Modification of Davissons method

Modification de la methode de Davisson

F.A. Baligh

G.E. Abdelrahman

ABSTRACT Davissons method (Offset Limit Method) is probably the best-known and most widely used method for predicting the ultimate pile load from load test results. This method when applied to the load-settlement curve of a tested pile, usually fails, unless the pile is loaded close to failure. Results of eighty static axial load tests on different types of piles carried out in Egypt were analyzed according to Davissons method to obtain the ultimate load. Only four out of the eighty tests satisfied the method and predicted the ultimate load. Elastic and plastic settlements were measured for each pile load test at test load. These values were compared with those obtained from their relevant parts in Davissons equation. Davissons equation, when applied at test load equal to either one and a half or twice the working load, gives values higher than those measured for both the elastic and plastic parts. Davissons equation was modified to accommodate the difference between the calculated and the measured values at the test load. The suggested form of Davissons equation allows the prediction of the pile ultimate load using the pile test load without having to extrapolate the load-settlement curve. RSUM La mthode de Davisson une des plus propag mthode pour prdire le poids le plus grand charge ultime pour les pieux et cette mthode quand elle faite sur le courbe du poids et la rupture du pieu examin toujours ne reussi pas sauf si que le pieu se met presque au poids dtructeur. Les rsultats de 80 expriences pour les poids verticales sur plusieurs genres des pieux ont taient faite en Egypte et elles taient calcules par la mthode du Davisson pour produire la charge ultime du poids du pieu , ilya4 expriences seulement des 80 expriences satisfes la mthode et la prdiction de la charge ultime. Le tassement de llastic et le plastic sont mesurs pour chaque exprience a une poids expriencelle . Ces valeurs sont compares avec ceux qui sont calcules de la quation de Davisson . Lquation de Davisson quand elle applique au poids expriencelle une fois et demi ou deux fois donne des rsultats plus grand que les mesures de lexprience de llastic et le plastic. Lquation de Davisson tait modifie pour accommoder le difference entre la valeur qui tait calcule et la valeur qui tait mesure au poids expriencelle. La forme propos du mthode de Davisson permit la prdiction du charge la plus grand ultime pour le pieu par utiliser le poids experiencelle a lallongement du poids et du tassement du pieu . 1 INTRODUCTION The working load or design load is the load a pile is designed to carry safely within a limited range of settlement; this limited settlement range depends on the nature of the building, its importance and also the properties of soil in which the pile is installed. Design load is pre-calculated using field and laboratory test results. Load tests are performed to prove the design load and check the pre-chosen factor of safety. A method to insure the above was proposed by Davisson (1972); which is easy to apply and has gained wide acceptance. The Offset Method defines failure as the intersection of the elastic stiffness of the pile drawn through an offset on the abscissa that depends on the pile diameter. Since the offset is defined by the pile diameter, the capacity is dependent on the pile diameter. The proposed modified method estimates the ultimate pile load from the axial pile load test by using a reduction factor for Davisson equation. For the purpose of this study, results of seventy six field axial load tests database on different pile types were compiled. The tested piles included driven, bored, and continuous flight auger piles with different diameters. The tests were carried out at various locations in different Egyptian soils. The soils for all the sites consisted of different layers of silty clay to fine sand along the pile length and the pile tip ended in medium to graded sand. 2 THE DAVISSON OFFSET LIMIT LOAD The method was proposed by Davisson (1972) as the load corresponding to the movement that exceeds the elastic compression of the pile (taken as a free standing column) by a value of 3.8 mm plus a factor equal to the diameter of the pile (in cm) divided by 10. The method is based on the assumption that ultimate capacity is reached at a certain small toe movement. It was primarily intended for test results from driven piles, but when applied to bored piles, it becomes impractically conservative. The offset limit criterion is intended for interpretation of quick testing, in which each load increment is held for periods not exceeding one hour. It can also be used when interpreting results from slow methods. This method also gained widespread use in phase with the increasing popularity of wave equation analysis of driven piles and dynamic testing. The load settlement curve is plotted to a convenient scale, so that the line OO1 makes an angle of about 20 degrees with the load axis, (Figure1). The line OO1 represents the relationship between the load and shortening of an elastic free axially loaded column which equals QL/AE. The line CC1 is drawn parallel to OO1 at an offset distance OC, where: D is in cm and OC equals (3.8+ 0.08 D) in mm. The intersection of CC1 with the loadsettlement curve gives the ultimate pile load Qult or 0.9 Qult according to Egyptian code of practice [4], as shown by the following equation: 2079

