Ménard Lecture The Pressuremeter Test: Expanding Its Use

th
Proceedings of the 18 International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
Mnard Lecture
The pressuremeter test: Expanding its use
Confrence Mnard
Lessai pressiometrique : largissement de son utilisation
Briaud J.-L.
President of ISSMGE, Professor, Texas A&M University, Zachry Dpt. of Civil Engineering, College Station, Texas, 77843-
3136, USA
ABSTRACT: The purpose of this contribution is to show how the use of the PMT can be expanded further than current practice. The
topics covered in a first part include the amount of soil testing necessary to meet a reliability target, the influence of the lack of tensile
resistance of soils on the PMT modulus, how to recreate the small strain early part of the curve lost by the decompression-
recompression process associated with the preparation of the PMT borehole, best practice for preparing the PMT borehole, commonly
expected values of PMT parameters, the use of the PMT unload-reload modulus, and correlations with other soil parameters. The
second part deals with foundation engineering and includes the use of the entire expansion curve to predict the load settlement
behavior of shallow foundations, the load displacement behavior of deep foundations under horizontal loading, foundation design of
very tall structures, long term creep loading, cyclic loading, and dynamic vehicle impact. Finally an attempt is made to generate
preliminary soil liquefaction curves base on the normalized PMT limit pressure.
RSUM : Le but de cette contribution est de montrer comment lutilisation du PMT peut tre tendu au-del de la pratique courante.
Les sujets abords dans une premire partie comprennent la quantit de reconnaissance de sol ncessaire pour atteindre un objectif de
fiabilit, linfluence de labsence de rsistance des sols la traction sur le module du PMT, comment recrer la partie de la courbe en
petites dformations perdue pendant la dcompression-recompression associe la prparation du trou de forage, les meilleures
pratiques pour la prparation du trou de forage, les valeurs communes des paramtres PMT, lutilisation du module dcharge-
recharge, et des corrlations avec dautres paramtres du sol. La deuxime partie traite des travaux de fondation et les sujets suivants
sont abords: lutilisation de la courbe dexpansion du PMT pour prdire le comportement des fondations superficielles, et le
comportement des fondations profondes sous charge horizontale, la conception des fondations des structures de grande hauteur, le
comportement de fluage, chargement cyclique, et chargement par impact de vhicules. Enfin, on propose des courbes prliminaires de
liqufaction du sol sur la base de la pression limite normalise du PMT.
KEYWORDS: pressuremeter, modulus, limit pressure, shallow foundations, deep foundations, retaining walls, liquefaction.
1 HOW I GOT INTERESTED IN THE PMT? later with the corrected manuscript again rather depressed and
telling Don, there is nothing left for me to do, everything has
The year is 1974 and I am a Master student at the University of been done. Don smiled and told me dont worry, there is much
New Brunswick, Canada working with Arvid Landva. I had more to be done on the PMT; I feel that it is still true today and,
learnt that the triaxial test was the reference test in the in fact, it is the topic of this lecture. So this is my story on the
laboratory. I had also read from Terzaghi that the action was in PMT and I have been a fan of the PMT ever since.
the field. So I sat down one late afternoon and tried to invent an
in situ triaxial test. I drew some complex systems with double 2 SPECIAL THANKS TO LOUIS MENARD
tube samplers and the pressure applied between the two tubes
on an internal membrane. It was very complicated and failed the I met Louis Menard (Fig. 1) on 15 December 1977, one month
Einstein test of optimum simplicity. I had also learnt from many before he died of cancer. I was a PhD student at the University
months behind a drill rig that anything complicated had very of Ottawa in Canada working on my pressuremeter research
little chance of success in the field so I kept searching and with Don Shields. I was coming back home for Christmas that
designing and then it dawned on me. What if I inverted the year and Louis Menard was kind enough to take some time
problem, drew an inside out triaxial test, and applied the from his very busy schedule to visit with me at the Techniques
pressure from inside the tube and pushed outward on the soil. Louis Menard in Longjumeau near Paris. I waited for 30
And so I designed my first pressuremeter. I was very excited minutes but finally got to meet the man who had invented the
