Jarl-Gunnar Salin
SP Technical Research Institute of Sweden
P.O. Box 5609, SE-114 86 Stockholm, Sweden
jarlgunnar.salin@sp.se
Keywords: Drying schedule, MC distribution, simulation
ABSTRACT
The quality of dried softwood timber depends of course on the drying process performance and in batch kilns especially on
the drying schedule. In the Nordic countries the drying schedule is almost always a time based schedule, i.e. the target
climate is defined as a function of drying time. The basic idea is that the average properties of each batch (same wood
species and board dimension) are almost constant and when a suitable drying schedule has been found, then it can be used
repeatedly without changes. In continental Europe the target climate is very often defined as a function of the actual
average MC of the batch. The basic idea here is to adapt the schedule according to the real properties of each batch.
This is a correct approach, but the problem is to measure the average MC accurately during the process. This is normally
done by pairs of electrodes driven into a limited number of boards for electric resistance measurements. Each board has
however individual properties and the measured MC differs normally from the average MC of the batch. In addition
the MC value based on the electric resistance is not fully reliable. This introduces a stochastic element in the performance
of the kiln control system. The aim of this paper is to analyse the accuracy of the MC based schedule control and compare
it to the time based schedule approach. The analysis is performed using a simulation program for the drying process.
Virtual boards with different properties (initial MC, heartwood content, density and ‘dryability’) are generated using
a random number generator. It has been taken into account that the properties mentioned are not statistically independent.
In this way a batch of timber is generated and the drying of each board is simulated in order to get the individual final MC
and further the MC standard deviation for the whole batch. The MC based drying schedule is applied according to
the calculated MC of a few randomly selected boards. The influence on the final MC variation, by the number of boards
selected for MC estimation, can thus be determined, as well as the influence of some other features. The results are then
compared to the corresponding results for a similar kiln using a time based drying schedule.
INTRODUCTION
Using simulation models for the prediction of the outcome from softwood drying processes has been
a standard procedure for a long time already. The basic method is to predict the average result
for the whole load in a batch kiln as a function of timber dimension, drying schedule etc.
This approach does not give any value for the spread in the results, such as the final MC standard
deviation. This limitation can be solved by performing simulations for each piece of a large group
of boards, where each board is given different properties according to the real variation in practice.
This method has been described and used in (Salin, 2002). The properties varied are: initial MC,
density, heartwood content and “dryability” (diffusion coefficient and related parameters).
These variables are considered the most important for softwood drying processes and also sufficient
for a reasonably reliable calculation. It should be observed that these variables are not independent
of each other. This is encountered for in the method described in (Salin, 2002). In this way a batch
of “virtual” boards is created and drying of each of these is simulated according to different drying
schedule types and approaches.
MC BASED DRYING SCHEDULE
An MC based drying schedule means that the climate in the kiln is a function of the measured
(estimated) average MC in the batch. The average MC is in practice based on the electric resistance
measured by pairs of electrodes driven into a number of boards in the batch. There are several ways
to do this by measuring the resistance at different depths, for instance close to the surface and/or at the
depth believed to represent the average MC of the board etc. Such solutions can improve the reliability
of the measurement. In the following theoretical calculations it is however assumed that
the MC measurement is 100% exact all the way for each board selected.
109
The example used in all the following simulations is characterized as follows. Scots pine (Pinus
silvestris) 50 x 150 with an average initial MC of about 73 % and density 430 kg/m3 (abs. dry/green
volume) is dried to either 16 % or 12 % in a batch kiln. The drying schedule used has a constant dry
bulb temperature of 80oC and a decreasing wet bulb temperature and the air velocity between board
layers is 4 m/s. Initially a time based schedule was chosen that gives a rather rapid drying, but with
a stress development (for all boards) clearly below the threshold value for checking. This time based
schedule was converted to an equivalent MC based schedule by investigating the MC development
in a board with average properties when dried with the time based schedule. Due to the relatively high
temperature, rapid drying and no final conditioning, the spread in the final moisture content is high –
which makes comparisons easier.
The simulations are performed for batches of 200 virtual boards. This number represents a rather small
kiln but is chosen for computational reasons. The final MC distribution for these 200 boards should
however represent a rather good approximation of the distribution from a full size kiln. Consider
as an example that three virtual boards are randomly chosen for control of the MC based schedule
(resistance electrodes). A typical final MC distribution for one single batch is shown in Fig.1
for a 16% MC target and in Fig.2 for a 12 % target.
0,3 0,3
0,25 0,25
0,2 0,2
0,15 0,15
0,1 0,1
0,05 0,05
0 0
8
8
10
12
14
16
18
20
22
24
10
12
14
16
18
20
22
24
Final MC, % Final MC, %
Figure 1. Typical final MC distribution for a Figure 2. Typical final MC distribution for a
single batch of 200 virtual boards. Target 16 % single batch of 200 virtual boards. Target 12 %
determined from 3 boards. determined from 3 boards.
0,2 0,25
0,18
0,16 0,2
0,14
0,12 0,15
0,1
0,08 0,1
0,06
0,04 0,05
0,02
0 0
7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37
7
9
11
13
15
17
19
21
23
25
27
29
31
33
35
37
Final MC, % Final MC, %
Figure 3. Final MC distribution for 17 batches Figure 4. Final MC distribution for 17 batches
of 200 boards each. Target MC 16 % deter- of 200 boards each. Target MC 12 % deter-
mined for each batch from 3 boards. mined for each batch from 3 boards.
It is clearly seen in both Figs.1 and 2 that the average MC in the batch is higher than the
corresponding target. (Actually these two distributions are taken from the same simulation, where the
16 % target case represents an intermediate result.) This shows that the three boards selected for
control, dried faster than the average of the batch. Now if several batches are dried in this way, then
some batches are dried too long and some for a too short time. When these batches are combined the
overall spread in the final MC will increase. Such a result is seen in Figs.3 and 4 for 17 batches (in
total 3400 boards) for MC targets 16 and 12 % respectively. If more than three boards are used for the
110
estimation
of the average MC in each batch then a narrower overall distribution will of course be the result.
The final MC standard deviation calculated for the total number of boards from 17 batches with
different number of control boards is presented in Table 1.
Table 1. Final MC standard deviation when several batches are combined
Number of control boards Target MC 16 % Target MC 12 %

[%] [%]
1 2,67 1,87
3 2,53 1,72
5 2,52 1,64
10 2,33 1,58
All 2,12 1,48
5, highest and lowest value neglected 2,41 1,57
10, highest and lowest value neglected 2,27 1,55
A perhaps surprising result seen in Table 1 is the relatively small influence the number of control
boards has on the standard deviation. An analysis shows that the standard deviation is a combination
of two parts; the (rather constant) spread in the final MC in a single batch and the error in the
estimated MC based on the control boards. It turns out that the first part is dominating. It should be
noticed that the values in Table 1 are rather high, which is due to a rapid drying process, without a
conditioning phase.
Although the total number of boards in the simulations is high, a random scattering of the results
in Table 1 have to be taken into account. A theoretical estimation of the final MC standard deviation
– with some simplifying assumptions – has been performed and is presented in Fig. 5.
3,5
Final MC standard deviation, %
3
Target 16 %
2,5
1,5
Target 12 % Calculation
1 Theory
Asymptote
Calculation
0,5 Theory
Asymptote
0
0 4 8 12 16
Number of control boards
Figure 5. Final MC standard deviation as a function of the number of control boards.

Fig. 5 shows clearly that the number of control boards does not influence the result very much.
It is also seen that omitting the highest and lowest measured MC values does not have a significant
impact on the standard deviation. This is true in this theoretical case – in practice the main reason
is to remove clearly erroneous, unsuccessful measurements. Due to the risk of such measurement
failures, the number of control boards should in practice not be too small, perhaps not less than five.
111
It should be taken into account that the control board measurements were assumed to be absolutely
exact in the simulations above. In reality the measurements are far from exact. In softwood boards
there are frequently both sapwood and heartwood present in the same board, with different MCs most
of the time during the drying process. There is further a steep moisture gradient in the board during
drying and it thus very important that the electrodes in a resistance measurement are driven
to the correct depth. If it is assumed that the standard error in a single measurement is 2 %-units
(Forsén, Tarvainen 2000) then the theoretical curves in Fig. 5 may be extended with a calculation
of the influence on the final MC standard deviation. The result is seen in Fig. 6.
3,5
Final MC standard deviation, %
Target 16 %
3
2,5
1,5
Target 12 %
1 Exact
2 % error
0,5 Exact
2 % error
0
0 4 8 12 16
Number of control boards
Figure 6. Influence of control board MC measurement error on the final MC standard deviation.
The standard deviation seen in Fig.6 has of course increased, but not so much, i.e. the spread caused
by the variation in board properties is still dominating.
TIME BASED DRYING SCHEDULE
The time based drying schedule is based upon the assumption that each batch of boards in the kiln has
the same distribution of properties and if the optimal drying procedure is found, then an optimal result
will be obtained for each batch. If this really is the case then a level slightly below the asymptotes in
Fig.5 would be achieved. This is however not a fair comparison as the main objective with the MC
based drying schedule is to adjust the schedule according to the real MC development for each batch.
A more reasonable procedure is to analyze how deviations from the assumption that each batch
is equal, affects the final MC standard deviation and compare this to the results for a MC based
schedule. An MC based schedule will in the long run produce an average MC close to the target
but the standard deviation depends on the ability of the control boards to estimate the real MC.
For time based schedules the opposite situation is found, i.e. the standard deviation is fairly constant
but the average value will deviate from the target as the changed properties of a batch are not taken
into account. A comparison between schedule types is thus not very easy. One possibility, which is
used here, is to instead of the standard deviation (which is based on the deviation from the average
value) base the calculation on the deviation from the target value. For the MC based schedule the
result will in principle be the same, but for the time based schedule the systematic deviation from
the target will be included and affects the result.
Three different changes in the batch properties are considered here; a pre-dried batch, a 1oC error in
the wet bulb depression and a 1 mm error in board thickness. In the first case it is assumed that
the (stickered) batch has been standing outdoors for a longer time than normal before loading
into the kiln. Boards with a high initial MC (much sapwood) will dry more during this period than
boards with a low initial MC and this feature is included in the simulation. The average decrease
of the initial MC for the whole batch is 8,0 %-units in this example. The second case was obtained
112
by decreasing the dry bulb temperature 1oC from the target level, simulating an error
in the temperature measurement. Finally in the third case, it is assumed that the green board thickness
has increased 1 mm without a change of the drying schedule. The result is seen in Table 2.
Table 2. Influence of a change in batch properties on the spread in the final MC with a time based schedule
Final MC target 16 % Final MC target 12 %

[%] [%]
Ideal case 2,05 1,48
Pre-dried batch 1,92 1,43
Wet bulb depression error 2,38 1,60
Board thickness error 2,08 1,51
It is seen in Table 2 that the influence of changes in the batch properties or drying conditions
on the spread in the final MC (standard deviation in relation to the target) is rather small.
It is remarkable to notice that in the case of pre-drying the spread actually decreased. This is
obviously due to the calculated effect of relatively slow pre-drying at low temperature before the
kilning.
If this is true in reality is not known. It could be argued that the changes in batch properties studied
in Table 2 are small. On the other hand, at least in the Nordic countries, it is a normal procedure
to check the final MC with a hand-held MC meter before emptying the kiln. At that stage fatal errors
would be noticed and the MC corrected if possible. It should finally be mentioned that although 3400
boards are simulated for each entry in Table 2, there are still some random scattering in the results.
CONCLUSIONS
As described above the variation in board properties has a dominating influence on the final MC
standard deviation. This means that a surprisingly low number of control boards with an MC based
drying schedule is enough, i.e. the benefit of additional control boards gradually disappears
in comparison with the total uncertainty. For the same reason, moderate changes in batch properties
or drying procedure, together with a time based drying schedule, will not influence the spread
in relation to the target MC very much.
If the results are compared to the new standards, it is found that with a MC based drying schedule,
the average MC of a batch (200 boards) will occasionally fall outside the accepted range for 1, 3 and
5 control boards. The portion of individual boards that fall within the ranges 12±3,6 % and 16±4,8 %
for the MC targets 12 % and 16 % respectively, fulfils the requirement (93,5 %) when several batches
are combined, but not for all single batches when the number of control boards is 1, 3 or 5. For time
based drying schedules the requirements are fulfilled. These results are based on the assumptions
that control board MCs are measured exactly and that the board property distribution is constant.
The results for the MC based drying schedule indicate thus that the number of control boards should
preferably be ten or more. This is illustrated by Figs.7 and 8.
0,25
Figure 7. Final MC distributions for a MC
0,2
MC based, 10 boards based drying schedule with 10 control boards
Time based
and for a time based schedule. Target final MC
0,15
is 16 %.
0,1
0,05
0
10
12
14
16
18
20
22
24
26
28
30
32
34
36
Final MC, %
113
Figure 8. Final MC distributions for a MC

0,3
MC based, 10 boards
based drying schedule with 10 control boards
0,25
Time based and for a time based schedule. Target final MC
0,2
is 12 %.
0,15
0,1
0,05
0
7
10
11
12
13
14
15
16
17
18
Final MC, %
As seen in Figures 7 and 8 the time based drying schedule produces a somewhat narrower distribution
in both diagrams, but the overall difference is rather small. The total number of boards in each case
is 3400 and in Fig.7 a few boards have a final MC of up to 36 %, but the portion of these extremes
is too small to be seen in the diagram.
Taking into account that especially the steep moisture gradient in a board during drying will increase
the MC measurement uncertainty – as seen in Figure 6 – it seems that a ten control board requirement
is a reasonable target. This indicates also that if a reliable average MC measurement method would be
available – that includes an essential part of the whole batch – then the MC based schedule would be
superior as it adapts to the real MC development of each individual batch.
On the other hand, in many laboratory tests where matched boards are dried in different batches
in order to compare different procedures, using a MC based schedule will introduce an unnecessary
random disturbance in the results. In such cases a time based schedule should be used.
REFERENCES
1. Forsén, H., Tarvainen, V., (2000): Accuracy and functionality of hand held wood moisture
content meters.
VTT Publications 420, Espoo, Finland. 79 p.
2. Salin, J-G., (2002): The timber final moisture content variation as a function of the natural
variation in wood properties and of the position in the kiln load. An evaluation using simulation
models.
Proceedings of the 4th COST Action E15 Workshop, Santiago de Compostela, Spain, May 30-
31.
114
Approach for an efficient Control of Grading Machine Settings
Markus K. Sandomeer1, Jochen Köhler2

1
ETH Zurich, Swiss Federal Institute of Technology, Zurich, Switzerland, sandomeer@ibk.baug.ethz.ch
2
ETH Zurich, Swiss Federal Institute of Technology, Zurich, Switzerland, jochen.koehler@ibk.baug.ethz.ch
Keywords: structural timber, machine grading, probabilistic modeling, Bayesian regression analysis
ABSTRACT
Machine grading systems operate according to similar principles; one or more strength related properties of the timber
to be graded are measured non-destructively by the machine and based on these measurements a population of un-graded
timber is subdivided into sub-populations of graded timber material. The grading acceptance criteria are formulated
in form of boundary values for the indicating property that have to be matched to qualify a piece of timber to a certain
grade. These boundaries are termed grading machine settings. The performance, i.e. the statistical characteristics
of the output of grading machines strongly depends on these settings, and in general very much attention is kept on how
to control them.
Currently applied procedures to control the grading machine settings are either machine controlled or output controlled,
however, both procedures do not facilitate the efficient use of information which is gained prior and during the operation
phase of the grading facility.
This presentation will highlight the existing methods for the control of grading machine settings according to the European
standard EN 14081. Moreover an alternative approach for the control of grading machine settings will be presented.
This approach is based on Bayesian regression analysis and allows a probabilistic modeling of timber material properties.
It can be seen as a combination of the machine control and output control approach. The alternative approach facilitates
the consistent consideration of new information gained during the whole grading process and will be summarized
in a coherent and implementable format. Possible benefits of its application in practice are discussed.
1 Background
Timber is by nature a very inhomogeneous building material. The mechanical and physical properties
of sawn timber vary on a large scale as a product of e.g. wood species and provenience. Regardless
of species and origin, timber properties diverge not only from tree to tree but also within a tree, over
the cross section and along the stem axis.
A pre-requisite for the use of timber in load-bearing constructions is that the relevant mechanical and
physical properties (strength, stiffness and density) are known and can be controlled to remain within
desirable limits. In comparison to other man-made structural materials, such as steel and concrete,
this cannot be achieved by changing the composition of the raw materials or by shifting some
of the environmental conditions. In order to use timber efficiently as a reliable structural material the
only way of obtaining controlled timber properties within required limits is by means of strength
grading procedures, either machine grading or visual grading.
Currently, there are many developments in the application of high-speed non-contacting scanning
systems which may offer the possibility of reduced grading costs, well-organized quality control
and more efficient utilization of structural timber. These highly innovative technical devices offer
a great amount of various information gathered during the grading process which is to date
not incorporated efficiently in the existing methods concerning the control of grading machine settings.
In this paper existing procedures for the control of grading machine settings according to the European
Standard EN14081 are discussed. Currently established methods can be either machine controlled
or output controlled, however, both procedures do not facilitate the efficient use of information which
is gathered during the operational phase of the grading facility.
Therefore, an alternative approach for the control of grading machine settings is introduced which
can be seen as a combination of the machine and output control procedure. It facilitates the consistent
consideration of new information gained prior and during the grading process. The approach
is summarized in a coherent and implementable format and possible benefits of its application
in practice are discussed.
115
2 Machine Grading of Structural Timber

Grading machines operate according to similar principles; one or more indicating properties
of the timber to be graded are measured by the machine and based on these measurements a sample
of ungraded timber is subdivided into subsamples of graded timber. However, the capability of indirect
measurements for predicting the strength can never be perfect and therefore, uncertainties are always
incorporated which have to be considered in the underlying grading model.
If more than one indicating property is measured by a certain type of grading machine, a combined
indicating property is identified, e.g. by regression analysis. In general grading acceptance criteria
are formulated for the indicating property in form of boundary values which have to be matched
to qualify a piece of timber to a certain grade. These boundaries are also termed grading machine
settings. The performance, i.e. the statistical characteristics of the output of grading machines strongly
depends on the settings, and in general very much attention is kept on how to control them. In Figure 1
observations on the indicating property and the property of interest are illustrated. Boundary values
of the indicating property (ip1 and ip2) subdivide the sample into sub-samples assigned to grade 1,
grade 2 and reject.
1
reject grade 1 grade 2
cumulative sample
property of interest
0.8
distribution
0.6
0.4
0.2
0
IP 1 IP 2
indicating property property of interest
Figure 1: left: Observations on the indicating property and the property of interest. The grading machine settings (ip1
and ip2) subdivide a sample into subsamples assigned to ‘grade 1’, ‘grade 2’ and ‘reject’.
right: Cumulative sample distribution of the property of interest of the graded subsamples.
Timber used for structural purposes is assigned to a specific strength class system (e.g. EN 338) which
constitutes the classification of timber samples based on requirements on their characteristic values
of the relevant material properties of interest. In EN 338 the required characteristic values for
the strength and density of the graded material are defined as the 5th-percentile values of the underlying
sample distribution. The characteristic modulus of elasticity (MOE) is specified by its mean value.
For the assignment of a timber sample to a certain grade by means of machine strength grading,
these mentioned characteristic values of the three relevant material properties (strength, stiffness
and density) have to be achieved. For this purpose machine grading becomes an important element
of quality control operations in the production line of structural timber elements.
3 Discussion of different Schemes for the Control of Grading Machine Settings