Sult = (0.08 *D+3.8) + (QultL/AE)

(1)

Where: Sult: settlement (mm) of pile at ultimate load, D: diameter of the pile at the pile tip (cm), Qult: ultimate load (ton), A: base area of the pile (mm2), E: modulus of elasticity of the pile material (2 ton/mm2) and L: length of the pile shaft (mm).
Load, Q

Table 1: Results of Axial Load Tests for One and Half the Working Load
Test No Dr1 Dr2 Dr3 Dr4 Dr5 Dr6 Dr7 Dr8 B9 B10 B11 B12 B13 B14 B15 B16 B17 B18 B19 B20 B21 B22 B23 B24 F25 F26 F27 F28 F29 F30 F31 F32 F33 F34 F35 F36 F37 F38 F39 F40 F41 F42 L m 13 17.5 14.75 22.25 19.3 15 26.5 13 22.85 13.1 11 12 14.55 13.35 13.5 17.25 12.5 13 22.5 21 14.8 20 20 13.35 18 18 18 14 14 14 13.6 13.6 12 13.2 12.2 12.2 15 12.2 21 21 15 25 D mm 450 500 530 430 400 400 600 400 800 500 500 500 600 450 800 600 800 600 600 600 400 600 600 450 650 600 600 600 600 600 400 400 600 400 400 400 400 400 500 500 600 400 Test load (ton) 1.5 W.L 67.5 67.5 120 75 60 60 195 80 262.5 105 75 85 90 52.5 187.5 150 240 153.75 120 75 45 150 157.5 52.5 210 180 180 120 120 120 45 45 90 30 30 30 30 30 97.5 97.5 150 55.5 Measured settlement at Test Load (mm) Elastic Plastic Total Sett. Sett. Sett.. 1.85 4.00 2.43 2.68 2.18 1.98 4.37 2.98 0.94 2.09 1.02 1.59 0.49 0.62 1.48 2.19 2.62 5.23 0.90 1.15 1.21 0.96 2.45 0.56 2.48 2.95 3.22 1.33 1.24 0.99 1.78 1.22 0.63 0.48 0.46 0.49 0.41 0.35 0.79 1.77 2.95 0.81 0.38 0.41 0.33 0.12 1.62 0.40 0.67 1.30 1.13 1.19 0.97 2.57 0.26 0.32 4.25 2.37 7.78 5.04 0.58 0.33 0.20 1.63 1.56 0.32 1.33 1.29 1.65 0.75 0.28 0.66 2.22 1.65 0.54 0.26 0.23 0.26 0.21 0.37 0.57 0.57 3.3 0.37 2.23 4.40 2.75 2.80 3.80 2.38 5.04 4.28 2.06 3.27 1.99 4.16 0.75 0.94 5.73 4.56 10.40 10.27 1.48 1.48 1.41 2.58 4.01 0.88 3.81 4.24 4.87 2.08 1.52 1.65 4 2.87 1.17 0.74 0.69 0.75 0.62 0.72 1.36 2.34 6.25 1.18