about my new invention and could not sleep that night. I waited tool I was so fond of. Around 7 oclock that day, I entered a
anxiously to go to the library the next morning to see what I huge deep office much like you see in castles. At the other end
could dig on this idea. I went to the library and there it was behind a big desk was Louis Menard waving at me to come
Louis Menard 1957, Jean Kerisel as his advisor, the Master in closer and take a seat. I introduced myself and we started to talk
Illinois with Ralph Peck, the development of the design rules, about the pressuremeter. Very quickly, I found myself enjoying
Sols Soils, 1963 and on and on. I came out of the library that the discussion and time flew by. We talked and argued and
morning, very disappointed that my idea had already been talked again and quoted data and theory and reasoning so much
invented. After much reflection that day, I finally decided that I so that at the end we had connected. I was mad because I
should be happy because it was obviously a good idea since it promised myself that I would take notes of what Menard was
had received that much attention. This is how I got interested in saying but in the heat of the action I forgot all about it and was
the pressuremeter. I then went to The University of Ottawa to left with no notes and it was already 8 Oclock. This is where I
work with Don Shields who was connected with Francois got really lucky. Louis Menard asked me: do you have any
Baguelin and Jean Francois Jezequel writing the pressuremeter plans tonight? I said no and he said: why dont you stay for
book. Don gave me the manuscript in early Sept 1976 and said dinner? Whaoh! That would be wonderful. We got up and he
read this and correct any mistake. I did and came back 3 months took his cane to walk from his office to his house which was a
107
1
Proceedings of the 18th International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
th
door away. The cancer was very advanced but he explained to the mean modulus measured on the soil samples has a 98%
me as we walked to the dining room that he had a slight illness confidence level of being within + or 20% of the true mean of
but that he would take care of that in no time! This is where I the modulus?
got my first clue of the remarkable strength of his will power, For this we recall the student t distribution. Consider a large
the steely determination of Louis Menard, a trait of character population (the big cube) of modulus E which is normally
which helped him win against all odds while creating some distributed with a mean p and a standard deviation p. Then
slight antagonistic situations. The dinner was a delight. consider a group of n randomly selected values of the modulus
Honestly, I cannot tell you what I ate but I certainly remember (E1, E2, E3, , En) from the population (results of the site
the stories that he told me with his wife and his children around investigation and testing). The mean modulus value of the group
the table. One stands out in my mind: his first encounter with E1, , En, is g and the standard deviation is g. Lets create
Ralph Peck. He said that he entered Professor Pecks office and many such groups of n modulus values (many options of where
Peck proceeded to explain to young Louis Menard that he to drill and where to test), each time randomly selecting n
would have to take a certain number of core courses to get his values from the larger population of modulus and calculating
Master degree. So Peck walked to the small blackboard in his the mean modulus g of the group. In this fashion we can create
office and wrote a list of these 4 or 5 courses, then went back to a distribution of the means g. It can be shown that the
his desk. Louis Menard got up, took the eraser and wiped the distribution of the means g has a mean g equal to p and a
courses out and said I am not interested in these courses; standard deviation g equal to p/n0.5. If we form the
however I am interested in these courses instead. Menard was normalized variable t:
indeed a very bright, very determined independent thinker. On g p
that day of 15 December 1977 he provided me with a wonderful t (1)
moment in my life, one that I will never forget. g / n
then the distribution of t is the student t distribution for n
degrees of freedom: t(n). The t distribution is more scattered
than the normal distribution of E, depends on the number n of
modulus values collected in each group, and tends towards the
normal distribution when n becomes large (Fig. 2).
Figure 1. Louis Menard (courtesy of Michel Gambin and Kenji