When considering the control of grading machine settings it is in general distinguished between two
main procedures, the machine controlled system and output controlled system. Focusing the European
Standard EN14081 the machine controlled system is represented by the so-called “cost matrix method”
(EN14081, part 2). A commonly applied method for output control systems is the “cumulative sum”
method (CUSUM) given in part 3 of the mentioned standard. An alternative approach will be
introduced which can be seen as a combination of the machine and output controlled systems.
116
3.1 Machine controlled System

The control of grading machine settings based on the machine controlled system was developed
in Europe almost 40 years ago and different methods have been used. A new approach was introduced
by Rouger (1996, 1997) in which the performance of the machine is compared with that of a ‘perfect’
machine capable of grading each specimen to its optimum grade. In the comparison of the assigned
grade and the optimum grade a cost analysis is used where weighting factors are applied to pieces
that are wrongly upgraded or downgraded. This method for machine control found its way into the
actual European standard EN 14081, part 2 and is denoted as the “cost matrix method”. Today, all
grading machines in Europe producing timber for the European market are operating with machine
control systems (Boström et al., 2000). Elaborate descriptions of the single steps of the machine
control procedure have been carried out earlier and can be found e.g. in EN 14081, part 2, Rouger
(1997), Bengtsson & Fonselius (2003), Köhler (2006) and Köhler & Steiger (2006).
In a machine controlled system all machines of the same type are said to be identical and any
combination of grading machine settings is applicable to all machines of that type and application
range. Therefore, the grade determining machine settings have to be derived by initial type testing
which requires an assessment for each machine type based on a very extensive experimental study.
In order to establish reliable settings for the different grades, sizes, species and origins often several
thousand pieces of timber are passed through the machine and thereafter tested to failure.
With this substantial assessment the grading machine settings are derived prior to the operation phase
of the machine and optimized to a representative ungraded timber sample which might be typical
for the daily use of the grading machine. Therefore the sampling and approving phases of the initial
derivation of grading machine settings claim the most efforts (Figure 2). Settings derived once,
are recorded in EN14081, part 4 and fixed for every machine of the same type and application range.
sampling phase approving phase operating phase
representation of testing of specimens grading with

timber population derivation of grading machine settings fixed settings
Figure 2: Different phases of the machine controlled grading procedure.
All grading machines together with their grade determining settings for the indicating properties
are enlisted in the European standard EN 14081, part 4, subject to the condition that they are approved
in accordance with the requirements given in the same standard. This list is updated approximately
every half a year, each time shortly after the Technical Committee CEN/TC124 has adjudicated upon
the approval of new grading machines or new application ranges of existing grading machines.
3.2 Output controlled System (CUSUM method)

Considering the control procedures of the CUSUM (cumulative sum) method given in EN14081, part 3
a limited number of specimens of each grade and dimension has to be selected a certain number
of times per working shift. The selected specimens have to be proof-loaded to check that the quality
of the timber is within desirable limits. If the timber survives the proof loading, it is deemed to have
acceptable strength properties. Successive values of a variable are compared with a defined target
or reference value, and the cumulative sum of deviations from this value is recorded in tabulation
and plotted on charts. If the accumulation reaches or exceeds a pre-determined decision interval,
this is taken to indicate that a change has occurred in the mean level of the variable and the process
is said to run out of control. The CUSUM approach assumes that the process is initially in control.
This can be ensured by some initial qualification or certification test which has to be passed prior
to the establishment of the continuous control program.
CUSUM operates with two separate types of charts, an attributes chart and a variables chart.
In the attributes chart the number of pieces that fails when proof loaded is recorded, while in
117
the variables chart the actual values of the modulus of elasticity are used and their mean value
is calculated for the assessment of the cumulative sum. Figure 3 shows an example of a graphical
combination of an attributes and variables chart. It can be observed, that the process runs out of control
after testing four test samples by means of proof loading.
Figure 3: Example of CUSUM values in a combined variables and attributes chart of the output control procedure.
It is observable that the values exceed the Y-boundary value for the first time after proof loading four test
samples and therefore the process is said to be out of control (Sandomeer & Köhler, 2007).
If the process runs out of control a different set of charts has to be applied until the process returns to
in control. When the process switches to out of control, in general, some check of the stress grading
process must be performed. If the process does not return to in control after a defined series
of test samples the production has to be stopped and investigation of the process and corrective action
may be undertaken before several nonconforming units are manufactured.
3.3 Alternative Approach – Combination of Machine and Output controlled

Systems
An alternative method for the control of grading machine settings was already proposed in Köhler
(2003) and Faber et al. (2004). This method facilitates the consideration of uncertainties throughout
the modeling of the grading process and probabilistic models of the material properties of graded
timber material can be derived. Furthermore the proposed model framework enables the (statistically)
efficient utilization of information gathered during the approval phase as well as during the running
phase of the grading machine; that is, the statistical characteristics of graded timber material properties
can be expressed and communicated by means of probability distribution functions, up to date
in regard to the currently available information. However, the method appeared rather theoretical for
practical application. Furthermore it is not clear how the statistical information about the graded timber
material properties should be communicated, since the derived probability distributions are not part
of a standard distribution family.
The basis of the presented method is a statistical representation of the relation of the material
properties of interest (poi) and the measurements of the grading machine (indicating property = ip)
by simple linear regression analysis as shown in the left illustration of Figure 4. The uncertainty
of the estimation can be consistently considered by the fact that the prediction of the material property
of interest is a normal distributed random variable. All parameters given in the equation (Figure 4 left)
are random variables representing the statistical uncertainties. Therefore all uncertainties
are incorporated into the grading model. Based on the relationship between the property of interest and
the indicating property e.g. two boundary values for the indicating property may be assessed to
subdivide the whole sample into different grades (Figure 4 middle). Beneath the x-axis the probability
density function of the indicating property is illustrated. With a given boundary value for the indicating
property and the given equation of the regression analysis, a conditional probability function for the
property of interest can be calculated without considering the particular sample values anymore
(Figure 4 right).
118
property of interest poi poi = α + β ip + ε poi poi
ip ip1 ip2 ip ip1 ip2 ip
indicating property
Figure 4: left: Regression analysis to indicate the relationship between the property of interest (poi) e.g. tension
strength and the indicating property (ip) measured by the grading machine.middle: Boundary values (ip1
and ip2) to subdivide the sample into different subsamples (grades). Calculation of the probability density
function of the indicating property.right: Based on the regression analysis and the probability density
function of the indicating property the conditional probability function of the property of interest can be
assessed.
The assessed conditional probability density function allows calculating a predictive cumulative
distribution function of the material property of interest. In Figure 5 the cumulative distribution
of the observed sample compared with the predictive cumulative distribution function is illustrated.
The 5th-percentile value of the tension strength predicted by the probabilistic model might appear quite
conservatively however, it should be noted that the model prediction represents the predictive value of
the 5th -percentile value containing all uncertainties due to model assumptions and due to lack of data.
Furthermore, variations in the timber material supply can cause considerable fluctuations
and uncertainties regarding the characteristic values estimated just on the basis of sample statistics.
The alternative approach based on Bayesian framework is capable to incorporate these uncertainties
and to facilitate the integration of new knowledge which might be obtained during the running phase
of the grading machine consistently.
model
= predictive distribution
cumulative distribution
sample
0.05
property of interest
Figure 5: Comparison of the cumulative distribution of the sample data and the probabilistic model. A shift in the 5th-
percentile value is observable (Sandomeer & Köhler, 2007).
Based on continuous observations of the indicating property without any additional costs the model
allows to formulate some decision criteria to indicate situations where variations in the input material
of the grading process are expected. If such a situation is indicated additional output control tests
(e.g. proof loading) should be initiated subsequently. Furthermore, the introduced method offers
the possibility to incorporate new information into the grading model continuously by updating
the Bayesian regression analysis and the resulting predictive distribution function of the material
property of interest.
119
DISCUSSION & CONCLUSIONS

The goal of the present paper is to give an overview of the existing procedures for the control
of grading machine settings given in the European Standard EN14081. The capability of the methods
to incorporate statistical uncertainties as well as model related uncertainties into the grading process is
of special interest. Furthermore, an alternative approach for the control of grading machine settings
is introduced which is capable to incorporate the different uncertainties into its probabilistic grading
model – an important pre-requisite for structural timber as a modern building material.
The machine controlled system given in EN14081, part 2 incorporates uncertainties into the grading
model only implicitly, i.e. by sampling different subsamples of different origins and by comparing
assigned grades with optimum grades. However, as long as the characteristic values of the derived
strength classes are assessed by means of sample statistics considering solely the underlying set of data
it is not possible to represent all uncertainties in a consistent matter.
The performance of the output control system by means of CUSUM control charts is observed to be
capable to detect aberrations in the quality of the material supply. Though very much attention is kept
on how to sample the test specimens randomly in order to manage the underlying statistical
uncertainties, the method is still based on sample statistics and therefore only capable to qualify shifts
of quality of the timber supply but not to quantify them.
The probabilistic approach shows to be a consistent combination of both methods described above.
It is based on Bayesian framework and therefore capable to incorporate all uncertainties into the
grading process and to facilitate the integration of new knowledge which might be obtained during
the running phase of the grading machine consistently. This backflow of information can be seen as
a crucial advantage in comparison to the established methods of the machine and output control
systems.
Furthermore, the introduced approach offers the chance to assess shifts in the material quality
by controlling the values of the indicating properties for the material properties of interest
continuously. Note, that this information is always available without additional costs. Assessed
aberrations of quality may be counteracted by means of updating the regression parameters
of the probabilistic grading model. In consequence of updating the regression model it is
a straightforward task to assign adjusted grading machine settings without additional substantial
test procedures.
The suggested probabilistic approach not only forms a very strong tool for the statistical quantification
of the material characteristics of timber but furthermore provides a consistent basis for quantifying
the efficiency of different quality control and grading procedures.
REFERENCES
Bengtsson C. & Fonselius M. (2003). Settings for strength grading machines – evaluation of the procedure according to prEN 14081, part 2. Proceedings of
the 36th Meeting, International Council for Research and Innovation in Building and Construction, Working Commission W18 – Timber Structures,
CIB-W18, Paper No. 36-5-1, Colorado, USA, 2003.
Boström L., Enjily V., Gaede G., Glos P., Holland C., Holmqvist C., Joyet P. (2000). Control of Timber Strength Grading Machines. SP REPORT 2000:11.
Faber M. H., Köhler J. and Sørensen, J. D. (2004). Probabilistic modeling of graded timber material properties. Journal of Structural Safety, 26(3), pp. 295-
309.
Köhler J. & Faber M. H. (2003). A probabilistic approach to cost optimal timber grading. Proceedings of the 36th Meeting, International Council for Research
and Innovation in Building and Construction, Working Commission W18 – Timber Structures, CIB-W18, Paper No. 36-5-2, Colorado, USA, 2003.
Köhler J. (2006). Reliability of Timber Structures. Dissertation ETH no. 16378. Swiss Federal Institute of Technology, Zurich, Switzerland.
Köhler J. & Steiger R. (2006). A discussion on the control of grading machine settings – Current Approach, Potential and Outlook. Proceedings of the 36th
Meeting, International Council for Research and Innovation in Building and Construction, Working Commission W18 – Timber Structures, CIB-W18,
Paper No.39-5-1, Florence, Italy, 2006.
Rouger F. (1996). Application of a modified statistical segmentation method to timber machine strength grading. Wood and Fibre Science. 28(4).
Rouger F. (1997). A new statistical method for the establishment of machine settings. Proceedings of the 30th Meeting, International Council for Research
and Innovation in Building and Construction, Working Commission W18 – Timber Structures, CIB-W18, Paper No. 30-17-1, Vancouver, Canada,
1997.
Sandomeer M. K., Köhler J., Linsenmann P. (2007). The efficient control of grading machine settings. Proceedings of the 40th Meeting, International Council
for Research and Innovation in Building and Construction, Working Commission W18 – Timber Structures, CIB-W18, Paper No. 40-5-2, Bled,
Slovenia, 2007.
EN 14081 parts 1-4: Timber Structures – Strength Graded Timber with rectangular Cross Section. Comité Européen de Normalisation, Brussels, Belgium,
2005.
EN 338: Structural Timber – Strength Classes. Comité Européen de Normalisation, Brussels, Belgium, 2003.
120
Bending strength and stiffness of aspen sawn timber

Harald Säll1, Bo Källsner2 and Anders Olsson3
1
Växjö University, School of Technology and Design, 351 95 Växjö, Sweden, harald.sall@vxu.se
2
Växjö University, School of Technology and Design, 351 95 Växjö, Sweden, and
SP, Swedish National Testing and Research Institute, Box 5609, 114 86 Stockholm, Sweden bo.kallsner@sp.se
3
Växjö University, School of Technology and Design, 351 95 Växjö, Sweden, anders.olsson@vxu.se
Keywords: Aspen, bending strength, modulus of elasticity, strength grading, structural timber
ABSTRACT
In order to investigate the possibility of using sawn timber of European aspen (Populus tremula L.) as a structural
material, 150 pieces of the dimension 45 x 120 mm2 were selected from a sawmill in the south of Sweden. The material was
visually strength graded using the Nordic standard INSTA 142 and the British standard BS 4978-1996, respectively.
The timber pieces were also subjected to flat-wise bending in a Cook-Bolinder strength grading machine. Finally all
specimens were tested in edge-wise bending according to the European testing standard EN 408. Global and local moduli
of elasticity as well as bending strength were determined.
The results indicate that the visual grading rules, intended for use on Norway spruce and Scots pine, seem to work fairly
well on European aspen. However, grading in a flat-wise bending machine like the Cook-Bolinder seems to give a lower
correlation between stiffness and strength for aspen timber than generally found for coniferous. Further, the correlation
between density and bending strength was found to be very low for the aspen timber tested. It was also noticed that
the aspen timber has a slightly higher bending strength and modulus of elasticity than is the case for normal qualities
of Norway spruce.
The study indicates that sawn timber of European aspen grown in southern Sweden can be visually graded and used
as structural material.
INTRODUCTON
Background
European aspen (Populus tremula L.) is one of the most widely spread species in the world, with
a natural range stretching from the Arctic Circle in Scandinavia to north Africa, and from Britain
across most of Europe and north Asia to China and Japan. In Sweden aspen is the second most
common hardwood species forming 2 % of the total wood supply. Only very little of the Swedish
aspen supply goes to the sawmill industry for processing and since there are no Swedish grading rules
and design values for aspen timber nothing of the timber is used for structural purposes.
Purpose
The purpose of this study was to investigate if sawn aspen timber originating from Sweden is proper to
use as a future structural material, to find out the possibilities of using visual and machine grading
and to determine some of the mechanical properties of the material in bending.
MATERIAL AND METHOD

Selection of boards
The material for the study, consisting of European aspen boards of the length 3000 mm, was selected
from the Werner Träförädling sawmill in Vimmerby situated in the south of Sweden. After the sawing,
the boards were kiln-dried to a moisture content (MC) of 7 %. A selection, in principle at random,
of 150 boards was carried out after this stage. The boards were finally planed to the nominal cross
sectional dimension 45 x 120 mm2.
121
Visual grading
After the material had been delivered to the laboratory the boards were visually graded according to
the Nordic standard INSTA 142 and the British standard BS 4978-1996. In the Nordic standard
the timber is graded into four classes T0, T1, T2 and T3. These classes are placed in the European
strength classes C14, C18, C24 and C30 respectively. At the visual grading of the boards
no consideration with respect to distortions were taken. In this paper only the results of the grading
according to the Nordic T-rules will be presented.
Machine grading
Prior to testing, all boards were passed through a Cook-Bolinder strength-grading machine
for recording of the flat-wise modulus of elasticity (MOE) distribution along each board. In order
to reduce the effects of vibrations in the machine, the boards were passed through the machine
at the low speed of 40 m/min.
Testing of boards in bending
The test arrangement used, see Figure 1, was in agreement with European testing standard EN 408
with two point loads acting in the third points. The distance between the supports was 18h where h is
the board depth. According to testing standard EN 384 the worst defect possible to test was placed
in the centre between the point loads and located randomly with regard to the compression and tension
side of the board. The position of the worst defect was determined based on visual inspection
of the boards.
In the testing standard it is stipulated that the MOE, here referred to as Elocal, shall be determined from
the curvature of the centre of the test piece over a gauge length of 5h. In this case the vertical
displacement was measured at the centre of the tension flange (point D relative to C and E).
In addition, the global MOE (Eglobal) was determined by measuring the mid-span vertical displacement
(point F) relative to the supports A and B.
After the MOE-values had been determined, the boards were tested in bending to failure.
For determination of moisture content ω and density ρ small specimens were cut out close to the cross
sections where failure occurred in the boards. In this paper we will only report the dry density ρ0,ω
of the wood material.
In the testing standard EN 408 it is stated that the test pieces shall be conditioned at 20 °C and 65 %
relative humidity. For softwoods this corresponds to a MC of 12 %. It should be noted that the timber
in the present investigation had a MC of about 7 % at testing.
F/2 F/2
F
C D E
A B
5h
6h 6h 6h
Figure 1. Test arrangement for determination of MOE and bending strength.

Basic statistics of measured parameters
Basic statistics of measured parameters, presented in Table 1, include minimum and maximum values,
arithmetic means, standard deviation (SD) and coefficient of variation (COV). In Table 2
the correlation matrix with respect to some of the parameters is given.
122
Table 1. Measured parameters. The values of the last column have no units.
Units Min. Max. Mean SD COV
Moisture content ω % 6.2 9.0 7.0 0.005 0.068
Dry density ρ0,7 kg/m3 325 538 434 42 0.097
Elocal MPa 7080 19540 13770 1990 0.145
Eglobal MPa 7400 16160 11930 1590 0.134
Ecook MPa 5110 14320 10500 1670 0.159
Bending strength fm MPa 16.0 94.3 55.1 13.8 0.25
Table 2. Correlation matrix of measured parameters.

r ρ0,7 Elocal Eglobal Ecook fm
ρ0,7 1 0.41 0.46 0.40 0.18
Elocal 0.41 1 0.84 0.61 0.50
Eglobal 0.46 0.84 1 0.73 0.49
Ecook 0.40 0.61 0.73 1 0.55
fm 0.18 0.50 0.49 0.55 1
Bending strength of ungraded boards

In Figure 2 the cumulative frequency of all bending strength data is shown together with a fitted
normal distribution function. The MC of the boards was, as already mentioned, about 7 % at the time
of testing. This low MC has probably caused that the compression strength of the wood material has
become somewhat higher than normal. The influence of the low MC on the tension strength
is probably fairly small. From this it can be concluded that the test specimens with dominating
compression failure, i.e. specimens with high bending strength, probably have obtained somewhat
too high values while test specimens with dominating tension failure, i.e. specimens with low bending
strength, are almost unaffected by the low MC.
1.0
Cumulative frequency
0.8
0.6
0.4
0.2
0.0
0 20 40 60 80 100
fm [MPa]
Figure 2. Empirical and fitted distribution functions of bending strength data.