0 0 O 2 4 6 8 C 10 12 14 16 18 20

50

100

150

200

250

300 Qult

350

Settlement, S

QL/AE

O1

C1

Figure 1. The Offset Limit Method (after Davisson 1972). In Davissons Offset Limit Method, the predicted failure load value tends to be conservative [3]. The actual limit line can be drawn on the load-settlement curve already before starting the test. The ultimate load can, therefore, be used as an acceptance criterion to proof tested piles in contract specifications. This method is not suitable for tests that involve loading and unloading cycles. 3 PROPOSED DAVISSON MODIFICATION Davissons method can not be applied unless the pile is loaded close to failure. The problem is that in most static load tests where the pile is loaded to one and a half or twice the design load, failure rarely occurs. For this reason it was found necessary to modify Davissons method so as to estimate the ultimate pile load from the test load. For the purpose of this study, results of seventy six axial load tests on different pile types (driven, bored and augered), lengths and diameters were considered as shown in table (1) for one and half the working load and table (2) for twice the working load. Brinch-Hansen, (1961, 1963) considered that the shape of the pile load-settlement curve is a parabolic curve which can be calculated by using the following equation: S=aQ
2

(2) (3)

a = 1/ (C1 S + C2)2

Where S is pile settlement at pile load Q, C1 and C2 are constants, for the same pile load test. At test load equation (2) can be written as follows: STL = a QTL2 (4)

Where STL is pile settlement at pile load test, and QTL is the test pile load. By dividing equation (2) at ultimate load by equation (4) the following relation was obtained: Qult2 = QTL2 [Sult / STL] (5)

Ultimate pile settlement, Sult, was taken according to Davisson (1972), as in equation (1) and pile settlement at test load, STL, can be written in a similar form: STL= (0.08 D+ 3.8) + (QTL L /A E) 2080 (6)

Table 2: Results of Axial Load Tests for Twice the Working Load
Test No Dr1 Dr2 Dr3 Dr4 Dr5 Dr6 Dr7 Dr8 Dr9 Dr10 Dr11 Dr12 Dr13 Dr14 Dr15 B16 B17 B18 B19 B20 B21 B22 B23 B24 B25 B26 F27 F28 F29 F30 F31 F32 F33 F34 L m 23 17 13.6 20 26 19.7 11.75 23.75 12 14.75 17.2 19.5 12 19.5 23.5 25.2 13.5 14 14 15.5 17.5 12.8 17 18 20 16.5 19.6 19.6 19.6 19.6 19.6 19.6 19.6 13.6 D mm 500 430 500 400 450 400 400 400 500 400 400 400 400 450 450 800 500 600 600 600 800 800 800 800 800 600 600 600 600 600 600 600 600 400 Test load (ton) 2 W.L 120 80 95 50 90 60 70 70 34 70 128 80 80 50 90 350 120 130 130 120 350 350 120 340 350 147 210 210 210 210 210 210 210 60 Measured settlement at Test Load (mm) Elastic Plastic Total Sett. Sett. Sett. 1.79 2.65 2.00 2.30 2.26 1.30 0.92 2.82 0.91 1.68 1.17 5.27 1.16 0.49 3.06 2.72 2.16 2.68 2.57 3.21 8.65 5.88 6.65 6.35 3.55 1.67 3.27 2.10 4.14 3.46 2.09 3.34 3.38 0.81 0.55 0.79 0.56 0.59 1.65 0.12 0.71 0.29 0.36 0.63 0.39 2.30 0.25 0.09 1.20 2.27 4.10 2.82 3.10 1.49 5.74 9.30 10.74 5.27 1.81 2.155 0.42 0.47 0.905 6.15 0.52 0.24 0.22 1.1 2.33 3.44 2.56 2.89 3.91 1.41 1.63 3.11 1.26 2.31 1.56 7.57 1.41 0.58 4.26 4.99 6.26 5.50 5.67 4.70 14.39 15.18 17.39 11.62 5.36 3.82

Comparing separately the two terms of equation (6) at test load to the measured plastic and elastic settlement at test load, the results are shown in Figures (2 to 5). It can be seen that the plastic part presented in Figure (2) and the elastic part presented in Figure (3) for test load at one and half the working load, and similarly in Figures (4) and (5) for the case where the test load is twice the working load, equation (6) over estimates the plastic and elastic pile behavior. In order to predict the modified Davissons equation using pile test load, QT.L, in case of the test load one and half, and twice the working load, the following equations were suggested respectively: STL= (0.06 D 1.9) + 0.5 (QT.LL/EA) STL= (0.12 D 4) + 0.85 ((QT.LL/EA) -2) (mm) (mm) (7) (8)