Mori)
3 INTRODUCTION Figure 2. The student t distribution

There are many different types of pressuremeter devices and The properties of the student t distribution together with Eq.1
many ways to insert the pressuremeter probe in to the ground. allow us to write:
This paper is limited to the preboring pressuremeter also called
Menard pressuremeter where a borehole is drilled, the drilling
g g (2)
P g t p g t 1

tool is removed, and the probe is lowered in the open hole. The ,n1 n ,n1 n
2 2
probe diameter is in the range of 50 to 75 mm and the length of
Where t(/2,n-1) is the value of t for n-1 degrees off freedom
the inflatable part of the probe in the range of 0.3 to 0.6 m. The
and a value of /2, is the area under the t distribution for
paper starts with a general observation regarding site
values larger than t (Fig. 3). Eq.2 expresses that there is a (1-)
investigations, then deals with many aspects of the
degree of confidence that the value of p is between the values
pressuremeter practice including the device itself, the
expressed in the parenthesis.
installation, the test, the parameters that can be obtained, and
For our example, we need to determine the number n of
their use in foundation engineering. In each topic, new
modulus values in the group (number of samples to be collected
contributions are made to expand the use of the PMT.
and tested during the site investigation) which will lead to a
high probability P that the predicted modulus (g) will be within
4 HOW MANY BORINGS ARE ENOUGH?
a target tolerance from the true mean modulus of the
What percentage of the total soil volume involved in the soil population (p). Therefore we wish to find the value of n which
response should be tested during the geotechnical investigation. will satisfy the probability equation:
This depends on many factors including the goal of the P g (1 ) p g (1 ) Ptarget (3)
investigation. This goal may be that there is a high probability
that the predictions will be within a target tolerance. As an
example of calculations, assume that the block of soil which
will be loaded by the structure is a cube 10 x 10 x 10 m in size.
Further assume that the goal is to predict the elastic settlement
of the structure with a precision of + or 20% and that the soil
cube has a modulus with a coefficient of variation equal to 0.3.
The question is: what percentage of the total volume of soil
must be tested to have a 98% probability that the predicted
settlement will be within + or - 20% of the true settlement (i.e.:
measured)? Since in this case the modulus is linearly
proportional to the settlement, the question can be rephrased to Figure 3. Definition of the parameter .
read: what percentage of the soil volume must be tested so that
108
Honour Lectures / Confrences honorifiques
th
That is to say we have a Ptarget % degree of confidence that p such anomalies, the amount of soil volume to test would
lies in the range g(1+or-). We can rewrite Eq.3 as increase. If we use the same approach for different volumes we
p g can generate the number of borings necessary to meet the
P Ptarget (4) criterion of 98% confidence of predicting within + or 20% for
g a soil with a coefficient of variation equal to 0.3. Fig. 5 shows

If the coefficient of variation of the population is , then we the number of borings required as a function of the soil volume
assume that the coefficient of variation of the group is also . involved in the response to the loading. The estimated line for
current practice is plotted on the same graph (based on the
p g authors experience) indicating that current practice does not

(5)
p g meet the criterion established. Note that the discrepancy
increases with the size of the project. Indeed the ratio between
Combining Eq.2, 4, and 5 we get.
2
the required number of borings Nr and the current number of
g

2
(6)
borings Nc increases with the size of the imprint.
g t or n t
2
, n 1 n 2 ,n 1
Eq. 6 is solved by iteration since n influences the value of t.
Student t distribution solvers are available on the internet. The
number n represents the number of soil samples to be tested in
order to obtain the value of the modulus within plus or minus
% from the exact answer with a Ptarget probability of success. If
we assume that a triaxial test sample to obtain a modulus value
has a volume of 10-3 m3, then the number n of samples gives the
volume of soil that must be drilled during the investigation to
satisfy the criterion. The percent volume tested becomes
Vs n 10 3
(7)
Vt Vt
In our example the initial volume was 1000 m3, so we can
calculate what percentage of the soil volume should be tested.
Fig. 4 gives the results and indicates that in order to be 98% Figure 5. Comparison of number of borings in current practice
sure that the answer will be within plus or minus 20% from the and number of borings required for a precision of + or - 20%
true value, the amount of sampling is 0.001 percent of the total with a 98% degree of confidence for a soil parameter coefficient
volume. of variation of 0.3.
5 WHAT CAN BE IMPROVED ABOUT THE PMT