The measured bending strength and density values are in Table 3 compared with results from
a Norwegian investigation of aspen timber (Fjærtoft et al 1998) and a Swedish investigation
of Norway spruce (Picea abies L.) timber (Johansson et al 1992). The dimension of the aspen timber
was for most boards 48 x 148 mm2 while the dimension of the Swedish spruce timber was 58 x 120
mm. The results presented in Table 3 show that the bending strength of Swedish grown aspen
is in good agreement with the one of Norwegian grown aspen. It can also be seen that the bending
strength of the aspen timber in both investigations is higher than normally found for Swedish grown
spruce. The density values of the aspen timber in Table 3 have been adjusted to ρ0,12 values (mass
123
determined at 0 % MC and volume at 12 % MC) assuming a linear volumetric shrinkage of 13.5 %

for the wood material between 28 % MC and 0 % MC.
Table 3. Comparison of fm, ρ0,12 and Elocal found in different studies.
Timber species Bending strength fm Dry density ρ0,12 Modulus of elasticity Elocal
Mean SD COV Mean SD COV Mean SD COV
[MPa] [MPa] [kg/m3] [kg/m3] [MPa] [MPa]
Aspen (Sweden) * 55.1 13.8 0.25 423 41 0.10 13770 1990 0.15
Aspen (Norway) * 53.0 12.5 0.24 440 37 0.08 12800 1800 0.14
Spruce (Sweden) ** 45.9 12.0 0.26 416 - 0.10 12790 0.21
* Measured dry density values have been adjusted to ρ0,12.
** Grade determining defect located on tension side of board.
Bending strength of visually graded boards

The results of the visual grading of the aspen timber can be seen in Figure 3 where the empirical
distribution functions of bending strength are shown for the material belonging to the different Nordic
T-classes.
1.0
Cumulative frequency
0.8 T0
T1
0.6
T2
0.4 T3
0.2
0.0
0 20 40 60 80 100
fm [MPa]
Figure 3. Empirical distribution functions of bending strength for timber graded into the Nordic T-classes.
The yield of the visual grading of the aspen timber into different classes and the bending strength
of these classes are given in Table 4. To the right in the same table some results are presented from
an investigation of Swedish grown spruce timber (Johansson et al 1992). The mean bending strength
of the different classes is higher for the aspen timber than for the spruce timber. For the two lower
classes T0 and T1 the coefficients of variation for the aspen timber seem to be slightly higher than
for the spruce timber.
Table 4. Yield and bending strength of visually graded aspen boards. Bending strength of visually graded spruce.
Swedish aspen Swedish spruce
Class n Min. [MPa] Max. [Mpa] Mean [Mpa] SD [Mpa] COV Mean [MPa] COV
T3 68 39.7 94.3 62.1 12.3 0.20 55.4 0.19
T2 39 19.3 78.0 54.3 10.8 0.20 47.0 0.23
T1 30 24.2 67.6 45.8 11.6 0.25 40.3 0.20
T0 11 16.0 62.0 42.0 14.1 0.34 34.0 0.29
Reject 2 26.4 55.4 40.9 - -
Of the 150 aspen boards tested only one board (see class T2 in Table 4) had a bending strength value
lower than the characteristic value of the class in question. In Table 5 the characteristic bending
strength fm,k of the different visual T-classes (for Nordic grown spruce and pine) are given
124
in the second column. In the last column the calculated characteristic bending strength of the tested
aspen boards fm,k,aspen are given. These later values represent 5-percentiles calculated at a confidence
level of 50 % assuming that the bending strength values within each T-class are of normal distribution.
The relatively high characteristic values found for the aspen timber indicate that visual grading
of aspen timber would be proper to use.
Table 5. Characteristic bending strength fm,k of T-graded Nordic spruce and pine timber. Estimated characteristic
bending strength fm,k,aspen of aspen timber tested.
Class fm,k fm,k,aspen
T3 30 41.9
T2 24 36.6
T1 18 26.7
T0 14 18.9
Prediction of bending strength

In order to study which parameters that can be used for prediction of the bending strength a multiple
linear regression analysis was carried out. The parameters used in the analysis were the dry density
ρ0,ω, the global modulus of elasticity Eglobal, the local modulus of elasticity Elocal and the modulus
of elasticity Ecook obtained by the Cook-Bolinder grading machine. The results of the analysis are,
for the aspen timber tested, shown in Table 6 in form of coefficient of determination (R2) and standard
error of estimate (SEE). As a comparison the corresponding values obtained in the investigation
of Swedish grown spruce (Johansson et al 1992) are given in the right part of the table. The R2-values
in Table 6, stating how large fraction of the variation in the bending strength (y-variable) that can be
explained by the different parameters (x-variables), are much lower for the aspen timber than for
the Swedish grown spruce timber. Note that the R2-value for the Cook-Bolinder grading machine is
0.55 for the spruce timber while it is only 0.30 for the aspen timber. None of the grading machines
used on the European market today will give such a low R2-value when predicting strength. It is also
notable that edge-wise bending (Eglobal and Elocal) gives lower R2-values than flat-wise bending (Ecook)
for the aspen timber which is in contrary to what is the case for spruce timber.
Table 6. Results of regression analysis for prediction of bending strength.
Swedish aspen Swedish spruce
2
Parameters R SEE [MPa] R2 SEE [MPa]
ρ0,ω 0.04 13.50 0.20 10.3
Eglobal 0.24 11.92 - -
Elocal 0.25 11.86 0.72 6.1
Ecook 0.30 11.54 0.55 7.8
Ecook and Elocal 0.34 11.15 - -
As an example on the weak relations between the measured parameters and the bending strength a plot
of the relation between Elocal and fm is shown in Figure 4.
In the investigation of Norwegian aspen timber (Fjærtoft et al 1998) R2-values are given in the range
between 0.15 and 0.25 for the parameter Elocal. This is even somewhat lower than the R2-value given
for Elocal in Table 6.
125
100
80
R2 = 0.2504
fm [MPa]
60
40
20
0
0 5000 10000 15000 20000 25000
Elocal [MPa]
Figure 4. Bending strength fm as function of Elocal.
CONCLUSIONS
The aspen timber tested has somewhat higher bending strength and MOE than is the case for normal
qualities of Norway spruce. The measured mechanical properties of the aspen timber are in fairly good
agreement with the ones obtained in a Norwegian study of aspen timber.
The Nordic visual grading rules, intended for use on Norway spruce and Scots pine, seems to work
fairly well on European aspen.
Machine grading in flat-wise bending using a Cook-Bolinder machine gives a lower correlation
between MOE and bending strength for the aspen timber tested than generally found for Norway
spruce and Scots pine timber. It is doubtful whether machine grading based on measurement of MOE
can be applied on aspen timber.
The correlation between density and bending strength was very low for the aspen timber tested. In
contrary to what is known for Norway spruce and Scots pine, the flat-wise MOE seems to be a better
parameter for prediction of bending strength of aspen than the edge-wise MOE.
The study indicates that sawn timber of European aspen grown in southern Sweden can be visually
graded and used as structural material.
ACKNOWLEDGEMENT
This investigation has been carried out by Anders Petersson M.Sc. and Tony Sjöö M.Sc. as a diploma
work (Petersson et al 2006) at the School of Technology and Design at Växjö University. This work
is part of a hardwood quality project financed by the Knowledge Foundation and the hardwood
sawmill industry of southern Sweden.
REFERENCES
1. Fjærtoft F., Eikenes B., Flæte P.O., Høibø O.A., (1998): Osp som konstruksjonsvirke –
sorteringskriterier og styrkeegenskaper (Sawn wood of aspen for construction purposes –
grading and strength properties). Report 2/98, Norwegian Forest Research Institute, Ås,
Norway.
2. Johansson C-J., Brundin J., Gruber R., (1992): Stress grading of Swedish and German timber.
A comparison of machine stress grading and three visual grading systems. SP Report 1992:23,
Swedish National Testing and Research Institute, Borås, Sweden.
3. Petersson A., Sjöö T., (2006): Böjhållfasthet och styvhet hos sågat virke av asp (Bending
strength and stiffness of aspen sawn timber). Diploma work no. TD002/2007, School
of Technology and Design, Växjö University, Sweden.
126
Visual grading in wood flooring sector in Greece

Michalis Skarvelis1, Antonios Papadopoulos2
1
N.AG.RE.F, Forest Research Inst. of Athens, Terma Alkmanos, 11528 Athens, GREECE, skmi@fria.gr
2
Tecnological Educational Institute of Kavala, Branch of Drama, Department of Forestry and Management of Natural
Environment, 1o km Dramas-Mikrochoriou, 66100 Drama, GREECE, antpap@teikav.edu.gr
Keywords: quality control, visual grading, wood flooring
ABSTRACT
The aim of this work was to investigate the classification systems used by wood flooring enterprises in Greece, using a
questionnaire. After 56 answers, the results showed that there are deviations in grades and in the criteria that are currently
used for the classification of wood flooring. There were reported very different grading systems, adopting 3 to more than
6 grading categories. The majority (40 %) was adopting 5 grading categories, but there were also enterprises using a few
main (3 or 4) and also a number of subcategories. Deviations were also observed among the producers who claim that
apply the same criteria and different trade names may be used for the same or similar qualities. All the above cause
difficulties in the product exchanges. Although it seems to be difficult, it is imperative need either the acceptance
of common national standards or the conformation with the European standards. Information and common education
of the evaluators is also a prerequisite, in order to achieve this goal.
INTRODUCTION
The application of classification systems used by solid wood conversing enterprises is based to a great
extend on visual characteristics and therefore it depends on subjective and personal factors.
These factors may include the clearness of the imposed criteria, the education and the experience
of the evaluators and the needs of the market. This causes difficulties in distribution and promotion
of the final products. The aim of this paper therefore, was to investigate the classification systems used
in one of the solid wood conversing sectors, as it is wood flooring enterprises. The work took place
in Greece during 2006, a year before the compulsory application of the EN 14342 standard (Wood
flooring - Characteristics, evaluation of conformity and marking), in order to discuss about the
problems that emerge from the lack of common standards for a specific product. A second target
was to point out problems that occur applying subjective and personal criteria, which in case of wood
may be too many (e.g. uniformity in colour, position and size of knots, slope of grain, appearance
of silver grain, etc.).
In Greece so far, there are no national standards for the classification of wood flooring, although
during the past some producers started themselves to apply a type of quality control. This attempt was
made in early 60’s and classified the wood flooring into three classes, namely A’, A-B’ and B-B’
(Kartasis, 1978). A second (semi-official) attempt was made recently, bearing in mind some CEN
standards, which also classified the wood flooring into three classes (YPEXODE, 2004). However,
it became clear very soon that this temporary solution was not applicable, since many producers and
merchants classified their products in to four or more grading categories; it is interesting to mention
that nowadays there are enterprises that classified their products up to eight categories.
Relevant examples exist also in other European countries, where it is very common to use three
grading categories for solid wood flooring, but there maybe found in the market up to six grading
categories in the same country. Some European countries were equipped with National Standards
many years ago and it is also common for some countries’ market to “borrow” Standards from
a different country, depending on (common) language or whether there was a large amount of trade
in the specific product.

The investigation of the classification systems took place during the period 2006 - 2007, using
a questionnaire which was distributed nationwide by personal interviews and by postal and electronic
means. A total of 56 questionnaires were collected, a significant high number for the Greek market.
127
The most effective way of data collection was found to be the personal interviews (85%). The recorded
data refer to: the property of the asked people, the size and the production of the enterprises and the
number of the quality classes used by them. Finally, a crucial question which included photographs
of five different grades of oak wood flooring, was incorporated in the questionnaire, in order
to estimate the effectiveness of the criteria that are currently used.
DISCUSSION
The property of the asked people is depicted in Figure 1. From this, it can be seen that the majority
of these were merchants (83.9%), followed by importers (69.6%) whereas a significant share were
producers (44.6%). At this point, it has to be mentioned that the asked people occupied more than one
property and therefore their opinion about the product and the current situation in the Greek wood
flooring market was more spherical.
100
Parquet producer
Importer
80
Merchant
Building Constructor
Technician
60
Other
%
40
44,6 69,6 83,9
20
21,4
0,0 0,0
0
Figure 1. Asked people’s property
Figure 2 shows, that the one third of the asked enterprises can be characterized as small ones (1 to 5
employees). This type of enterprise is a family type enterprise and is very common in Greece.
|It is interesting to mention that the small size is the main characteristic of the Greek wood industry
since 69% these employ less than 5 people (Kakaras, 1999). The medium enterprises (5 to 50
employees) were found to be more uniformly distributed. It is quite impressive that the percentage
of the big enterprises (more than 50 employees) is very low. However, the number of employees of the
enterprises did not seem to affect their annual production, which shows a uniform distribution
as Figure 3 reveals. This is due to the fact that the wood flooring is not the only sector that the big
enterprises are being involved.
>50
31 - 50 9%
22%
1-5
35%
16 - 30
17%
6-15
17%
Figure 2. Number of employees in asked enterprises
128
As far as the raw material is concerned, it was found that the most popular wood species used by
Greek enterprises are oak and iroko. A total of thirteen wood species of the temperate zone were
recorded (oak, pine, cypress, pitch-pine, black locust, beech, walnut, chestnut, cherry, eucalyptus, ash,
cedar, larch) and twenty four tropical woods.
Most popular length of the wood flooring was found to be the 25 – 30 cm, whereas figures of 2 m and
even 6.5 m (in softwoods) were also mentioned. Width varied from 3.5 to 20 cm. A significant number
of enterprises (82.2%) produce wood flooring 22 mm in thickness, although figures of 18 mm (26.7%)
were also observed. This contradicts earlier findings where the common thickness was 18 and 20 mm
(Kartasis, 1978).
As depicted in Figure 4, there are significant deviations in quality classes that are currently used for the
classification of wood flooring. There were reported different grading systems, adopting 3 to more
than 6 classes. The majority (36.4%) was adopting 5 classes, but there were also enterprises using
3 (3.6%) or more than 6 classes (7.3%). From the data processing, it is also revealed that serious
deviations exist even among the enterprises which apply the same number of quality classes; in these
cases the significance of each criterion used varies, since the criteria are numerous (silver grain
existence, sap existence in lower face, sap existence in upper face, knots in lower face, knots in upper
face, pin knots in upper face, flat sawn, quarter sawn, slope of grain, diagonal grain, uniformity
in colour).
>100001 -2500
50001 - 100000 11% 16%
33%
2501 - 5000
2%
5001 - 10000
9%
10001 - 25000
25001 - 50000
13%
16%
Figure 3. Annual production (in m2)
It is worth to mention, that only one case was recorded where the enterprises which apply the same
number of quality classes, they used the same criteria.
5 classes
36,4%
4 classes
25,5%
6 classes
27,3%
3 classes >6 classes
3,6% 7,3%
Figure 4. Number of quality classes used by the enterprises
129
All the enterprises that classified their products in to higher classes applied mainly desirable
and aesthetic criteria; this is more noticeable when 5 or more classes were adopted. On the other hand,
when the products were classified in to lower classes, non desirable criteria were applied (knots, silver
grain, sap existence). In order to estimate the effectiveness of the criteria that are currently applied,
a crucial question which included photographs of five different grades of oak wood flooring,
was incorporated in the questionnaire. The enterprises asked to order them in classes according to their
own criteria and the data analysis showed that:
The adopted criteria were not applied completely by the enterprises. This was obvious since
the collected answers were not given by the director of the quality control sector. It is expected
therefore lower deviations in real life than those recorded by this research, but we have to point
out that personal decisions may affect the result since a lot of criteria could not be measured
exactly.
Even when the adopted criteria among the enterprises were similar, they were not applied with
the same magnitude.
It was found that some enterprises which claim to apply 3 or 4 classes they also apply a number
of subcategories for the classification. This causes difficulties in the market and increases
the number of classes to 5 or 6 and therefore leads to non comparable results.
It is impressive that for the classification of the five different grades of oak wood flooring,
35 different classification answers were recorded by the 56 asked enterprises.
Another problem which leads to non comparable results seems to be the name of the classes,
even if the applied criteria are the same. The name of the classes is usually given by some
physical characteristics of wood (e.g. “Straight grain”, “Sap”) or imposed by the needs
of the market (e.g. “Royal”, “Rustic”). However, there were cases where a simple trade name
is used (“AA”, “AB”) or a combination of the above (“Extra -Silver grain”).
It was found that the driving force for the classification systems that are currently in use,
is the needs of the market and therefore the promotion of the sales. This is true for both big and
small enterprises, as depicted in Figure 5, since there was found no deviations in the number
of classes used.
In some cases where criteria -although not the same - were similar, very similar results
were given. It was noticed mainly where 5 or 6 classes were used. The opposite was also
noticed: in some cases with very similar criteria, very different results were reported.
So, we believe that in many cases there is very varying approach on the same feature
and a need for common education of the evaluators is present, in order to minimize subjective
decisions.
Enterprises stated that they are forced to these different classification systems through market
conditions and their experience, but to a percentage of 22% they confess that they use them for sales
promotion purposes. This seems to be true for all enterprises: according to their production we named
them “Big enterprises” (> 50.000 m2/year) and “Small enterprises” (<50.000 m2/year).
With the exception of enterprises applying 5 classes (Fig. 5), where “small” enterprises are much
more, in all other cases there are no serious differentiations among small and big enterprises.
Finally, it is generally admitted to a percentage of 70% that the different classification systems cause
difficulties in distribution and promotion of the final products. Although it seems to be difficult,
it is imperative need either the acceptance of common national standards or the soonest conformation
with the European standards. Information and common education of the evaluators is also
a prerequisite, in order to achieve this goal.
130
16
12
0
Bigger enterprises
se
s
ss
e
Smaller enterprises
s
ss
e
cla
s
ss
e
cla
es
ss
3
cla
ss
4
cla
5
c la
6
>6
Figure 5. Number of classes used by “small” and “big” enterprises
(Production/consumption less or more than 50.000 m2/year respectively).
CONCLUSIONS
Investigating the classification systems used by wood flooring enterprises in Greece, using a
questionnaire, the results showed that there are deviations in grades and in the criteria that are
currently used for the classification of wood flooring. There were reported very different grading
systems, adopting 3 to more than 6 grading categories. The majority (40 %) was adopting 5 grading
categories, but there were also enterprises using a few main (3 or 4) and also a number of
subcategories. Deviations were also observed among the producers who claim that apply the same
criteria and different trade names may be used for the same or similar qualities. All the above
mentioned causes difficulties in the product exchanges. Although it seems to be difficult, it is
imperative need either the acceptance of common national standards or the conformation with the
European standards. Information and common education of the evaluators is also a prerequisite, in
order to achieve this goal.
REFERENCES
CEN. 2005. ΕΝ 14342: Wood flooring - Characteristics, evaluation of conformity and marking. CEN,
May 2005, p. 24.
Kartasis Ι. 1978. Xylina Dapeda (Parketa). Greek Min. of Agric., Forest Serv., Indiv. Edit. No. 41,
1978, p. 83.
YPEXODE. 2004. Temporary National Technical Standards: 01 Wooden nailed floors. Greek Min. for
the Envir., Physical planning and Public Works, Inst. of Construction Economy, Athens, 2004, p. 41.
Kakaras I. 1999. Furniture and wood construction industries in Larisa area. Existing situation and
perspectives. Proceedings of the International Conference on Wood Science, London, UK. Pp: 218-
224.
131
Improved log sorting combining X-ray and 3D scanning – a preliminary study