Where: D = pile diameter (cm), QT.L= pile load test at one and half the working load (ton) L = pile length (mm), E = pile Youngs modulus (2 ton/mm2), and A = pile cross section area (mm2). By substituting in equation (5) in case the test load is one and half the working load the ultimate pile load Qult may be calculated using pile test load, QTL, as follows:

Q ult = QTL

[0.08D + 3.8] + Qult L 2

EA QTL L [0.06 D 1.9] + 2 EA

(9)

Similarly, if the test load is twice the working load the ultimate pile load Qult may be calculated using pile test load, QTL, as follows:

Q ult = QTL

EA QTL L [0.12 D 4] + 0.85[ 2] EA

[0.08D + 3.8] + Qult L

(10)

Comparisson Between Suggested, Measured, and Davisson Plastic Behavior 15

3.69 2.57 5.047 9.61 2.61 3.58 3.6 1.91

12 9 6 3 0 0 100
Test Load Q ton

Measured

Suggested
Davisson
Measured trendline
Suggested trendline
Davisson trendline

200

300

400

Dr: driven pile cast in situ. B: bored pile. F: continuous flight augered pile. L: pile length. D: pile diameter.

Figure 2. Plastic behavior for test load one and half the working load.

2081

Comparisson Between Suggested, Meassured, and Davisson Elastic Behavior 15 12

Comparisson Between Suggested and Measured Elastic Settlement 10

Suggested

Suggested Settlement. mm

9 6 3 0 0 100 200 Test Load Q ton 300

M easured
Davisson
Suggested trendline
M easured trendline
Davisson trendline

8 6 4 2 0 0

elastic settlement Plastic settlement

400

Figure 3. Elastic behavior for test load one and half the working load.
Comparisson Between Suggested, Measured, and Davisson Plastic Behavior 15
Pile Settlement S mm

2 Measured Settlement mm 4 6

10

Figure 6. Measured and suggested elastic and plastic pile settlement in case the test load is one and half the working load.
Comparisson Between Suggested and Measured Elastic Settlement 10
Suggested Settlement. mm

12 9 6 3 0 0 100 200 Test Load Q ton 300 400

Measured Suggested Davison Suggested trendline Davisson trendline Measured trendline

8 6 4 2 0 0

Elastic settlement Plastic settlement

Figure 4. Plastic behavior for test load twice the working load.
Comparison Between Suggested, Measured, and Davisson Elastic Behavior 15
Pile Settlem ent S m m

Measured Settlement mm

10

Figure 7. Measured and suggested elastic and plastic pile settlement in case the test load is twice the working load.

12 9 6 3 0 0 100 200 Test Load Q ton 300 400

Suggested Measured Davisson Suggested trendline Measured trendline Davisson trendline

REFERENCES
Canadian Geotechnical Society, 1992. Canadian Foundation Engineering Manual. Davisson, M.T. 1970. Design Pile Capacity. Proc., Conf. on Design and Installation of Pile Foundations and Cellular Structures, Lehigh Univ., Envo Public. Co. pp. 75-85. Davisson, M.T. 1972. High Capacity Piles. Proceedings, Lecture Series, Innovations in Foundation Construction, ASCE, Illinois Section, Chicago, March 22, pp. 81-112. Egyptian Code of Soil Mechanics and Foundation Engineering, 1995. Part 4.

Figure 5. Elastic behavior for test loads twice the working load. A comparison was carried out between measured and suggested elastic and plastic pile settlements as shown in Figure (6) in case the test load is one and half the working load and Figure (7) in case the test load is twice the working load. 4 SUMMARY AND CONCLUSIONS

Davissons method needs the pile to be loaded near to failure to be applicable. Davissons equation when applied for test load it highly over estimated the elastic and plastic settlements. The suggested form of Davissons equation allows the prediction of the pile ultimate load using the pile test load without having to extrapolate the load-settlement curve. Since equations (9 and 10) are equations of the second degree in Qult they can be easily solved to obtain the ultimate pile load.

2082