EQUIPMENT?
Only a few things, I think. We are at the point of maturity in this
area. If anything, we need to be able to run controlled stress
tests or control strain tests equally well. Controlling strain or
volume has the advantage of not having to guess at the limit
pressure to decide on the pressure steps. Controlling pressure
has the advantage of not having to wait for a long time if the
hole is too big. The devices which control stress require
compressed gas bottles which can be dangerous. Control
volume devices are safer in that respect and still allow control
stress tests. Most civil engineering structures apply stress
control steps.
With regard to the issue of the three cells versus mono-cell
probes, it has been shown (Briaud, 1992) that for probes with a
length to diameter ratio longer than 6, the difference between
the expansion of the mono-cell and the expansion of an
Figure 4. Required volume of soil to be tested as a percent of infinitely long cylinder for an elastic soil are within 5 % of each
the total volume involved in the soil response to predict a soil other. Therefore as long as the probe has a length to diameter
property with a 98% confidence level and within a percent error ratio of 6 or more, there is no need for three cells in a
for given coefficients of variation of the soil property. pressuremeter probe.
The diameter of the probe has an impact on the quality of the
Consider now an 8 story building which is 40 by 40 m at its test for the following reason. The thickness of the ring of
base. The volume of soil involved in the response of the disturbed soil created by the carving or washing process during
building to loading is at least 40 by 40 by 40 m or 64000 m3. drilling is approximately constant regardless of the diameter of
The required sampling is 0.001% or 0.64 m3 which corresponds the drill bit. As such, the larger the pressuremeter diameter is,
to 640 triaxial tests. Further assuming that we will drill 40 m the less influence this disturbed zone will have on the
deep borings allowing us to conduct 20 triaxial tests per boring, pressuremeter curve. Therefore, it is best to increase the
this would require some 32 borings. In practice, we would diameter of the pressuremeter probe. A larger diameter will also
typically drill 4 or 5 borings for such a building. This shows that have a positive impact on the reliability of the borehole
we do not test the soil enough in our current soil investigations diameter as it is much easier to drill a well calibrated 150mm
to meet the set criterion. Note that the assumptions made in the diameter hole than a 50mm diameter hole. Using lightweight yet
student t distribution calculation include the assumption that the rugged 150 mm diameter, 1 m long PMT probes will improve
soil is uniformly variable. In other words, there are no PMT test quality.
heterogeneity trends or anomalies in the soil mass. If there were
109
th
6 MAKING A QUALITY BOREHOLE IS THE MOST r z (9)

IMPORTANT STEP
r z
Er E Ez
This is the most important and the most difficult step in a z (10)
quality pressuremeter test. Much has been tried and written on z
rz r z
Er E Ez
the best way to prepare the hole. Special training is required for
drillers to prepare a good PMT borehole as drilling for PMT Where r, , z are the normal strains in the r, , and z
testing is very different and almost opposite to drilling for soil directions, r, , z are the normal stresses in the r, , and z
sampling (Table 1). Table 2 gives some general directions, Er, E, Ez are the modulus in the r, , and z
recommendations to obtain a quality borehole with wet rotary directions, and r, r, zr, rz, z, z are the Poissons ratios.
drilling which I would recommend in most cases. Because of the symmetry rules, the following equations must
also be satisfied
Table 1. Differences between drilling for PMT testing and Er r E r (11)
drilling for soil sampling E z z E z (12)
DRILLING FOR PMT DRILLING FOR Er zr E z rz (13)

TESTING SAMPLING Here it is assumed that a compression modulus E+ acts in the
Slow rotation to minimize Fast rotation to get to the radial and vertical direction and a much reduced tension
enlargement of borehole diameter sampling depth faster modulus E- acts in the hoop direction.
Care about undisturbed Dont care about borehole (14)
Ez Er E
borehole walls left behind the bit walls left behind the bit
Dont care about soil in front Care about undisturbed soil E E (15)
of the bit in front of the bit +
Where E is the modulus of the soil when tested in compression
Advance borehole beyond and E- is the modulus of the soil when tested in tension. The
testing depth for cuttings to settle Stop at sampling depth problem is further simplified by assuming that
in
Clean borehole by running
rz zr 1 (16)
Do not clean the borehole by
running the bit up and down in
bit with fast mud flow up and z r 2 (17)
down in open hole; avoids
the open hole; this will increase
unwanted cuttings in sampling z r 3 (18)
the hole diameter
tube The plane strain condition of the cylindrical deformation gives
Dont care about borehole
Care about borehole diameter
diameter
z 0 (19)
The definition of the strains is, in small strain theory
Table 2. Recommendations for a quality PMT borehole by the du
wet rotary method. r (20)
dr
Diameter of drilling bit should be equal to the diameter of the probe u
Three wing bit for silts and clays (carving), roller bit for sands and
(21)
r
gravels (washing)
Now the equilibrium equation gives
Diameter of rods should be small enough to allow cuttings to go by
d r r (22)
Slow rotation of the drill (60 rpm) 0
Slow mud circulation to minimize erosion dr r
Drill 1 m past the testing depth for cuttings to settle Using Eq. 8 to 22 leads to the governing differential equation
One pass down and one withdrawal (no cleaning of the hole) where the displacement u is the variable. The boundary
One test at a time conditions are a displacement equal to zero for an infinite radius
and a pressure equal to the imposed pressure at the cavity wall.
7 THE PMT PARAMETERS The solution is a bit cumbersome:
7.1 PMT Modulus and tension in the hoop direction