Johan Skog1,2, Johan Oja3
1
SP Technical Research Institute of Sweden, Wood Technology,
Skeria 2, SE-931 77 Skellefteå, Sweden, johan.skog@sp.se
2
Luleå University of Technology, Department of Wood Technology,
Skeria 3, SE-931 87 Skellefteå, Sweden
3
SP Technical Research Institute of Sweden, Wood Technology,
Skeria 2, SE-931 77 Skellefteå, Sweden, johan.oja@sp.se
Keywords: 3D scanning, automatic grading, quality sorting, sawlogs, X-ray scanning
ABSTRACT
Quality sorting of sawlogs is becoming more and more common. This is the result of increasing production of customer
specific products in combination with high raw material prices.
Today, log quality sorting is being based on either 3D or X-ray scanning techniques. Previous research has shown that
sorting accuracy is improved when using multivariate models to combine variables from both 3D and X-ray scanners.
There is however a potential of further improving the sorting if 3D and X-ray data are combined at an earlier stage; from
the measured 3D shape a better estimate of the X-ray path lengths through the log may be found, thus enabling
the calculation of a log density profile from the measured X-ray attenuation.
The development and evaluation of such a technique is the topic of current research at SP Trätek and Luleå University
of Technology in Skellefteå. Preliminary results show that the method is good at calculating large scale properties such as
heartwood content and heartwood and sapwood densities. When looking for smaller geometric objects, e. g., knot whorls,
extra care must be taken so that observational errors from the 3D scanner do not compromise the X-ray data.
Software simulating industrial X-ray scanner data from CT-scanned logs has also been developed. A very good agreement
was found between simulated data and actual data from an industrial installation. This underlines that such a simulation
tool is very valuable when developing algorithms for industrial X-ray scanners.
INTRODUCTION
Grounds for quality sorting of sawlogs
For a long period of time, sawmills have had a strong focus on productivity. The standard methods
of improving profits has been either making efforts to improve volume recovery or increasing
the volumes of sawlogs passing through the mill. Today, the competition for sawlogs is fiercer and raw
material costs are rising, comprising an increasingly large amount of the total costs of the mill.
Consequently, it is has become ever more lucrative for sawmills to focus on value recovery and quality
of sawn products rather than just outgoing volumes. One way of doing this is sawing products carrying
specific combinations of dimension and grade, better corresponding to customer demands and thus
yielding better value. In order to achieve this, keeping an efficient raw material use, it is important
to get it right from the beginning, identifying the right sawlogs for each product; hence quality sorting
of logs before sawing is becoming more and more common.
Quality sorting using optical 3D scanners
Quality sorting of sawlogs requires the ability to predict quality of sawn goods from log measure-
ments. Many sawmills already sort the timber based on length and diameter measures obtained from
optical three-dimensional (3D) surface scanners (Figure 1). These scanners also yield information
about taper, bumpiness and other variables that may be used to estimate the log quality (e.g. Grace
1994, Jäppinen and Nylinder 1997, Oja et al. 1999). Such predictions, using partial least squares (PLS)
modelling on 3D data in the quality sorting software Kvalitet On-Line (Anon. 2007b), has proven
successful and become widely spread in Sweden.
133
Figure 1. Schematic description of an optical 3D surface scanner (Dashner 1993).
Quality sorting using X-ray scanners

Still the quality information obtainable from 3D scanners is limited, since it is based solely on outer
shape properties of the log. An extensive evaluation (Grundberg et al. 1990) of different measurement
techniques showed that inner properties of logs are best measured using X-rays.
X-ray scanning by computed tomography (CT) is too slow for industrial applications but may be used
to obtain precise measurements of wood density (Lindgren 1991). The high quality images captured
have proven very useful for research purposes, e.g., the Swedish pine stem bank, a collection
of CT images gathered by Grönlund et al. (1995).
In order to improve speed, industrial X-ray scanners only use a limited number of fixed measurement
directions (e.g. Aune 1995, Grundberg and Grönlund 1995). Many authors have developed algorithms
analyzing images from such detectors, including calculation of knot structure (Pietikäinen 1996),
annual ring width (Wang et al. 1997), outer shape (Oja et al. 1998, Skatter 1998) and strength of sawn
products (Oja et al. 2005). Figure 2 shows the principle of a successful two-directional X-ray
LogScanner for Scots pine (Pinus sylvestris L.) developed by Grundberg and Grönlund (1995), being
in use at seven sawmills as of October 2007. Today the use of X-ray scanners is increasing and, likely,
more installations will follow.
Figure 2. Schematic description of the X-ray LogScanner developed by Grundberg and Grönlund (1995),
image from Oja et al. (1998).
134
Quality sorting combining X-ray and 3D scanners

Oja et al. (2004) made a comparison between the grading performances of X-ray and 3D scanning
techniques and also investigated possible benefits from using a PLS model combining parameters from
both methods. The study showed that 57% of sawn boards were correctly graded when using
3D scanning, 62% when using X-ray LogScanner and 66% when combining data from both scanning
methods. The highest possible result, with ideal log grading, was 81%. The study concludes that
the combination of 3D and X-ray scanning seems very promising and suggests that future studies
should focus on fully utilizing the possibilities of combining the two techniques. The combination
could also be expected to give better diameter measurements, joining the ability of the X-ray
LogScanner to handle varying bark thickness with the ability of the 3D scanner to handle irregularly
shaped cross-sections (Oja et al. 1998). Yet another reason for investigating the benefits of such
a combination is that a 3D scanner is already present at most mills installing an X-ray LogScanner,
and thus do not present any substantial extra investment costs.
By combining raw data from 3D and X-ray scanners at an earlier stage it would be possible to obtain
improved density profiles of the logs. A density profile of the log can be obtained if the LogScanner
signal is compensated for the varying travel path lengths of the individual photons through the wood
(Grundberg et al. 1990). The best travel path compensation is found from the real shape of the log,
which in practice is not known. Instead, algorithms developed should use the best shape information
available, namely the shape measured by the 3D scanner. The hypothesis is that such a combined
technique would lead to improved assessments of heartwood and sapwood densities and would allow
for better identification of, among other things, heartwood content, knot whorls, annual ring distance
and rot.
The aim of this study is consequently to develop algorithms combining X-ray and 3D data using travel
path compensation and to evaluate whether the methods developed have a potential of improving
quality sorting at sawmills.

Data collection
In order to enable the development of quality sorting algorithms based on 3D and X-ray data, it was an
essential first step to collect a well defined data set. A large sawmill in the north of Sweden, which had
recently installed a one-directional LogScanner, was chosen as host for the project and a total of 435
Scots pine sawlogs, originating from 13 different diameter classes in the range of 150 to 300 mm,
were picked out. Each log was individually ID marked and had its top heartwood diameter and annual
ring distance manually measured. The logs were then sent through the log sorting station, where data
from the RemaControl X-ray LogScanner and the MPM Engineering 3D surface scanner
were recorded.
Once all available log data had been collected, the logs were sawn using suitable 2X-patterns, forming
a total of 870 centre yield planks. The grades of the sawn planks with respect to knots were established
manually by the mill’s lumber grader as well as automatically using a FinScan Boardmaster
equipment. Eventually, the dried planks were sent to a finger jointing facility where their knot free
zones were determined.
Combining 3D and X-ray data using travel path compensation
The development of algorithms combining raw data from the 3D and X-ray scanners using travel path
length compensation and was performed using MatLab 7 (Anon. 2007c) and Visual Studio 2005
(Anon. 2007a).
Data were combined on cross-section level according to the principle shown in Figure 3. For each
X-ray cross-section, the corresponding 3D cross-section was located and inserted into a simulation
model of the X-ray scanner. Since both scanners are located along a common carrier line, right next to
each other, it was assumed that the rotational position of the logs did not change between the scanners.
135
Vibrations do however introduce an uncertainty in the exact horizontal and vertical position of each
cross-section. Thus the positions of the 3D cross-sections are being individually adjusted in order
to achieve best possible matching with X-ray cross-sections.
Once the 3D cross-section has been properly positioned in the simulation model, the travel path
lengths through the wood for photons hitting each detector pixel are being calculated. With knowledge
of the radiated X-ray spectrum as well as the absorption and travel length along each ray path,
the average density at each ray path is being calculated (cf. Grundberg et al. 1990) and arranged
together as a density profile of the cross-section.
Figure 3. Principle for density profile calculation from an X-ray (black) cross-section and a 3D (gray) cross-section.
The 3D cross-section is inserted into a simulation model of the X-ray LogScanner and the travel path
lengths through wood of photons hitting each detector pixel are calculated (gray). A density profile may
then be found by combination of measured X-ray absorption and calculated travel path length.
Simulating an X-ray LogScanner from CT data

In order to validate the travel path algorithms and the density profiles calculated, as well as to allow
for further algorithm development classifying knot whorls, rot and other internal artefacts, well defined
source data is essential. A data set suitable for this purpose is the Swedish pine stem bank, a collection
of more than 600 CT-scanned Scots pine logs. These logs are well defined regarding knot structure
and other quality parameters and CT images are available for every 10 mm of the log.
When developing the X-ray LogScanner, Grundberg and Grönlund (1995) wrote algorithms that could
simulate LogScanner data using CT images from the Swedish pine stem bank. The implementation
however had become obsolete due to computer platform change and due to higher resolutions now
being available for both CT images and LogScanners. Thus a new simulation code was written, better
suited for the demands of this project.
The new simulation code was verified by the X-ray scanning of a knot rich sawlog at the host sawmill.
The sawlog was then brought to the tomography lab at Luleå University of Technology in Skellefteå
where it was scanned with the same settings used for the Swedish pine stem bank. Rotational position
of the log was kept track of, so that it could be scanned at the same position in both equipments.
LogScanner data could then be simulated and compared to the data actually collected at the sawmill.
The access to the full set of CT, X-ray LogScanner and 3D data for the same log also allowed
for comparison between calculated density profile and actual density of the log.
136

Travel path compensation
The study presented in this report has focused on the development of algorithms that combine raw data
from 3D and X-ray scanners at an earlier stage in order to obtain improved density profiles of the logs.
A comparison between raw X-ray LogScanner data and travel path compensated data is shown
in Figure 4. A preliminary evaluation of the method show advantages as well as disadvantages
of performing calculations on travel path compensated data rather than raw data.
Great advantages include the possibility to compensate for log shape irregularities and that the method
enables direct reading of density from the images. This should make it easier to compare logs
of different sizes and facilitate the development of general algorithms. It should also be easier
to determine the heartwood border and the heartwood content and better measures of heartwood
and sapwood densities can be expected.
The main difficulty lies in the step of finding the right 3D shape for a given X-ray cross-section.
Firstly, there is some uncertainty in the identification of correct 3D cross-section and secondly
there is also an uncertainty in the exact position of the log at the X-ray scanner. Furthermore
the method transfers measurement errors from the 3D scanner into the X-ray data, e.g.,
an underestimate of the log diameter would cause an overestimate of the density profile of the cross-
section. For this reason, extra care must be taken when looking for small artefacts ranging over
a few slices, such as knot whorls, so that any localized measurement errors do not compromise the
data.
Figure 4. (a), (b): Raw LogScanner data, X-ray absorption depends on density as well as travel path. (a) shows the full
data range whereas (b) shows a zoomed data range, better exposing internal artefacts as well as the varying
absorption along the log due to varying diameter. (c): Travel path compensated LogScanner data showing
the density profile of the log. Uncertainty in the combination of 3D and X-ray data can be seen as bright
or dark radial lines at the edge of the log, where the relative uncertainty of the travel path is greatest.
Density calculations
Figure 5a presents a density profile of a single cross-section within the log as well as an average over
five neighbouring cross-sections. The average green density along a ray path in the sapwood area
of the log is 1.0 g/cm3, which corresponds quite well to the true green density measurements obtained
from the CT image (Figure 5b), ranging from 0.90 g/cm3 (inner sapwood) to 1.0 g/cm3 (outer
sapwood).
137
1.1
1
Average density [g/cm3]
0.9
0.8
0.7
0.6
60 80 100 120 140 160 180 200 220 240 260
Detector pixel
Figure 5. (a): Green density profile of a log cross-section, the dotted line shows a single cross-section and the
solid line shows an average over five neighbouring cross-sections. (b): CT image showing the corresponding
log cross-section. Dark pixels represent low density while bright pixels represent high density.
The average green density, ρa, along rays passing through the heartwood area of the log is around
0.83 g/cm3. This figure is a combination of heartwood, ρh, and sapwood, ρs, densities along the rays:
ρ a = h ⋅ ρ h + (1 − h ) ⋅ ρ s
The difficulty lies in finding a good estimate of the percentage of path being travelled through heart-
wood, h. The algorithm finding this percentage is yet to be completed but preliminary calculations may
be performed using estimated values. Looking at the density curve (Figure 5a), heartwood diameter
can be estimated to be around 45% of the total diameter (90 pixels vs. 200 pixels). Thus, green density
of heartwood close to the pith would be around 0.62 g/cm3. For a ray passing further out from
the centre of the log, e. g., having an h value of 33%, heartwood green density would be about
0.48 g/cm3. Measurements within the CT image reveal a true heartwood density ranging from
0.50 g/cm3 close to the sapwood up to 0.60 g/cm3 close to the pith.
Although being performed only at a single position of a single log, these preliminary calculations
suggest that green density values obtained by the method are not unreasonable. Further material
for testing may be obtained by CT-scanning small parts of industrially scanned logs or by simulations
from the stem bank (see below).
LogScanner simulations from CT
Figure 6 shows a comparison between industrially gathered and simulated X-ray LogScanner data,
the agreement between the images being very good. This strongly supports the stance that simulation
of LogScanner data from well defined CT images is a suitable method of obtaining source data for the
development of X-ray LogScanner algorithms.
Figure 6. LogScanner data from an industrial installation (left). LogScanner data simulated from CT images (right).
138
CONCLUSIONS
This preliminary study has concluded that the combination of X-ray and 3D data using travel length
compensation is a very promising technique for determining a density profile of the log. The technique
will also facilitate the development of general algorithms for characterization of inner properties
of sawlogs.
Preliminary green density values obtained for both heartwood and sapwood seem reasonable. How-
ever, more work need to be spent on development of rigorous methods for finding the heartwood
density and final algorithms must be tested on a larger data material.
It has also been concluded that X-ray LogScanner data simulated from CT images well correspond
to data from actual industrial installations. Even though being unaffected by vibrations and other
disturbances experienced in industrial environment, such simulations constitute a very good tool
for the development of X-ray LogScanner algorithms.
ACKNOWLEDGEMENTS
The work was financially supported by the SkeWood research programme through TräCentrum Norr
and through the Swedish Agency for Innovation Systems (VINNOVA).
REFERENCES
1. Anon., (2007a): Microsoft Corporation, Redmond, WA. http://www.microsoft.com/
2. Anon., (2007b): SP Trätek, Borås, Sweden. http://www.sp.se/en/areas/wood/
3. Anon., (2007c): The MathWorks Inc., Natick, MA. http://www.mathworks.com/
4. Aune J. E., (1995): An X-ray Log-Scanner for sawmills. In Lindgren, O. (ed.). Proceedings
from the 2nd international seminar/workshop on scanning technology and image processing on
wood. Luleå University of Technology, Technical Report 1995:22 T, pp. 52-64.
5. Dashner B., (1993): 3D Log scanning – the next generation. Proceedings of the 5th Int.
Conference on Scanning Technology and Process Control for the Wood Products Industry,
Atlanta, 25-27 Oct. 1993.
6. Grace L. A., (1994): Design and evaluation of an optical scanner based log grading and sorting
system for Scots pine (Pinus sylvestris L.) sawlogs. Dissertation. The Swedish University of
Agricultural Sciences, Uppsala, pp. 7-21.
7. Grundberg S., Grönlund A., Lindgren O., (1990): Noggrannhet vid detektering av stockars inre
kvalitet. TräteknikCentrum Rapport I 9005020. (In Swedish with English summary.)
8. Grundberg S., Grönlund A., (1995): The development of a LogScanner for Scots pine. In
Lindgren, O. (ed.). Proceedings from the 2nd international seminar/workshop on scanning
technology and image processing on wood. Luleå University of Technology, Technical Report
1995:22 T, pp. 39-50.
9. Grönlund, A., Björklund, L., Grundberg, S., Berggren, G., (1995): Manual för furustambank.
Luleå University of Technology, Teknisk rapport 1995:19 T, 25 pp. (In Swedish.)
10. Jäppinen A., Nylinder M., (1997): Automatic sorting of spruce (Picea abies (l.) Karst.) sawlogs
by grade. Holz als Roh- und Werkstoff 55: 301-305.
11. Lindgren L. O., (1991): Medical CAT-scanning: X-ray absorption coefficients, CT-numbers
and their relation to wood density. Wood Sci. Technol. 25: 341-349.
12. Oja J., Grundberg S., Grönlund A., (1998): Measuring the outer shape of Pinus sylvestris saw
logs with an X-ray LogScanner. Scand. J. For. Res. 13: 340-347.
13. Oja J., Broman O., Lindfors S. E., (1999): Projektrapport: Timmerinmätningsstöd. Teknisk
Rapport 15, Luleå University of Technology. (In Swedish with English summary.)
139
14. Oja J., Grundberg S., Fredriksson J., Berg P., (2004): Automatic Grading of Sawlogs: A
Comparison Between X-ray Scanning, Optical Three-dimensional Scanning and Combinations
of Both Methods. Scand. J. For. Res. 19: 89-95.
15. Oja J., Källsner B., Grundberg S., (2005): Predicting the strength of sawn wood products: A
comparison between X-ray scanning of logs and machine strength grading of lumber. Forest
Prod. J. 55(9): 55-60.
16. Pietikäinen M., (1996): Detection of knots in logs using x-ray imaging. Dissertation. Technical
Research Centre of Finland, Espoo. VTT Publications 266, 88 pp.
17. Skatter S., (1998): Determination of cross-sectional shape of softwood logs from three x-ray
projections using an elliptical model. Holz als Roh- und Werkstoff 56: 179-186.
18. Wang X., Hagman O., Grundberg S., (1997): Sorting Pulpwood by X-ray Scanning.
International Mechanical Pulping Conference, Stockholm, Sweden, pp. 395-399.
140
Development of twist on boards of Norway spruce during kiln drying

Aleš Straže*, Robert Kliger**, M. Johansson**, Ž. Gorišek*
*University of Ljubljana, Biotechnical Faculty, Rožna dolina, C. VIII/34, SI-1000 Ljubljana, Slovenia;
Email: ales.straze@bf.uni-lj.si; zeljko.gorisek@bf.uni-lj.si
**Chalmers University of Technology, Department of Structural Engineering, Steel and Timber Structures, 41296,
Göteborg, Sweden; Email: robert.kliger@chalmers.se; marie.johansson@chalmers.se
Keywords: grain angle, kiln drying, moisture content, shrinkage, warp
ABSTRACT
Development of twist of Norway spruce boards (Picea abies Karst.) during normal temperature kiln
drying was researched. Oriented tangential boards (18 × 80 × 800 mm) were sawed from diametrical
radial planks, with known axial shrinking, density, growth dynamics and spiral grain angle (SGA).
Unrestrained boards were later on kiln dried at normal temperature, to determine moisture content
(MC) and twist development and in 2 days drying interval. Twist was induced in most cases around
fibre saturation point, and raised proportionally with decrease of MC towards to the highest values
at the end of the process. Radial position of boards had the strongest influence on twist. End values of
twist amounted from 10 and 20 °/m close to the pith, and decreased to less than 4 °/m at 70 mm from
the centre of logs. SGA, generally decreasing in radial direction, significantly influenced the twist,
especially at low MC, bellow 16%. Tangential shrinkage had also pronounced impact on twist
variation, especially at MC above 20%, which is ascribed to the likely presence of tangible
MC gradient as a consequence of drying procedure.
INTRODUCTION
Warp, mainly twist, crook and bow, of sawn wood of Norway spruce (Picea abies Karst.) frequently
cause serious problems in wood construction industry. Twist often causes the most severe problems
(Perstorper et al. 1995, Forsberg 1997).
The two parameters most commonly associated with twist are distance from pith to the centre of the
cross-section of the board, which is an indirect expression of the annual-ring curvature, and spiral
grain angle (SGA). Significancy of these parameters was studied by some analytical models (Stevens
et al. 1960, Booker 2005, Bäckström et al. 2006), statistical surveys on data from stationary
conditioning experiments (Johansson et al. 2001, Warensjo et al. 2004) and at numerical modelling
(Ormarsson et al. 1998). In most cases ring curvature was the best predicting variable or had
the strongest impact on twist variation, followed by SGA. Recently, additional explanation of twist
of sawn wood was achieved by including of SGA gradient (Forsberg et al. 2001).
A lot of research was done to prevent twist of sawn wood during experimental and industrial drying.
Therefore different drying schedules were studied at normal temperature, varying drying time,
temperature and climatic conditions (Milota 2000). Improved shape stability of dried sawn wood was
achieved using additional steaming of boards, varying climatic conditions during drying or with
increasing of top loads on kiln stacks (Mackay 1973, Arganbright et al. 1978, Kliger et al. 2005,
Frühwald 2006). The impact of green moisture content variation and especially moisture content
gradient at dried boards is less investigated. Some research in this field was done by finite elements
modelling or with experimentation by static conditioning (Bäckström et al. 2006).
This research was designed to follow the development of twist in sawn wood of Norway spruce (Picea
abies Karst.) during unrestrained normal temperature kiln drying. The aim of the study is to determine
some material characteristics as growth rate, density, axial shrinking and spiral grain, to asses their
variation and to analyze the applicability of material data in twist modelling during drying process.
141

Ten Norway spruce trees (Picea abies Karst.), 50 to 70 years old, were cut down for the survey
purposes. Afterwards, two succesive logs, 0.4 and 0.8 m in length, were sawed off from breast height
of individual tree. Material characterization was carried out on shortest logs, wherefrom samples
for density determination, growth dynamic, axial shrinkages ( L, R, T) and spiral grain angle
(SGA) were prepared in radial direction of logs. Material data was determined using standard test
procedures (DIN 52182; DIN 52184), whereas SGA was determined by hand breaking of, 3 mm thick,
oriented and radially positioned lamellas, in longitudinal-tangential plane.
Diametrical radial planks, 60 mm thick, were sawed from 0.8 m long logs. Planks were ripped through
the pith and radially sawed into 18 mm thick, tangentially oriented lamellas, with known radial
position and material characteristics (Fig. 1). 170 lamellas were later on kiln dried, without restrain,
using common normal temperature drying schedule, with step raised temperature (from 67 to 80 °C)
and increasing drying gradient (from 3 to 4.3). During the drying, moisture content (MC) and twist
development were successively determined in 2 days interval, using 2 h breaks of the drying process.
We measured MC gravimetrically, by weighting of individual lamellas (EN 13183-1).
Twist on lamellas was measured by aluminium wedge, with 0.5 mm precision.
41%
23%
60
17%
15%
6%
A B CD E
60
18
800
800
Figure 1 Saving of sawn lamellas from a long log (A to E – growth ring decades with volume share)

Material data
In general, longitudinally aligned wood tissue significantly deviated from the tree axis. In all examined
trees the helix was left-handed in juvenile region with the maximum angle, from 2 to 6°, close
to the pith (Figure 2). As a rule, reduction of SGA with distance from the pith was present in all
examined trees, where three typical radial patterns were observed: a - grain direction may stay left
handed, b - grain direction tend to become almost straight or c - grain can change direction to right
hand. Determined SGA patterns caused great variability at the outer part of different stems.
These observations coincide well with results from other studies of Norway spruce (Danborg 1994,
Forsberg et al. 2001, Perstoper et al. 2001, Sepulveda 2001).
142
Total tangential shrinkage of this material is similar to what is reported in the literature. The overall
mean value is 8.97% (CV = 0.145), with significantly lower value, namely 8.54 (CV = 0.107), close to
the pith (p = 0.05) (Figure 3). One of the reasons for this difference is most probably the influence
of growth ring curvature on the measurements close to the pith, since measurement length is 20 mm.
However, small increase of tangential shrinkage is also noted in similar studies (Cown et al. 1983,
Perstoper et al. 2001).
7
6
5
4
Spiral grain angle [°]
3
2
1
0
-1
-2
-3
-4
-5
0 50 100 150 200 250
Distance from the pith [mm]
Figure 2 Spiral grain angle variation of Norway spruce wood with distance from the pith.
Figure 3 Variability of tangential shrinkage in radial direction of Norway spruce wood.
Drying kinetics
Position of lamellas in cross section of trees significantly influenced their green moisture content
and its variation, as well as their drying progress. Generally, low MC values, from 40 to 80%
were determined in central part of logs. Contrary, lamellas from the circumference of logs had much higher green
MC, always above 100% (
Figure 4). The green state has strong impact on drying kinetics
in followed procedure. During the drying, central and intermediate lamellas reached fibre saturation
point already in first 2 days of the process, whereas at least doubled time was needed to reach a similar
143
MC at peripheral lamellas (Figure 5). Additionally, because of high MC at the beginning, higher
MC gradient at cross section of peripheral lamellas is expected during the whole drying process.
200
180
160 GREEN
140
120
MC[%]
100
80 2 days
60
40 4 days
20 7 days
0 9 days
-200 -150 -100 -50 0 50 100 150 200 250
Distance from the pith [mm]
Figure 4 MC distribution with distance from the pith in green (──), partly dried (···) and in kiln dried state (- - -) of
Norway spruce boards (Log #1).
200 20
180 18
160 16
140 14
Twist [°/ m]
120 12
MC[%]
100 10
80 8
60 6
40 4
20 2
0 0
0 1 2 3 4 5 6 7 8 9
Time [days]
Figure 5 Typical drying curves and twist development at kiln dried boards of Norway spruce from sapwood (──),
intermediate wood (- - -) and heartwood (···).
Twist development
Twisting of lamellas, irrespective of their radial position in log, generally started at reaching
of MC around fibre saturation point. According to expectations, more pronounced twisting was
developed on lamellas positioned near the centre of logs, especially in radius of approx. 60 mm.
Succesive sampling in hygroscopic range confirmed almost straight-line warping of lamellas
144
with decreasing of their moisture content (Figure 1). Linear dependency of twist with decrease
of MC was confirmed in similar drying experiments (Mackay 1973), or during static climatic
conditioning (Johansson et al. 2001).
The largest distortions were present at the lowest MC, at the end of the procedure, where lamellas
close to the pith reached mean twist angle of 17 °/m. Fast reduction of maximum twist angle with
distance from the pith was confirmed, where at location of 70 mm from the centre reached constant
mean value, between 1 and 3 °/m (Figure 6). Similar values of twist, and radial tendency has also been
shown for conifer species in some experimental studies (Balodis 1972, Shelly et al. 1979, Johansson
et al. 2001).
Figure 6 Twist distibution on boards of Norway spruce in radial direction at the end of drying process
(MC = 5.0%, CV% = 6.2%).
Modelling twist and regression analysis

Stepwise regression analysis, using determined material data, was made in order to investigate to what
extent different parameters influenced the magnitude of twist. The stepwise regression analysis
is a good tool since collinearity between parameters is taken into account. Using this method,
60% of the variation in twist in whole MC range could be explained in resulting governing equation
of step 3 (constant is excluded from the model):
1
Twist = 0.057 + 0.313 ⋅ ε T + 0.104 ⋅ θ
r
Denotation:
r = distance from the pith [mm],
εT = realized tangential shrinkage [%] (FSP = 30%) and
θ = spiral grain angle in the middle of lamella [°].
To better understand the twist development the same analysis was repeated at successive MC interval
(∆MC = 2%) in the narrower hygroscopic range, where the twist was induced. Values of variable
estimates with explanation of experimental data variation in the stepwise regression model
are presented in Table 1. The ring curvature was confirmed as the most important variable in the whole
145
MC range, and explained more than 80% of twist variation at low values (MC < 8%). Some variability
of twist, namely 1.8 to 4.5%, explains also tangential shrinkage, but only at MC above 18%.
At lower MC, below 16%, gave SGA also low contribution to the multiple regression model (R2 ≤ 2%)
(Figure 7).
Table 1 Stepwise regression analysis of twist variation in hygroscopic range of Norway spruce wood.
Variable /
<6 <8 < 10 < 12 < 14 < 16 < 18 < 20 < 22 < 24 < 26 < 28 < 30
MC
1/r 0.122 0.113 0.112 0.112 0.112 0.099 0.099 0.084 0.08 0.075 0.075 0.075 0.072
St.error 0.006 0.004 0.005 0.005 0.005 0.004 0.003 0.003 0.003 0.003 0.003 0.003 0.003
%model 83.1 82.3 79.9 79.9 79.9 77.5 77.2 67.7 65.4 62.8 62.8 62.8 61.0
εT 0.125 0.169 0.179 0.195 0.195 0.195 0.207
St.error 0.02 0.022 0.022 0.022 0.022 0.022 0.022
%model 1.8 3.3 3.7 4.5 4.5 4.5 5.1
θ 0.289 0.264 0.281 0.281 0.281 0.244
St.error 0.066 0.048 0.048 0.048 0.048 0.039
%model 1.7 1.8 2.0 2.0 2.0 1.8
N 153 269 308 308 308 430 457 515 552 566 568 572 581
2
R 84.8 84.1 82.0 81.9 81.9 79.3 79.0 71.0 69.1 67.2 67.2 67.2 66.1
St.err.of est. 2.196 2.127 2.236 2.236 2.236 2.184 2.148 2.415 2.427 2.473 2.469 2.46 2.485
90 9
Variable contribution: Grain angle, T-shrinkage [%]

80 8
Variable contribution: 1/r, total [%]
70 7
60 6
50 5
40 4
30 3
20 2
10 1
0 0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
MC[%]
Figure 7 Explanatory values of variables in stepwise regression model of twist variation in the hygroscopic
range of Norway spruce wood (─ total; - - - ring curvature; ··· spiral grain angle; - · - tangential shrinkage).
These results confirmed the best stepwise regression modelling at MC bellow 18%, where ring
curvature and SGA explained more than 80% of experimentally determined twist variation.
Many studies on twist, using statistical modelling, brought forward similar findings (Forsberg et al.
2001, Johansson et al. 2001). Regression modelling at MC above 20% was less succesful, even
with significant explanatory value of tangential shrinkage. Most likely, significant impact on this
finding can be assigned to presumably presence of tangible MC gradient and induced drying stresses.
146
Measuring of MC gradient, especially determing MC of board’s circumference, would certainly

improve predicting of twist at high MC, as was confirmed for instance in experimental drying studies
(Mackay 1973) and in FEM models of twist (Ormarsson et al. 1999). The improvement of twist
prediction at high MC would also be possible with determining of drying stresses and analysis of
possible relaxation, as a consequence of drying time, temperature, present MC and reological
properties of wood (Mackay 1973, Price et al. 1980, Pang et al. 2004).
CONCLUSSIONS
The research confirmed significant influence of some material properties, like ring curvature, variying
spiral grain angle with distance from the pith and tangential shrinkage on development and magnitude
of twist in sawn wood of Norway spruce. Proportionaly increasing twist with moisture content
decrease was the most pronuonced closed to the pith at the end of the drying, with lower, almost
constant values at least 70 mm from centre of logs. Stepwise regresion analysis of twist with drying
time confirmed varying impact of determined material properties on experimentaly determined twist
at specific moisture content of sawn wood of Norway spruce.
REFERENCES
Arganbright, D., Venturino, J., & Gorvad, M. (1978). Warp reduction in young-growth Ponderosa pine
studs dried by different methods with top-load restraint. Forest Products Journal , 28 (8), 47-52.
Bäckström, M., & Johansson, M. (2006). Analytical model of twist in Norway spruce (Picea abies)
timber. Scand. J. For. Res. , 21, 54-62.
Bäckström, M., & Kliger, R. (2006). Restraining moisture related twist in timber structures.
Holz als Roh- und Werkstoff , 64, 235-242.
Balodis, V. (1972). Influence of grain angle on twist in seasoned boards. Wood Science , 5, 44-50.
Booker, R. E. (2005). Geometric model to predict twist in unrestrained boards. Wood Science
and Technology , 39, 269-289.
Cown, D.J.; McCounchi, D.L. (1983). Radiata pine wood properties survey (1977-1982). New Zeland
Forest Service. New Zeland Forest Service.
Danborg, F. (1994). Spiral grain in plantation trees of Picea abies. Can. J. For. Res. , 24, 1662-1671.
Forsberg, D., & Warensjö, M. (2001). Grain angle variation: A major determination of twist in sawn
Picea abies Karst. Scand. J. For. Res. , 16, 269-277.
Frühwald, E. (2006). Improvement of shape stability by high-temperature treatment of Norway spruce.
Effect of drying at 120 °C with and without restraint of twist. Holz als Roh- und Werkstoff , 64,
24-29.
Johansson, M., Perstorper, M., & Kliger, R. (2001). Distortion of Norway spruce timber.
Part 2: Modelling twist. Holz als Roh- und Werkstoff , 59, 155-162.
Kliger, R., Bengtsson, C., & Johansson, M. (2005). Comparison between HT and LT-dried spruce
timber in terms of shape stability and dimensional stability. Holzforschung , 59, 647-653.
Mackay, J. (1973). The influence of drying conditions and other factors on twist and torque
in pinus radiata studs. Wood and Fiber Science , 4, 264-271.
Milota, M. R. (2000). Warp and shrinkage of hem-fir stud lumber dried at conventional
and high temperature drying. Forest Products Journal , 50 (11/12), 79-84.
Ormarsson, S., Dahlblom, O., & Petersen, H. (1998). A numerical study of the shape stability of sawn
timber subjected to moisture variation. Part 1: Theory. Wood Science and Technology , 32, 325-334.
147
Ormarsson, S., Dahlblom, O., & Petersson, H. (1999). A numerical study of shape stability of sawn
timber subjected to moisture variation. Part 2: Simulation of drying board. Wood Science
and Technology , 33, 407-423.
Pang, S., & Pearson, H. (2004). Experimental investigation and practical application of superheated
steam drying technology for softwood timber. Drying technology , 22 (9), 2079-2094.
Perstorper, M., Pellican, I., Kliger, R., & Johansson, G. (1995). Quality of timber products from
Norway spruce. Wood Science and Technology , 29, 157-170.
Price, E., & Koch, P. (1980). Kiln time and temperature affect shrinkage, warp and mechanical
properties of southern pine lumber. Forest Products Journal , 47 (1), 41-47.
Sepulveda, P. (2001). Measurement of spiral grain with computed tomography. J. Wood Sci.,
47, 289-293.
Shelly, J., Arganbright, D., & Birnbach, M. (1979). Severe warp development in young-growth
Ponderosa Pine studs. Wood and Fiber Science , 11, 50-56.
Stevens, W. C., & Johnston, D. D. (1960). Distortion caused by spiral grain. Timber Technology,
217-218.
Warensjo, M., & Rune, G. (2004). Effect of compression wood and grain angle on deformations
on studs from 22-year-old Scots pine trees. Scand. J. For. Res. , 19 (5), 48-54.
148
COST E 53 Conference - Quality Control for Wood and Wood Products, 15th - 17th October 2007, Warsaw
Online Wood Industries Database