1
u
ro s12 ( s21 s12 ) ( s21 s12 ) 2 4 s11 s22 o
2
r
(23)
o
A number of parameters are obtained from the PMT. One of the Where s11, s22, s12, s21 are defined as follows
most useful is the PMT modulus Eo from first loading This
modulus is calculated by using the theory of elasticity. One of E 1 22
s11 (24)
the assumptions in elasticity is that the soil has the same
modulus in compression and in tension. This may be true to
1 2 2
2 1 1 1
some extent for clays but unlikely true for sands. When the E 2
s12 (25)
PMT probe expands, the radial stress increases and the hoop
stress decreases to the point where it can reach tension. In 1 2 22 1
elasticity, the increase in radial stress is equal to the decrease in E 2
hoop stress, so if the pressure in the PMT probe is 500 kPa, the s21 (26)
hoop stress at the borehole wall is -500 kPa (neglecting the at 1 2 22 1
rest pressure). The soil is unlikely to be able to resist such E 1 1
tension and using elasticity theory in this case is flawed. The s22 (27)
following derivation shows the influence of having a much 1 2 2
2 1
weaker modulus in tension than in compression. Eq. 23 is to be compared with the equation for the isotropic
The general orthotropic elastic equations are solution which is
(8) E u
r r r zr z ro o o (28)
Er E Ez
1 ro
110
th
Consider the case where the ratio E+/E- = 10, 1 = 3 = 0.33, expansion is defined as the radius at which and r are 1/10th
then 2 equal to 0.033. Then Eq.23 and Eq. 28 give respectively: of the value at the cavity wall, that radius of influence is 100.5Ro
u (29) = 3.16Ro. Within this radius of influence, the average strain
ro 0.309 E o can be calculated as follows
ro
3.16 Ro R
2
1 (33)
uo (30) av o o
dR 0.316 o
ro 0.752 Eo 3.16 Ro Ro Ro R 2
ro
where av is the average hoop strain within the radius of
Therefore, E+ = 2.43Eo (31) influence of the pressuremeter test, o is the hoop strain at the
+ -
This can be repeated for different values of E /E to obtain Fig. wall of the cavity, Ro is the initial radius of the cavity, and R is
6. The inverse of the modulus ratio is consistent with the values the radial distance in the soil. The modulus was mentioned as
recommended by Menard for the values in settlement analysis being associated with a strain level at the cavity wall o
as shown in Fig.6. This observation about the tension in the typically in the range of 2 to 6%; this means that the average
hoop direction also impacts PMT tests in hard soils and rock strain av will be 0.6 to 2%. For the two Texas sites mentioned
which are sound enough to exhibit significant tensile strength. above, the average strain would be close to 1% (3.53% x 0.316).
In this case, the PMT curve shows a break in the expansion Note that this range of strain is consistent with the strain level
curve (Fig. 7) at a pressure p where the hard soil or rock breaks associated with foundation engineering but is much higher than
in tension. This pressure is such that (Briaud, 1992): the range of strain associated with pavement design or
t p 2 oh (32) earthquake shaking where a very low strain modulus is used.
Where t is the soil tensile strength and oh is the horizontal The fact that the small strain modulus is absent from the
stress at rest before the PMT is inserted. beginning of the PMT curve and that the strain range is between
0.6 to 2%, is created in part by the recompression of the soil
which was decompressed horizontally by the drilling process.
This recompression makes the small strain part of the stress
strain curve disappear as shown in the PMT test on Fig. 8. In
this test, an unload-reload loop was performed by decreasing the
pressure to zero and increasing it again to simulate a first
expansion curve. Then a second unload-reload loop was
performed over a much smaller pressure range. This test shows
that the recompression modulus varies tremendously depending
on the extent of the unloading. This test also shows that the low
strain information is lost in the decompression and
recompression loading process. Can we find a way to recreate
the early part of the PMT curve from the information gathered
during the test.
Figure 6. Correction of PMT modulus for low tension soils CONCEPT
ACTUAL TEST
1400
Figure 7. Tensile strength from PMT test 1200