László Szalai1, János Boros1, Ákos Bausz1
1
Institute of Informatics and Economics, Faculty of Wood Sciences, West Hungarian University
Bajcsy-Zsilinszky. u 9. 9400 Sopron, Hungary
wyx@inf.nyme.hu
Keywords: database, online, wood
ABSTRACT
In our everyday life, we are surrounded by tools that aid us in our work or improve our comfort. A significant proportion
of these products are made of wood or some wood based composite. In the past 10 years, the wood products industry
was significantly restructured in Hungary. The number of small plants and joineries where domestic and tropical species
are used is on the rise. Superior quality and durability became important factors. To achieve this, being able to access the
newest research results and innovations in this area, as well as trade information (both as consumers and suppliers),
is in the interest of woodworking companies. Some professional journals do publish scientific findings, but there is a lack
of comprehensive, widely accessible sources. Not to mention that industry professionals and trades people are often too
busy to browse the literature, and find that searching for and processing relevant information is too time-consuming.
Wood industry is a strategic trade in our region. Establishing a service to store and, later, publish the professional
knowledge accumulated at the Faculty of Wood Sciences, University of West Hungary, as well as the practical experience
generated in various plants, could be very helpful in achieving the goals. The creation of a professional database,
the goal of which was the assessment of parameters belonging to various specialties, has started. The goal of the project
is creating a database encompassing a wide range of information for the industry, which allows authorized industry
professionals to access and extend the content of the database. The extension of the database is aided by special tools
as well. The present structure of the database was established based on the advice of many of the experts at the Faculty
of Wood Sciences. Since the onset of the project, the configuration of the physical database needed for storing data
in the area of wood materials, wood mechanics and wood based panels has been created. This contains parameters
defined by the experts. The structure was shaped through multiple consultations, and it may also be extended
and modified, according to needs arising later. The connections necessary for use have been defined in the database.
This database may be used as a prototype. The web-based user interface for the management of the wood material
database is also ready. The newest available web based technology was used in programming the web-based interface.
Our further goals include finalizing the structure of the further databases besides that of the wood materials,
and extending the functionality of the web interface. Another goal is designing the service content of the knowledge base.
INTRODUCTION
The West Hungarian University is the most important Hungarian research institute of forestry
and wood industry. The project of Online Wood Industries Database was started in 2006
in cooperation with RET (regional university knowledge centre) and the Institute of Informatics
and Economics. The main goal was to create a database structure and a web surface, which is clear
and can be used easily. During the planning and the collection of data, we surveyed the expectations
of the wood industry experts who work at our university. We can say that the completed work meets
their needs because there was no comprehensive online database of wood industry before. The wood
industry experts and students can store, retrieve and publish their research results.
Structure
The users can connect to the database and to the application surface on the Internet, with
the usage of any browser. We use the Apache HTTP web server to publish the data (Apache HTTP
Server 2.0.55). The advantages of this system are that it is free, frequently used and the minimum
hardware specification is low.
The user can access the actual database construction and the stored data on the SQL Server
2005 through the mentioned web server. There is a HTTPS connect between the client and the web
server which makes our system safer. Besides, there is a built-in firewall between the client and
the server. You can see this structure on Figure 1. The safe configuration, developed in this way,
protects the private, uploaded data in the database from the potential Internet attacks.
The web surface was created basically in PHP, which is embedded in HTML code. (PHP 5.1.2)
7
. The advantage of this script-language is its speed as the code is not interpreted by a browser, it runs
149
on the server. It is perfectly suitable for making dynamic web pages, processing forms and also
communicating with databases6. To improve it, we used CSS (Cascading Style Sheets), with which
we could create style sheets so the appearance and the content could be separated. We used JavaScript
and DHTML 9 in the interest of the easier usage, so we can change the content of the pages without
downloading them1. We realized the prototype of client-server architecture with the help
of the mentioned devices and technology.
Figure 1
Our professors’ and teachers’ notes helped to create a good structured database 5; 8. We can
separate the database into three units according to the area of the specialties. We can split these
specialties into other smaller, logical components.
- Wood science specialty
o Wood species and characteristics
o Wood protection and characteristics
- Wood composites specialty
o Composites and characteristics
o Adhesives and characteristics
- Mechanics and physics specialty
Functionality
The system has basically three functions: uploading the data, modification of the uploaded data
and query the data from the database. Nevertheless, there is a registration and an authorization system
to handle the users’ accounts safely and clearly. Both of the web surface and the database support
the multi-language structure.
Data uploading
The “upload module” is an integral part of the web surface, users can upload data
to the database through these forms. Naturally, we place the functions, which contain data about
the database structure (e. g. SQL commands), in separated files, unreachable by the web server in order
to make the surface safer.
However, users can select another type of uploading. If they have a lot of data with the same
structure, they can upload them with the help of a Word or an Excel file.
Data modification
Users can also modify data with the help of forms. If a user has the rights to modify data
(see also authorization system) then the data can be modified. The queried data are represented in
a table. However, there is a Modification button in the last column. After pressing it, you can see
a surface which is similar to the uploading form but you can not change the basic data, only
the attributes. Users can delete the data also.
150
Data query
There are two ways of querying data. On the one hand, there is a conventional characteristic
query. Users have to select from a list which data they would like to query and then these data will be
shown in tables. It is possible to save the resulted data into Excel format. The displayed data
are pageable if need be. There is an option to hide the source of the data. On Figure 2 you can see
the result of a query - characteristics of the speed of sound (sound velocity).
Figure 2
On the other hand, we can query data through a hierarchical query page. The base attributes
and the characteristics make a tree structure on the query page. Users can open the characteristics,
and the previously uploaded data. You can see this method on Figure 3.
Figure 3
151
The multi-language surface
The web surface supports other languages (not only Hungarian), so users can select German
and English language too. The system is prepared to building in other languages, but then the surface
and the database structure need to be modified. You can switch between the languages by clicking
the flags on the main page. For example if we select English, then all of the form texts will be English
on the surface, but the language of the queried data will not change. It is possible to store and visualize
some data in foreign languages (wood species, parameters etc.).
Registration module
On the registration form only the most important data need to be given: a username,
a password, the user’s real name and an email address. Because of the automated attacks
on the Internet, users have to type the characters can be found on the security image. After giving
the data, the system sends an email to the given email address. The user must confirm the registration
with using the link in the email to activate the account. The registration and the log in are important 2;3.
Authorization system
The handling of the privileges is related with the registration system. After registration,
there are two ways users can get privileges. On the one hand they can get privileges individually;
on the other hand they can get them of the group which they belong to. We have created groups for
an easier handling 4. These groups are the ’developers’, ’administrators’, ’authors’, ’readers’ etc.
We have created an admin surface to manage the users and groups.
CONCLUSION
The framework of the online wood industries database and the connected web surface
is working in test environment. Our further tasks are to improve the database and create the framework
of further special science specialties. The web surface has a number of functions. Experts can upload,
modify or delete their data in user-friendly environment and they can query in several ways with
a number of additional facility. Besides, the program is able to work safely and reliably.
Our further task is uploading more and more data with experts, improving the functions
of the surface and completing it with the missing special fields. Besides, it is important to put
the program into an environment which can be used widely by users and researchers.
REFERENCES
Hodicska Gergely (2004): Újratöltés nélküli adatcsere böngészıben.
http://weblabor.hu/cikkek/jsadatcsere/
1. Hodicska Gergely (2004): Munkamenetkezelés alapjai.
http://weblabor.hu/cikkek/munkamenetkezeles1/
2. Hodicska Gergely (2004): Munkamenetkezelés biztonsági kérdései.
http://weblabor.hu/cikkek/munkamenetkezeles2/
3. Hodicska Gergely (2005): Jogosultság kezelés. http://weblabor.hu/cikkek/jogosultsagkezeles/
4. Kovács Ferenc (2002): Adatbázis-kezelı rendszerek tárgy, elıadásanyag. Budapesti Mőszaki
és Gazdaságtudományi Egyetem.
5. Peter Moulding (2002): PHP Haladóknak, Fekete Könyv. Perfact-Pro Kft. ISBN 963-0095-58-
0.
6. PHP online kézikönyv. http://hu.php.net/.
7. Szitás Zoltán (2002): Adatbázisok tárgy, elıadásanyag. NYME, Sopron, Hungary.
8. Dr. Takách Géza (2006): Webes rendszerek programozása tárgy, elıadásanyag. NYME,
Sopron, Hungary
152
Quality assessment of old-growth Scots pine stands in Poland

Stefan Tarasiuk1, Grzegorz Jednoralski2, Krzysztof Krajewski3,
1 2
Department of Silviculture, Forestry Faculty, Warsaw University of Life Sciences - SGGW, Department of Forest
3
Utilisation, Forestry Faculty, Warsaw University of Life Sciences - SGGW, Department of Wood Preservation, Faculty of
Wood Technology, Warsaw University of Life Sciences - SGGW
Key words: Arbotom device, decayed wood, live trees, large dimensioned trees, nondestructive testing, Phellinus pini,
Poland, Scots pine, standing trees, visual quality assessment, visual grading
ABSTRACT
In this paper an overview is presented of the standing trees (stands) of Scots pine quality assessment in Poland as
carried out in the state owned forest (80% of Poland total wood resources). Because of the growing age of pine stands in
Poland and the resulting hazard of timber quality deterioration, there is urgent need to implement non-destructive
techniques for testing of wood in stabding trees; some aspects of field and laboratory investigations of the in living trees
with use of ARBOTOM device are presented.
INTRODUCTION
Public forest constitues 82.3 % of Poland's forest resources. The absolute majority is under
administration of the State Forest Enterprise "State Forests" (PGL LP) - 78.2 % (Raport 2006). The
centralized management over the forests has been legally based in the 1991 Forest Act. One of the
most substantial statements of the Act it that saying that PGL LP are obligatorily self-financing. The
most important territorial units of PGL LP are forest ditricts (about 430 nowadays). Forest districts
have to plan their economic activity for 1 year periods. As the self-finacing of steate forest is in more
than 80% based on the income form wood sells, it is absolutely vital for the plan making that the
assessment of qualitative and quantitative (and thus: financial result) structure of forest stands planned
for harvest in the next year be done possibly accurate. The long practice has shown that the accuracy
of planning is + /- 3 % in terms of total volume, whereas in terms of particualr quantitative-qualitative
classes the error may be well above 10 %. In the middle-to-old-growth stands, every tree considered
to be cut is subject to individual assessment (Oktaba 1999). The assessment criteria follow strictly the
country-wide standards, with particular attention given to defects visible along, usually 4 meter (in
some cases 6 or 7 m), bottom part of the stem. The mentioned standards: (PN-79/D-01011 Drewno
okrągłe. Wady.; PN-92/D-95017 Surowiec drzewny. Drewno wielkowymiarowe iglaste), and the
instructions defining the general technical conditions to be fulfilled by particular timber assortments
make the assesment results comparable country wide. Based on the visual inspection of the bottom part
of the stem, the whole tree is assessed as either of the following four quality classes of large dimension
timber: WA class (superior quality timber, potentially veneer and/or top quality sawn timber), WB
class (potentially plywood and/or top quality sawn timber), WC class (potentially sawn timber),
or WD class (poorest quality large dimension timber).
Those invisible (hidden under bark of the trees) defects are detected by the experienced wood
quality inspectors by means of indirect inspection like, eg, knocking around the stem, considering the
shape of the bottom part of the stem, considering the bark cracking patterns, etc. Besides, also the local
153
terrain knowledge (sites, history etc.) may prove useful in the detection of hidden defects of wood of
standing trees.
Nowadays, when the model of sustainable forestry with all its consequences has been strongly
forced and implemented in Polish forestry sector, it is the environmental functions and the social
functions of the forest that have become the leading ones. The wood production function is less
emphasized. It's becoming more and more frequent that the old growwth stands are not subject to the
normal rotation ages but continue to grow. The continuously growing average age of forest stands
leads to the growing share of large dimensioned trees/stands that may in theory partly supply the
wood market with highest quality, top priced timber. This leads to the search of better methods and
techniques of standing trees quality assessment, more preecise than the eyes and experience of the
inspector.
The adequate knowledge of the health status of the bottom part of the stem would enable a
more precise and realistic planning of harvesting of particular volumes of the most valuable wood
assortments and particularly so, veener raw material, plywood raw material and the niche products
like, eg, shingle. Timber with the rot in the inner part of the stem, may be used for the production of
plywood (assuming the diameter is sufficient). If the rot, or hollow, makes the whole of the stem
classified WD class, it can still be perfect material for the shingle production, if only the outer
defectless sphere of the stem is at least 10 cm.
The modern methods of wood harvest with use of harvesters and forwarders, even makes the
assortment classification procedure of standing trees easier. The machine operator marks (cf, photos
below) the logs of exceptionally high quality (and thus: unit price) in the clear cut area. The colouring
accessories are mounted in the heads of hte harvesters, enabling thus individual marking of particular
logs (different colors mean different assortments). The wood harvested and marked as above
described is willingly purchased by customers who are often ready to pay more per m3 as compared
with the wood cut and extracted from the clearcut area with use of the traditional technologies (chain
saw and skidder tractors extracting the whole uncut trees). The only technical problem is the
neccesity to use large and powerful machinery, able to copy with the large trees (often a single pine
tree may have several m3 of round timber.
Heartwood decay of trees infested by Phellinus pini leads to deterioration of timber and limits
the range of its possible uses. The ability to detect the decayed wood in live trees characteristic of high
silvicultural value or historically important trees, would facilitate the rational decision making,
increasing thus the efficiency of economic and/or conservation activity. Simple acoustic techniques of
standing tree inner structure assessment have long been practiced in forestry (eg, hitting the stem with
use of an axe) those are however heavily biased assessments, with high subjectivity of interpretation,
resulting in the high risk of misinterpretation when deciding upon the need to cut the tree. This is
especially the case when no other hints are available such like the presence of specific symptoms of
the tree being infested with fungus, and the tree itself seems vital. The risk of wood decay grows
rapidly with the age of mature pine trees (Bernadzki 2003). In the State Forests of Poland the area of
old-growth Scots pine stands (that is aged more than 100 years) equals 450 thousand hectares
(Anonymous 2003a, b). The standing volume of the forests is about 130 mln m3. It is especially
important to have a non-destructive tool for detection of possible wood defects in the old-growth trees
like, eg, protected seed stands, reserve stands, plus trees etc. If an operational detection technique
applicable for the standing trees has been developed that would be an invaluable instrument for the
154
decision-making in the management of old-growth pine stands, allowing, eg, for early enough
initiation of the regeneration process.
Instrumental methods of standing trees wood assessment have been recently becoming more
and more popular, but to make them suitable for the practical use needs a large body of study and
testing focused at improving their objectivity and reliability.

The acoustic wave based methods of wood examination have been rapidly developing in the
last decades (Bulleit, Falk 1985). The presence of decaying spaces in wood, or zones of reduced bulk
density as a consequence of fungal activity, affects the propagation mode of acoustic waves, leading in
particular to decreased velocity of their movement on the way from the emitter and to the receiver
(Konarski, WaŜny 1977, Ross et al. 1997). The time recorded of the acoustic wave passage through the
wood enables the calculation of wave velocity. The repeated measurements over a given surface
increase the accuracy and facilitate the graphic visualization of the results. The interpretation of
tomogram images is useful both in studying live (standing) trees and in a number of applications
connected with the wood industry (Soma et al. 2000, Wang et al. 2001). The utilization of non-
destructive acoustic-based techniques has become well developed and more and more frequently used
in the assessment of monument trees and valuable park trees (Chuang and Wang 1997, Rust 2000).
One of the available techniques gaining popularity in recent years is so called „impulse tomography”.
The theoretical rudiments of the method and successful attempts of its application are reported by
Schwarze et al. (2004). In their study carried out on broadleaved tree species the technique proved
useful for the assessment of wood decay in stems of Fagus sylvatica, Aesculus hippocastanum and
Acer platanoides trees infested by several fungi species causing wood decay.
The purpose of the present paper is to make a tentative usability assessment of an acoustic-based
technique for the study of inner damage of Scots pine stems as caused by Phellinus pini. The empirical
material was sampled from an old-growth tree in the Kampinos National Park (central Poland). The
tree subjected to study was 165 years old. It was 26m high, and 37 cm diameter at 0.5m above the
ground level. At about 450 cm above the ground level, there was a fruitbody of P. pini visible on the
stem. The tree was cut and cross-sectioned, and examined with the ARBOTOM device immediately
after in the field conditions.
The device set consists of a compact steering and power supplying unit, connected with the
open chain of emitter-receiver sensors, and is attached to the laptop computer supplied with the
software able of data processing
The acoustic impulses are produced by repeated yet delicate hits done with a little hammer into
each of sensors. At the time of one sensor being hit, all the others are receivers. The velocity of wave
movement is automatically calculated based on the registered time of impulse passage between pairs
of sensors. The upgraded software Arbotom ver. 1.51, has the option of selection of one out of 45 tree
species/genera and individual analysis of the results. The data matrices enable the continuous control
of particular measurement data, their mean values and statistical variability of the results. The software
enables also the elimination of apparent measurement errors. The device processes so the data that the
computer image of wood structure is eventually produced.
155
The measurements of acoustic waves velocity were conducted twice: the field measurements
were done on May 13, 2004 immediately after the tree was cut and cross-sectioned. The second
measurement was conducted after the wood was transported to the laboratory; this repeated recording
was intended to verify the field results accuracy under the fully controlled indoor condition. On both
occasions, a set of 12 sensors were applied, installed in two adjacent surfaces distanced 40 cm from
one another, 6 sensors on each. A total of nine plains were studied, labeled as H20, H60, H100, H140,
H180, H220, H260, H300, and H340 cm. The sensors were fixed to the horizontal surface with use of
tiny pins with their heads inserted about 1 cm in the wood tissue.
RESULTS
After comparing the results obtained in the field and in the laboratory it was proved that for the
entire 9 plains studied the results were very much similar for every horizon. This gave the rationale to
the conclusion that the field obtained results may be repeated under the laboratory condition, assuming
the naturally high humidity of freshly cut wood has been maintained.
Of the total of 9 horizons analyzed, three levels are analyzed representing the H60, H180 and
H300 levels. They reflect precisely the spatial pattern within the entire set of plains.
Mean velocities of the acoustic waves transmission are presented in Table 1. The values given
are means calculated from at least 15 elementary observations; the data are accompanied by the per
cent error values.
The average velocity of acoustic wave passage from one sensor to another ranges from 590 to
906 m/s at level H60; from 693 to 951 m/s at H180, and from 763 to 1214 m/s at H300. The analysis
of the data shows that the mean velocity increases continuously with growing height and thus –
distance from the fruitbody level. Considering the passage velocity in the context of direction (A to B
versus B to A) within a single surface, noteworthy the velocity between the same pair of sensors is not
equal. It is to conclude that if the difference is rather large this may suggest the asymmetric location of
wood zones characteristic of the changed structure (fungi damaged spaces) between the heads.
The maximum error of wave velocity did not exceed the 10% level that is, it was within the
acceptable range (as recommended by the ARBOTOM device manufacturer). The mean per cent error
of the acoustic wave velocity in particular levels (plains) was maintained at: 3.2% (H60), 3.5%
(H180), and 3.9% (H300).
The computer images of wood cross sections clearly detect the anomalies in the stem’s inner
structure. The heaviest damage/destruction was observed in the plain nearest to the level of the fungus
fruitbody. With growing distance from the fruitbody level up the stem the detected defect was
diminishing. The computer images were then verified by comparing them with the real cross-sections.
The healthier the studied wood is the greener its computer images is colored; while damaged
wood images range from light yellow (weakly damaged) and to red color (the heaviest destroyed
wood). The computer images are not identical with the actual macroscopic state of the wood, they are
its approximate projection. The limited number of sensors applied for analysis of a single plain (6), did
not allow for the higher accuracy of images. In future studies, it is recommended do use higher number
of sensors, eg, 12 on each plain, similarly as done by Schwarze et al. (2004). The graphical
presentation of wood structure as obtained following the application of the acoustic method, resembles
156
the visually observed pictures of the wood cross-sections, nonetheless to interpret the images the
expert knowledge and experience is necessary.
Table 1. The acoustic wave characteristics (velocities and runtime errors) obtained for the three levels:
H60, H180, and H300 along a stem of a 160 years old Scots pine tree infected with Phellinus pini in
the Kampinos National Park (Poland)
Sensor Velocity [m/s] Error [%]

number
1 2 3 4 5 6 1 2 3 4 5 6
H60
1 - 737 748 708 763 809 - 2 2 2 2 3

2 780 - 704 813 837 906 5 - 6 4 3 4
3 710 590 - 720 745 719 3 6 - 5 3 3
4 735 811 819 - 649 797 2 2 3 - 4 2
5 845 873 862 698 - 828 2 2 2 3 - 2
6 787 842 736 740 699 - 4 3 2 3 4 -
H180
1 - 764 896 800 822 803 - 3 3 2 3 3

2 713 - 806 838 833 886 3 - 3 2 2 2
3 896 872 - 756 924 923 4 5 - 5 4 3
4 787 866 734 - 820 871 2 2 3 - 3 2
5 794 847 886 799 - 693 6 5 6 9 - 7
6 820 951 931 907 748 - 3 3 2 2 4 -
H300
1 - 851 997 1106 1086 806 - 3 2 2 2 3
2 842 - 774 1058 1214 1077 9 - 10 7 7 7
3 1092 905 - 892 1206 1172 2 3 - 2 2 2
4 1086 1040 763 - 974 1050 2 3 3 - 3 2
5 1047 1175 1049 955 - 786 3 3 3 4 - 4
6 929 1174 1140 1158 902 - 6 5 4 5 6 -
157
Fig.1. Phenotypically superior trees in terms of quality and dimensions are marked standing
(letter A in this photo) before starting the harvest
Fig. 2. The mounting of containers with bright color paint in the harvester's head gives the opportunity
for the individual marking and separate piling of every log marked standing
158
Fig. 3. From the quality timber, high value products are obtained
Veneer production line in Pisz (Poland))
CONCLUSIONS
1. The quality assessment of standing trees as practiced in Poland has proven to be an efficient tool
useful in optimizing the after harvest timber processing.
2.The destruction of Scots pine stem wood caused by the activity of Phellinus pini may successfully be
detected with use of the ARBOTOM impulse tomograph.
2. The device is useful for the identification and damage degree assessment of Scots pine wood in the
stems of standing trees and in fallen trees.
3. The results of measurements conducted under field conditions immediately after cutting the tree and
in the controlled laboratory conditions are comparable to a high degree, assuming wood humidity
(saturation degree) is maintained at a near to natural level.
5. The extent of practical application of the tested technique is limited because of it is rather time-
consuming, nonetheless its use is fully justified when examining trees of high economic importance, or
conservation value, or monument trees.
159
REFERENCES
Anonymous. 2003a. Wyniki aktualizacji stanu powierzchni leśnej i zasobów drzewnych w Lasach
Państwowych na dzień 1 stycznia 2002r. PGL LP – DGLP, Warszawa grudzień 2002.
Anonymous. 2003b. GUS Leśnictwo 2003. Warszawa, wyd. GUS.
Bernadzki, E. 2003. Struktura wieku i zagroŜenie zgnilizną drewna starych drzewostanów sosnowych.
Sylwan 5: 3-12.
Bulleit W.M., Falk R.H., 1985. Modeling stress wave passage times in wood utility poles.
Wood Science and Technology 19, (2): 183-191.
Chuang S.T., Wang S.Y., 1997. Feasibility of evaluating lumber and tree quality by using stress wave
method (I). Forest Products Industries 16, (4): 681-696.
Konarski B., WaŜny J. 1977. Zusammenhang zwischen Ultraschallgeschwindigkeit und den
mechanischen Eigenschaften pilzbefallenen Holzes. Holz als Roh- und Werkstoff 35(7):341-5.
Ross R.J., DeGroot R. C., Nelson W.J., Lebow P.K., 1997. The relationship between stress wave
transmission characteristics and the compressive strength of biologically degraded wood.
Forest Products Journal 47 (5): 89-93.
Rust S., 2000. A new tomographic device for the non - destructive testing of trees. Proc. 12th Int.
Symp. on Nondestructive Testing of Wood, University of Western Hungary, Sopron:13-15.
Schwarze F., Rabe C., Ferner D., Fink S., 2004. Schalltomografische Untersuchungen an
pilzinfizierten Bäumen. Interpretation der Schalltomogramme. AFZ-Der Wald 59(2): 55-60.
Soma T., Shida S., Arima T., 2000. Application of Ultrasonic Scanning in the Wood Products
Industry. Proc. 12th Int. Symp. on Nondestructive Testing of Wood, University of Western Hungary,
Sopron.
Wang J., Biernacki J. M., Lam F., 2001. Nondestructive evaluation of veneer quality using acoustic
wave measurements. Wood Science and Technology, 34 (6): 505 – 516.
Oktaba J., Jednoralski G., Staniszewski P. 1999. Szacunek brakarski drzew na pniu w PGL-LP.
Biblioteczka leśniczego zeszyt 113. Wyd. Świat Warszawa ss. 15
Raport 2006 Lasy w Polsce 2006. Centrum Informacyjne Lasów Państwowych. Warszawa ss. 47
PN-79/D-01011 Drewno okrągłe. Wady.
PN-92/D-95017 Surowiec drzewny. Drewno wielkowymiarowe iglaste.
160
Pre-grading of sawn timber in green condition

Helene Unterwieser 1, Gerhard Schickhofer 2
1
holz.bau forschungs gmbh, Inffeldgasse 24, 8010 Graz, helene.unterwieser@holzbauforschung.at
2
Institute for Timber Engineering and Wood Technology, Inffeldgasse 24/I, 8010 Graz, gerhard.schickhofer@tugraz.at
Keywords: eigenfrequency measurement, green condition, grading, ultrasonic runtime measurement
ABSTRACT
This paper deals with the investigation of the possibilities of timber grading in the green condition and the comparison
of the chosen measurement techniques. For this purpose, a total of 310 ungraded spruce boards with different cross
sections and sawing patterns were investigated by two methods for determination of the dynamic properties. Based on the
green condition of the boards, five test series at different moisture contents were carried out. As result of literature studies
and previous investigations, the focus was laid on the method of eigenfrequency measurement, which seems to be best
suited for application in an industrial production process. The measurement of the eigenfrequency was carried out with an
industrially implemented optical vibration device (ViSCAN, MiCROTEC). For comparison reasons, an ultrasonic runtime
measurement with ‘Sylvatest’ was conducted as well. With these two dynamic parameters and the density it’s possible
to calculate the dynamic modulus of elasticity, which is a good indicator for the static properties for stiffness and strength.
The analysis of a total of 1550 test values for density, eigenfrequency and ultrasonic runtime showed interesting results
of their behavior regarding the moisture content. The dynamic modulus of elasticity, determined out from the
eigenfrequency measurement, demonstrated no dependence on the wood moisture above the fiber saturation point.
A very strong linear correlation between the MoE in green condition and the MoE at a MC of 12 % was found (r² = 0.96).
The dynamic modulus of elasticity, calculated from the ultrasonic runtime measurement in contradiction, did not deliver
such a strong correlation between the dynamic values in green and dry condition (r² = 0.88). It was clearly shown that the
ultrasonic runtime was influenced by the existing moisture content. On the basis of the dynamic investigations the dynamic
MoE, calculated with the eigenfrequency, seemed more qualified to predict a reliable grading result. To prove the pre-
grading with the adjusted dynamic MoE, the static properties in tension and bending of the boards were determined.
The series were divided into a lower and a higher quality class by a defined limit value of the static MoE. This separation
could be reproduced by the adjusted dynamic MoE; only 5 % of the boards were classified wrong.
INTRODUCTION
In many cases companies buy ungraded sawn timber. After the drying process the boards will be
graded prior to the production. Due to the lack of knowledge of the quality, a high percentage of sawn
timber accumulates, which cannot be processed any further. This unprocessed sawn timber needs to be
used differently and therefore often has to be transported to other plants, which generates further
transportation costs. An optimized pre-grading, based on the green state of the timber could ease
the situation by a quality assured grading process in the early stages of the production chain. The most
important thing is to find a grading method of timber in green condition, which can deliver reliable
grading results regarding the mechanical properties of the dried sawn timber. Another point is
the practicability of the economical meaningful separation of low-grade and high-grade timber
qualities concerning the recovery.
Due to literature studies and pre-investigations, the focus was laid on the method of measuring the
eigenfrequency, which seems to be best suited for the application in an industrial working process. The
dynamic modulus of elasticity, which is calculated from the eigenfrequency and density, can be
determined as grading parameter and shows a strong correlation to material properties like the modulus
of elasticity. For comparison the technique of measurement of the ultrasonic runtime will also
be carried out.
The investigation comprised four series of 310 ungraded spruce boards, which were supplied in green
condition (MC > 30 %). The sawing patterns were chosen as centercuts (with and without pith) and
sideboards. The cross sections were typical cross sections of the project partner and ranged from 202
mm to 98 mm in the width and from 49 mm to 41 mm in the depth (determined dimensions in green
condition), whereas the length was 4,500 mm in every case. The material storing took place in a closed
hall of the project partner at temperatures between 21°C and 24°C and a relative humidity of approx.
161
58 %. With these parameters, a slow and gentle drying of the boards could be guaranteed. For the final
measurement of the dynamic properties, a moisture content of 12 % was desired. To meet this aim
the boards were kiln-dried for 10 days. Samples for kiln-drying of each test specimen were taken
before the first and after the last dynamic measurement, to get the information of the moisture content
in the different states of measurement.
The investigation was composed of two parts. The first part included the determination of the dynamic
properties of the boards in green and dry condition. A total of five measurements of the
eigenfrequency and the ultrasonic runtime were accomplished within a period of 12 weeks.
The measurement of the eigenfrequency was carried out on an industrially implemented optical
vibration device (ViSCAN, MiCROTEC); for the measurement of the ultrasonic runtime
the ‘Sylvatest’ was in use. The mass of the boards was determined at each measurement for calculation
of the density. The dynamic modulus of elasticity for both measurement techniques was calculated
using the following equations:
MoE dyn , eigenfrequ ency = ( 2 ⋅ L ⋅ f ) 2 ⋅ ρ
L
MoE dyn ,ultrasonic runtime = ( ) 2 ⋅ ρ
t
The second part of the investigation deals with the determination of the static properties for strength
and stiffness. Each series was split into parts for tension and bending tests according to the standard
EN 408. Besides the tension and bending strength, a global modulus of elasticity (reference length =
2,943 mm) and a local MoE in tension (reference length = 5 · b), as well as a local MoE in bending
(reference length = 5 · b) were determined.
RESULTS
In the following section the results of the analysis of the dynamic properties, determined from five
measurements, will be shown.
The determined mean moisture contents at measurement 1 lay between 48 % and 89 %. The density
of the centerboard series decreased by 23 %. The sideboard series showed the highest values for the
density, but also the highest decrease (35 %) in ρ, which can be explained with the high moisture
content of the sideboards in green condition. The average statistical dispersion decreased from 16 %
to 11 %.
Results of the eigenfrequency measurements
The mean values of the eigenfrequency increased by 118 Hz or 26 %. The average statistical
dispersion decreased from 12 % to 7.5 %. The calculated mean values of the dynamic modulus of
elasticity (MoEdyn) increased by approximately 15 %, whereas the average statistical dispersions stayed
nearly constant over all five measurements. During the measurements 1 to 4 no change in the mean
values of MoEdyn was noticed.
After the data analysis, a strong correlation between the MoEdyn in the green conditions (measurements
1 to 4) and the MoEdyn at a MC of 12 % was found (Figure 1-le). A linear function could be determind,
which allows the calculation of a modulus of elasticity for a wood moisture of 12 % (coefficient of
determination r² = 0.96), based on the MoEdyn at the (unknown) existing moisture content u:
MoEdyn,12 = 1.09 ⋅ MoEdyn, u + 636
The observed strong correlation could be explained by the behavior of the input-parameters of the
MoEdyn, eigenfrequency f und density ρ concerning the moisture content u. The different effects of the
moisture content on these parameters nearly compensated each other in the area above the fiber
saturation point (‘scissor’-effect, Figure 1-ri).
162
Figure 1. Correlation of dynamic modulus of elasticity in green and dry condition (le); Correlation of moisture
content and the ratios of dynamic properties and density (ri)
Figure 1-ri shows the ‘quality-free’ diagram (ratios, determined by division of the green and dry
properties) of the mechanical properties density, eigenfrequency and dynamic modulus of elasticity
in respect of the moisture content. The graphs clearly displays the described compensate effects
on the dynamic modulus of elasticity, which shows no change of the MoEdyn above the fiber saturation
point (r² = 0.05). It can also be seen that the ratios of the observed values are different below
(gray coloured) and above the area of fiber saturation.
Results of the ultrasonic runtime measurements
The mean values of the ultrasonic runtime t decreased by 176 µs or 18 %. The average statistical
dispersion decreased from 11.6 % to 6 %. The calculated mean values of the dynamic modulus of
elasticity (MoEdyn,US) increased by approximately 8 %; the average statistical dispersions stayed nearly
constant over all five measurements.
The analysis of the relation of moisture content and dynamic modulus of elasticity did not offer such
a strong correlation between the observed properties in green and dry condition. The calculated
coefficient of determination r² expresses this fact with a value of 0.88 (Figure 2-le).
Figure 2. Correlation of dynamic modulus of elasticity (US) in green and dry condition (le); Correlation of moisture
content and the ratios of dynamic properties (ri)
An explanation is the influence of the moisture content on the input parameter ultrasonic runtime t
above the fiber saturation point (Figure 2-ri). It can be clearly seen that the correlation of ultrasonic
runtime ratios and moisture content is not as strong as for the determined eigenfrequency f shown
in Figure 1-ri. This means a dependence of the relative change (ratio) of the dynamic property t on
the existing moisture content at time of measurement; therefore density and ultrasonic runtime do not
163
compensate the moisture effects in the same way as observed at the eigenfrequency measurements.
The graph in Figure 2-ri shows this fact in the relation of moisture content and ratios of MoEdyn,US. A
linear correlation with a coefficient of determination r² = 0,463 could be calculated.
Another indication for the potentially better appropriateness of the eigenfrequency measurement
and the dependence of the ultrasonic runtime from the wood moisture respectively is shown in Figure
3.
Figure 3. Trends of the dynamic MoEdyn -ratios in relation to the moisture content
The diagrams in Figure 3 contain the ratios of the dynamic modulus of elasticity for both techniques in
relation to the moisture content. For this purpose, moisture groups were defined (MC steps: 10 %) for
representing a possible trend of the mean values of the relative changes of the dynamic modulus of
elasticity. The ratios of the MoEdyn out of the eigenfrequency measurement show no dependence on the
moisture content; the average values of the moisture groups oscillate around the global factor of 1.15,
which was determined by considering all test series and all measurements. Within the moisture content
of 30 % and 180 %, the calculated mean values vary between 1.13 and 1.17. Compared to that, it is
clearly shown that the ratios of the MoEdyn,US, calculated from the ultrasonic runtime measurement,
are decreasing with increasing moisture content. Within the moisture content of 30 % and 180 %
the calculated average values decrease from 1.15 to 1.02, combined with less spread of the ratios.
The determined global factor of 1.078 does not represent the trend of the ratios very well above fiber
saturation.
DISCUSSION
As described in the paragraphs before, the modulus of elasticity, calculated from the eigenfrequency
measurement, is more qualified for a pre-grading in green condition because of the independence from
the wood moisture above the fiber saturation. The reliability of the defined equation of a dynamic MoE
for dry condition, based on green condition, has to be proved in practice. For this purpose the static
properties were required for classifying the test specimens in quality classes.
For the examined pre-grading, the requirement concerning the numbers of grading classes was
the separation by the static modulus of elasticity into two quality classes, a low-grade and a high-grade
timber quality class. For this purpose the mean value of the static MoE (global MoE in tension, local
MoE in bending) was fixed with 9,000 N/mm² for the quality separation.
A total of 154 boards were tested in tension. 32 boards (20.8 %) matched the lower quality class
concerning the global modulus of elasticity in tension. After calculation of the dynamic MoE with
the given linear function the same ranking of the boards was attempted. The class limit from
164
the dynamic MoE could be found with 10,000 N/mm². 37 boards (24 %) had to be arranged
in the lower quality class; this represented a failure classing of 4.6 %.
155 boards were tested in bending according to EN 408; 34 boards (21.9 %) had to be classified
in the lower quality class concerning the local modulus of elasticity in bending. After calculation
of the dynamic MoE with the given linear function, the same ranking of the boards was attempted.
The class limit of the dynamic MoE could be found with 10,600 N/mm², whereas 45 boards (29.7 %)
had to be arranged in the lower quality class. This represents a incorrect classing of 14 %.
The classing with the dynamic MoE for the bending series did not deliver such reliable results
as for the tension series. One of the reasons is the weaker correlation of the dynamic and static bending
modulus of elasticity in comparison to the strong relations of dynamic MoE and modulus of global
elasticity in tension (Figure 4).
Figure 4. Relations of static and dynamic MoE for tension (le) and bending (ri)
It is clearly shown that the correlations of the MoE in tension and dynamic MoE are strong for each
series (0.87 < r² < 0.97). The reference length for calculating the global modulus of elasticity is given
with 2,943 mm, which nearly covers the entire board length (3,500 mm). Both MoE, the global MoE
in tension and the dynamic MoE display the whole board with its characteristics; thereof
the correlation of the two properties is stronger than that for the MoE in bending. For the bending
series a wider spread of the test values is displayed, especially for series C (sideboards). For this series
a coefficient of determination r² of 0.65 could be reached. The local bending MoE with the reference
length of 5 · b does not display the entire board with its characteristics.
CONCLUSIONS
The investigation showed that a pre-grading of boards in green condition is possible. After analysis,
the technique of eigenfrequency measurement could be more suitable for the pre-grading purpose
because of the independence of the dynamic modulus of elasticity from the moisture content above
fiber saturation. The input parameters density and eigenfrequency compensate their moisture
influences by reason of their contrary behavior. With a simple linear function it is possible to calculate
a ‘dry’ dynamic modulus of elasticity, based on green condition, without knowledge of the existing
wood moisture content. The strong correlation between the green and dry dynamic MoE is displayed
by a coefficient of determination of r² = 0.96. For comparison, the measurements of the ultrasonic
runtime deliver modulus of elasticity, which do not show such a good correlation in green and dry
condition (r² = 0.88). The effects of moisture on density and ultrasonic runtime do not compensate
each other in the same way as the input parameters of the MoE, determined from the eigenfrequency
165
measurement. The relative changes of the dynamic MoE are influenced by the existing moisture
content in the wood. Thus, a reliable pre-grading will be hindered. Another point to prefer
the technique of eigenfrequency measurement was observed during the investigation. Faulty
measurements of the ultrasonic runtime were observed, when a special industrial noise was present
during the measurement. The measurement had to be repeated.
After the determination of the static properties in tension and bending, the pre-grading with
the calculated dynamic MoE could be carried out. For this purpose the four series were divided into
a lower and a higher quality class, whereas the limit value of the static MoE was defined
with 9,000 N/mm². The separation of the boards was tried to reproduce with consideration
of the dynamic MoE from the green condition, which was adjusted with the given linear function
to a wood moisture of 12 %. From a total of 309 spruce boards 10 % were classified wrong.
Thereof, 5 % of the boards showed very low stiffness properties, but were classified in the higher
quality class.
The aim of the current investigation was a pre-grading of sawn timber in green condition and not
a substitute of the grading process in dry condition. Most important for industrial use is the definition
of the desired grading or quality classes. With the investigated stiffness-pre-grading, it is possible
to separate low quality boards from boards with high (strength) stiffness properties. For definition
of several grading classes, the pre-grading with the stiffness properties only is possibly not sufficient
for prediction of strength properties. Additional grading techniques have to be accomplished. Further
investigations should deal with that task.
REFERENCES
1. Blaß H.J., Frese M., (2004): Sortierverfahren für die kombinierte maschinelle und visuelle
Festigkeitssortierung. Forschungsbericht, Lehrstuhl für Ingenieurholzbau und
Baukonstruktionen, Universität Karlsruhe (in German)
2. Hanhijärvi A., Ranta-Maunus A., Turk G., (2005): Potential of strength grading of timber
with combined measurement techniques. Report of the Combigrade-project-phase 1,
VTT Publications 568, Espoo
3. Steiger R., (1996): Mechanische Eigenschaften von Schweizer Fichten-Bauholz bei Biege-,
Zug-, Druck- und kombinierter M/N Beanspruchung. IBK Bericht Nr. 221,
Institut für Baustatik und Konstruktionen, ETH Zürich (in German)
4. Unterwieser H., Fritz S., (2006): Steifigkeitssortierung von Schnittholz im Frischzustand.
Arbeitsbericht der Vorversuche, holz.bau forschungs gmbh (in German)
166
Timber grading machine using multivariate parameters