1000
7.2 PMT first load modulus
800
P (kPa)
The PMT first load modulus Eo also called the Menard modulus
is obtained from the initial straight line part of the PMT curve. 600
This straight line exists over a range of relative increase in 400
cavity radius which varies from one soil to another but is
typically in the range of 2 to 6 % relative increase in cavity 200
radius. At two sites in Texas, one in stiff clay the other in dense 0
sand, the average range of 15 PMT tests was 3.47% for the clay 0.00 0.04 0.08 0.12 0.16 0.20
site and 3.59% for the sand site. This refers to the value of dR/R0
R/Ro at the cavity wall. The average radial strain in the soil
mass involved in the response to the cylindrical cavity Figure 8. PMT stress strain curve with unload reload loops
expansion is much smaller and averages 0.316 R/Ro as shown
in the following. The hoop strain and the increase in radial 7.3 PMT modulus at small strain
stress r decrease away from the wall of the cavity at a rate of A soil modulus depends on several factors (Briaud, 2013) one
1/R2 where R is the radial distance into the soil mass (Baguelin of which is the strain level. The PMT curve is a stress strain
et al., 1978). If the radius of influence of the pressuremeter curve where the stress is the radial stress r (measured pressure
111
th
in the PMT) and the strain is the hoop strain (relative increase
in cavity radius). It is therefore possible to define a secant a. REZEROED PMT CURVE
modulus as a function of strain from the PMT curve (Fig. 9).
b. HYPERBOLIC CURVE FITTING
Figure 9. PMT stress strain curve and secant modulus
It can be shown in elasticity that the shear modulus is given

by:
1 ro (34)
G
2 o
If we call Go the shear modulus associated with the straight
portion of the curve, we can normalize the modulus at any strain c . NORMALIZED SECANT SHEAR MODULUS VS
with respect to Go. We calculate the secant shear modulus G1, STRAIN
G2, G3 and so on corresponding to points 1, 2, and 3 on the
pressuremeter curve (Fig. 9). Then we can plot the ratio G1/Go,
G2/Go, G3/Go as a function of the corresponding strain 1, 2,
3. Note that is the strain at the cavity wall but that the mean
strain mean induced in the soil within the zone of influence is
only about 32% of that value (Eq. 33).
The curve linking G/Go vs. mean is shown on Fig. 10c and
10d. From zero strain to the strain value corresponding to the
end of the straight part of the PMT curve (AB on Fig. 10a), the
G/Go vs. mean curve is flat on Fig. 10c and 10d because within
that strain range the modulus G is constant and equal to Go.
In order to generate the non linear beginning of that curve
d . NORMALIZED SECANT SHEAR MODULUS VS
(EB on Fig. 10a), it is convenient to assume a hyperbolic model
as proposed by Baud et al. (2013) of the form LOG OF STRAIN
(35)

1

2Gmax pL
This equation defines a hyperbola which describes the PMT
curve with the limit pressure pL as the asymptotic value and
2Gmax as the initial tangent modulus. The hyperbolic model has
been shown to be very successful in describing the stress strain
curve of soils (Duncan, Chang, 1970). In Eq. 35, pL is known
and all the points on the PMT curve, after excluding the points
on the straight line part, can be used to find the optimum value
of Gmax by best fit regression. This can be done by plotting the
data points as / vs. and fitting a straight line through the data Figure 10. Normalized secant shear modulus vs. strain
points (Fig. 10b). Then 1/2Gmax is the ordinate at = 0 and 1/pL
is the slope of the line. Estimates of Gmax were calculated independently by using
1 (36) correlations proposed by Seed et al. (1986) based on SPT blow
count for sand, Rix and Stokoe (1991) based on CPT point
2Gmax pL
resistance for sand, and Mayne and Rix (1993) based on CPT
Then Eq. 35 gives the complete curve. This technique was used point resistance and void ratio for clays. These estimates of
at two sites, a stiff clay site near Houston, Texas, and a medium Gmax were consistently much higher than the values obtained by
dense sand site in Corpus Christi, Texas. Example results are the hyperbolic extension of the PMT curve; 25 times larger for
presented in Fig. 11 which shows that the data fits well with a the stiff clay and 44 times larger for the dense sand. This
hyperbolic equation. For these two sites, the average ratio indicates that this hyperbolic fit to the PMT curve does not lead
Gmax/Go was 1.75 for the stiff clay and 1.27 for the dense sand. to accurate very small strain moduli.
112
th
a. PMT CURVE STIFF CLAY e. PMT CURVE DENSE SAND
f. HYPERBOLIC CURVE FITTING

b. HYPERBOLIC CURVE FITTING
g. NORMALIZED SECANT SHEAR MODULUS VS STRAIN

c. NORMALIZED SECANT SHEAR MODULUS VS STRAIN
h. NORMALIZED SHEAR MODULUS VS LOG STRAIN

d. NORMALIZED SHEAR MODULUS VS LOG STRAIN
Figure 11. Examples of hyperbolic extension of the PMT curve