based on ultrasonic and density measurement
Sandoz Jean-Luc, Benoit Yann
CBT SA, Rue des Jordils 40, 1025, Saint Sulpice, Switzerland
Keywords: grading, machine, non-destructive evaluation, wood
ABSTRACT
The latest European standards for timber grading (EN 338) include now up to ten resistance classes, based on the modulus
of rupture MOR, the modulus of elasticity MOE and the density ρ of wood.
The visual inspection being unable to give an accurate grading regarding these new norms, non-destructive technologies
are developed to answer to these fundamental concerns.
This paper relates a new industrial machine performing all these new characteristics – Triomatic – based on the previous
industrial equipment using ultrasonic – Sylvamatic – and on a new parameter: the density measurement.
If Sylvamatic succeeded in the first step of the Combigrade project managed by VTT – Finland, the upgraded version
Triomatic has participated to the second phase of the project and has followed the calibration process for French species
under the CTBA’s supervision for a reliable European Grading Assessment.
INTRODUCTION
CBS-CBT Group [1] is specialized in timber engineering and non-destructive technology for wood
quality assessment.
In 1985, the ultrasonic technology was transferred into a patented portative device, Sylvatest, thanks to
a thesis work done at the Swiss Federal Institute of Technology, Lausanne [2].
In 1998, a newer portative version has been developed – Sylvatest Duo – using not only the ultrasonic
measurements, but also the acoustic phenomenon to increase the evaluated results reliability [3].
Concerning the industrial equipment, CBS-CBT was among the firsts to propose a reliable industrial
technology with the Sylvamatic machine.
In parallel, the European standards EN338 [4] were developed until the proposal of new classes of
resistance, and especially the recognition of high performance timber with the class C40 (table 1).
If the visual inspection can be used from the class C14 to C30 (assuming hazardous reliability), non-
destructive technologies are now needed in order to answer to the new standards concerns.
Moreover, next year, in 2007, the timber industry will be asked to assume the traceability for each
structural element, including the resistance class defined in the norms.
Industrial non-destructive grading is then a fundamental topic today in Europe. It is for this reason why
VTT, in Finland, launched last year a comparison campaign between the different existing
technologies [5].
If Sylvamatic has performed excellent results, a new machine has just been developed by CBS-CBT,
adding the density measurement to the initial Sylvamatic parameters.
This paper relates the development and the European grading assessment process of this new industrial
machine called Triomatic.
167
Table 1: Mechanical values for timber design, according to the European standards EN 388.
Strength classes
Characteristics Symbol C 14 C 16 C 18 C 22 C 24 C 27 C 30 C 35 C 40
Bending [N/mm2] fm,k 14 16 18 22 24 27 30 35 40
Tension || [N/mm2] ft,0,k 8 10 11 13 14 16 18 21 24
Tension ⊥ [N/mm2] ft,90,k 0,3 0,3 0,3 0,3 0,4 0,4 0,4 0,4 0,4
Compression || [N/mm2] fc,0,k 16 17 18 20 21 22 23 25 26
Compression ⊥ [N/mm2] fc,90,k 4,3 4,6 4,8 5,1 5,3 5,6 5,7 6,0 6,3
Shear [N/mm2] fv,k 1,7 1,8 2,0 2,4 2,5 2,8 3,0 3,3 3,8
Elasticity modulus [kN/mm2]
– parallel average E0,mean 7 8 9 10 11 12 12 13 14
– parallel f5% E0,0.5 4,7 5,4 6,0 6,7 7,4 8,0 8,0 8,7 9,4
– perpendicular average E90,mean 0,23 0,27 0,30 0,33 0,37 0,40 0,40 0,43 0,47
Shear modulus average [kN/mm2] Gmean 0,44 0,50 0,56 0,63 0,69 0,75 0,75 0,81 0,88
Density minima [kg/m3] ρk 290 310 320 340 350 370 380 400 420
Density average [kg/m3] ρmean 350 370 380 410 420 450 460 480 500
THE NDE TECHNOLOGY
The non-destructive technology used by the Triomatic is based on the following parameters:
• Acousto-Ultrasonic phenomenon
• Local density measurement
• Moisture content measurement
This paragraph describes the measurements of these different parameters.
The Acousto-Ultrasonic phenomenon

The ultrasonic measurement
From 1985, the ultrasonic method for the measurement of the mechanical performances of timber – the modulus
of elasticity: MoE|| and the bending resistance: σb – has been validated for the wood as a structural material [6].
At the end of the 80’s, a technology transfer has been realised with the Sylvatest device [6] through the results
of a thesis work [2]. This device was based on the measurement of the speed of propagation of a low frequency
wave transmitted in the longitudinal axis of the wood (figure 1) as shown on the equation (1):
Sender VUS Receiver
Ultrasonic wave transit
Figure 1: Longitudinal ultrasonic measurement in a wood piece.
C LL MoE //
VL = = [1]
ρ ρ .1.82
For a species such as spruce, a calibrated model giving the MoE|| and the bending strength σb can be written as
follow (equations 2 and 3):
MOE // = α 1VL + β 1 [2]
σ b = α 2VL + β 2 [3]
Where:
VL: the waves’ velocity in the longitudinal axis [m/s];
MoE//: the modulus of elasticity parallel to the grain [N/mm2];
ρ: the density of the material [kg/m3];
σb: the bending modulus of rupture [N/mm2];
α And β: calibration parameters.
168
The acousto-ultrasonic measurement
In 1998, IBOIS, the laboratory for timber construction of the Swiss Federal Institute of Technology in
Lausanne, has developed and improved the ultrasonic technology for wood.
The new generation of device, Sylvatest Duo, deals with the analysis of the acousto ultrasonic
response of the wood [3] (figure 2).
V
exp(-ax+b)
max peak
t0 t1 t2
t
Cumulative Signal Energy

min peak
Stress Wave methods Acousto-ultrasonic methods
Figure 2: Sylvatest-Duo and its transducers (left) and analysis of the acousto-ultrasonic signal with the measurements
of the speed and of the maximal peak of energy of the transmitted waves (right).
The system always measures the speed of the transmitted low frequency wave (22 kHz), and measures
too the maximal peak of energy of these waves thanks to the equipment presented by the figure 2.
The speed of propagation is still correlated to the modulus of elasticity (MoE||), but the energy is
correlated to the local singularities (knots, grain direction, degradation area…).
In fact, the energy damping of the waves is directly dependant of local singularities. The maximal
value of the peak of energy represents thus a measurement of the acoustic response of the wood which
translates faithfully the damping function.
This new generation of device, able to measure and manage the two acousto-ultrasonic variables,
allows working in the wood natural axis: the longitudinal, radial and the transversal ones.
The density measurement

In order to estimate the wood’s density, an extra measurement module can be installed. This module is
based on the local density measurement composed by two pins screwed on a load sensor.
A jack pushes the system into the wood and the compression load is measured as described by the
figure 3.
1 2 3
Jack
Load sensor
Pins
Wood
Figure 3:
Density measurement module. Two pins are screwed on a load sensor. The compression load is measured in order to
evaluate the wood’s density
169
The compression load measured when the pins penetrate the material is correlated to the wood local
density as illustrated by the equation 4.
ρ = αC + β [4]
With:
ρ: The local wood density [kg/m3]
C: The compression load measured when the pins penetrate the wood [V]
α and β: Calibration parameters
The moisture content measurement
Affecting the speed of the ultrasonic waves and the compression load value the wood moisture content
must be measure
This parameter can be evaluated thanks to a common Gann as illustrated by the figure 4.
Moisture content
meter
Wood
Figure 4: Moisture content measurement principle
THE INDUSTRIAL MACHINE: TRIOMATIC

Presentation
Triomatic is the non-destructive machine using the parameters described in the previous chapter:
• Acousto-ultrasonic measurement
• Local density measurement
• Moisture content measurement
In order to increase the process, several pairs of transducers are possible (figure 5). Moreover, the
density measurement and the moisture content measurement can be done within the same module.
To avoid the side effect for the plank, two density measurements are realized: one is above the plank,
and the second makes a measurement from under the plank, like a jaw (figure 6).
170
Figure 5: Ultrasonic module composed here by two pairs of transducers. Triomatic Ecolam, Belgium.
More pairs could be added in order to increase the process if needed.
Figure 6: Density and moisture content measurement module. A jaw composed by two load sensors on which
a pair of pins is screwed measures the wood local density from above and from under the plank. The moisture content
is simultaneous measured once the pins are in the wood. Triomatic, Ecolam, Belgium.
Concerning the knots, a maximal value of compression load is considered. If this maximal value is
reached or overtaken, the parameter “knot” is taken into account. Then, a special algorithm manages
the density measurement, comparing the both measurements (from above and from under the plank)
plus the consideration of the eventual knots (equation 5).
ρ a = αC a + β ρ + ρu
If Ci > Cmax, then Ci = Cknot ρ = a [5]
ρ u = αC u + β 2
With:
ρa: Density measured from above the plank [kg/m3]
ρu: Density measured from under the plank [kg/m3]
ρ: Evaluated local wood density [kg/m3]
C a: Compression load measured from above the plank [V]
Cu: Compression load measured from under the plank [V]
Cmax: Maximal compression load considered [V]
Cknot: Compression load considered if knots are present [V]
171
Calibration
Some results of the first step of the Combigrade project can be presented. The results of the second
step should be published during autumn 2006.
The first step of the Finnish project was involving 100 planks of Spruce and 100 planks of Pine both
from Finland and Russia.
The presented results consider the correlations between the failure tests giving the modulus of
elasticity MOE and the modulus of rupture MOR versus the Triomatic evaluation taking into account
the speed of ultrasound, the density of the planks and the wood moisture content.
Figure 7 presents the results for Spruce and figure 8 for Pine.
Table 2 summaries the results about the reliability of the non-destructive evaluation.
70
17000
60
15000
MOE Actual
MOR Actual
13000 50
11000 40
9000
30
7000
20
5000
4000 6000 8000 10000 14000 18000 20 30 40 50 60 70
MOE Predicted P<.0001 RSq=0.68 MOR Predicted P<.0001 RSq=0.45
RMSE=1410.1 RMSE=9.3866
Figure 7: Results of the non-destructive measurements given by the Triomatic(X axis) machine versus the failure
tests (Y axis) for Spruce from Russia and Finland MOE, left, MOR right. From Combigrade project 1
20000 80
18000 70
16000 60
MOR1 Actual
MOE1 Actual
14000
50
12000
40
10000
30
8000
6000 20
4000 10
4000 8000 12000 16000 20000 0 10 20 30 40 50 60 70 80
MOE1 Predicted P<.0001 RSq=0.78 MOR1 Predicted P<.0001 RSq=0.64
RMSE=1426.9 RMSE=10.069
Figure 8: Results of the non-destructive measurements given by the Triomatic machine (X axis) versus the failure
tests (Y axis) for Pine from Russia and Finland MOE, left, MOR right. From Combigrade project 1
Table 2:Coefficient of determination r2 results of the non destructive evaluation operated by Triomatic versus the
failure tests done on 100 planks of Spruce and 100 planks of Pine, both from Russia and Finland.
Spruce Value r2
MOE 0.68
Triomatic
MOR 0.45
Pine Value r2
MOE 0.78
Triomatic
MOR 0.64
CONCLUSION
Triomatic is a machine available on the timber strength grading market. Its good results and its cost are
very interesting factors for sawmills and gluelam factories.
Based on the measurements of ultrasonics, density and moisture content, all the characteristics
expected by the norms can be properly evaluated: MOE, MOR and density of each wooden beam.
By adding the density measurement to the ultrasonic value, the gain on the reliability is very
interesting as illustrated by the table 3.
172
Table 3: Gain obtained by adding the density measurement (Triomatic) to the single ultrasonic value (Sylvamatic).
Pine
r2 MOE MOR
Sylvamatic 0.67 0.51
Triomatic 0.78 0.64
Gain 16.4% 25.5%
Triomatic is ready for the CE label which should appear from the next year in 2007 all over Europe.
A new calibrating project has been launched with the collaboration of the CTBA, France, in order to
validate the machine for French species.
Samples of a same species are studied according to their growth location (region + altitude), the log’s
diameter, and the plank’s position of the in the original log.
If Triomatic is fully automatic, a mobile semi-automatic version can also be proposed, especially for
smaller timber companies.
REFERENCES
[1] Group CBS-CBT, www.cbs-cbt.com
[2] Sandoz, J.-L., 1990: “Grading and reliability of construction timber, validation of the ultrasound
method”, thesis n°851, EPFL, IBOIS, Lausanne, Switzerland
[3] Sandoz, J.-L.; Benoit, Y.; Demay, L., 2000: “Wood testing using acousto-ultrasonic”, World
Conference on Timber Engeneering, Whistler, Canada.
[4] European Standards for Timber Construction EN338. Mechanical properties for sawn wood used as a
structural material
[5] Hanhijärvi, A.; Ranta-Maunus, A.; Turk, G., 2005:”Poteantial of strength grading of timber with
combined measurement techniques”, Report of the Combigrad project, phase 1.
[6] Sandoz, J.-L., 1989: “Grading of construction timber by ultrasound”, Wood Science and Technology
23:95-108
173
COST is supported by the EU Framework Programme
ESF provides the scientific, technical and administrative
secretariat of COST
Founded in 1971, COST is an intergovernmental

framework for European Co-operation in the field
of Scientific and Technical Research, allowing
the co-ordination of nationally funded research
on a European level.
COST Actions cover basic and pre-competitive
research as well as activities of public utility.
The goal of COST is to ensure that Europe holds
a strong position in the field of scientific
and technical research for peaceful purposes,
by increasing European co-operation and interaction.
More info: http://www.cost.esf.org

ISBN 978-83-7244-904-7

COSTE53 ConferenceWarsaw Presentations All PDF

Uploaded by

Copyright:

Available Formats

COSTE53 ConferenceWarsaw Presentations All PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

COSTE53 ConferenceWarsaw Presentations All PDF

Uploaded by

Copyright:

Available Formats

Quality Control

Faculty of Wood Technology

THE FIRST CONFERENCE

Relevance of wind-induced compression failures regarding 1

Wood and wood products quality evaluation in Slovakia 75

Relevance of wind-induced compression failures regarding bending strength and

MATERIALS AND METHODS [1]

MOR = a0 + a1⋅DENS + a2⋅USMIN + a3⋅KMAX + a4⋅KCLU + a5⋅CW + a6⋅FUNG + a7⋅CFIND + ε1 (1)

RESULTS AND DISCUSSION

Effect of the size of CF

DEVELOPMENT OF QUALITY MANAGEMENT IN POLISH WOOD

DEVELOPMENT OF QUALITY MANAGEMENT SYSTEMS AROUND THE WORLD

Figure 2. Growth of certification of Quality Management Systems according to ISO 9001:2000

POLISH WOOD INDUSTRY

SECTOR total 0-9 10-49 50-249 250-999 1000 and more

The first sector includes:

Average paid Average monthly gross

A current condition of wood industry in Poland in 2007 was presented in table 2.

Gross Domestic Product (GDP) 10 527,0 bln € 268,7 bln €

Value of the furniture sold production ** 93,8 bln € 5,6 bln €

Furniture production share in GDP 0,9% 2,1 %

Manufacture of wood and wood, straw and wicker products

Manufacture of furniture; manufacturing n.e.c.

DEVELOPMENT OF QUALITY MANAGEMENT SYSTEMS

Bark Measurements with X-Ray Technology

Keywords: bark, spruce, x-ray

MATERIALS AND METHODS

RESULTS AND DISCUSSION

Statistical distribution of single bark measurements

Measurements for whole logs

Dimensional and Shape Quality Control. Modern Equipment

K ey w o rd s : processing accuracy, measurement methods, modern equipment.

EQUIPMENT AND MEASURING METHOD

Figure 1. CADesQ measuring equipment

Figure 2. Interface of the measuring program on CADesQ equipment

Figure 3. OptoDesQ equipment

Figure 4. Interface on OptoDesQ program

Figure 5. Results generated by OptoDesQ program

Figure 6. LaminatDesQ equipment

Ultrasonic evaluation of mechanical properties of wood in standing trees

Keywords: mechanical properties, non-destructive testing, standing trees, timber, ultrasound

MATERIALS AND METHODS

VtL = 0,7267 VL + 2310,7

Wave velocity along the standing tree trunks VL [m/s]

4) Irregularity of a shape or non-centric distribution of inner rot can be determined by a cross-wise

Moisture buffering capacity

Measurement of wood surface layer drying (MARVINGCA project)

Water accumulation at the coating/wood interface (MARVINGCA project)

Water evacuation from a curing glue line (MARVINGCA project)

MATERIALS AND METHODS

RESULTS AND DISCUSSION

Curing of the glue line

Assessment of Early Stage Fungal Decay of Wood

Karin Fackler1,2, Barbara Hinterstoisser3, Manfred Schwanninger4, Cornelia Gradinger1,2, Ewald

Keywords: beech wood, chemometrics, decay fungi, FT-NIR, spruce wood

MATERIALS AND METHODS

RESULTS AND DISCUSSION

NIR 2nd derivative

-9.0E-06 native beech