(stiff clay, dense sand)
113
th
7.4 PMT modulus long term creep, and cyclic loading a. FOOTING LOAD-SETTLEMENT CURVE
It is relatively easy to maintain the pressure constant during a LOAD (MN)
PMT test while recording the increase in radius of the cavity 0
0 2 4 6 8 10 12
(Fig. 12). A pressure holding step of 10 minutes is not very time

consuming and can lead to very valuable information if the -20
structure will be subjected to long term loading (e.g.: building,
-40
retaining wall). The pressure held for 10 minutes should be
SETTLEMENT (mm)
higher than 0.2pL because below that threshold the influence of -60
the decompression-recompression effect and the disturbance
-80
effect is more pronounced (Briaud, 1992). The evolution of the
secant modulus Et during the pressure holding test is well -100
described by the following model:
n -120
t (37)
Et Eto -140
to
-160
Where t is the time after the start of the pressure holding step, to
is a reference time after the start of the pressure holding step b. FOOTING SETTLEMENT VS TIME CURVE
usually taken as 1 minute, Et and Eto are the secant modulus 0.06
corresponding to t and to respectively, and n is the creep 6.23 MN
LOG DISPLACEMENT LOG10 (S/S1)

exponent. The value of n is obtained as the slope of the plot of 0.05
7.12 MN
log Et/Eto vs. log t/to. The creep exponent n increases with the 8.01 MN
stress applied over strength ratio and depends on the soil type 8.9 MN
0.04 9.79 MN
and stress history. It has been found in the range of 0.01 to 0.03
10.24 MN
for sands and in the rnage of 0.03 to 0.08 for clays (Briaud,
0.03
1992). For clays, the lower values are for overconsolidated
clays while the higher values are for very soft clays.
Measurements on large scale spread footings on an unsaturated 0.02
silty sand (Briaud, Gibbens, 1999) demonstrated that the power

law model works very well (Fig. 13) because the log settlement 0.01
vs. log time curve was remarkably linear. These experiments
also indicated that n increases with the load level but is 0.00
significantly reduced by unload reload cycles. PMT tests with 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
creep steps were performed next to the footings (Fig. 13c and LOG TIME, LOG10(t/t1)
13d); the parallel between the footing and the PMT is striking. c. PMT STRESS VS STRAIN CURVE
a. CREEP TEST
d. PMT MODULUS VS TIME CURVE
b. CYCLIC TEST
Figure 13. Creep response of a 3m by 3m spread footing and a

PMT test (Briaud, Gibbens, 1999, Jeanjean, 1995).
Similarly, one can conduct cyclic loading during the PMT

Figure 12. Creep and cyclic PMT test test. A series of 10 cycles is not very time consuming and can
lead to very valuable information if the structure will be
114
th
subjected to significant repeated loading (e.g.: large wave a. PILE LOAD-DISPLACEMENT CURVE
loading). The evolution of the secant modulus EN to the top of
cycle N is well described by the following model
EN E1 N m (38)
Where N is the number of cycle using number 1 as the first
loading cycle, EN the secant modulus to the top of the Nth cycle,
E1 the secant modulus to the top of the first cycle (first time that
the pressure is decreased), and m is the cyclic exponent. The
value of m is obtained as the slope of the plot of log EN/E1 vs.
log N. Fig. 14 shows a parallel example of a pile subjected to
cyclic horizontal loading and a cyclic PMT test. As can be seen
the power law model of Eq.38 describes the evolution of the
deformation with the number of cycles (straight line on log-log
scales) very well and the parallel between the pile and the PMT
is striking.
7.5 PMT unload-reload modulus

The unload reload modulus Er is obtained by performing an
unload reload loop during the PMT test. The main problem with
Er is that, unlike Eo, it is not precisely defined. Indeed it b. PILE STIFFNESS VS NUMBER OF CYCLES
depends on the strain amplitude over which the loop is CURVE
performed and to a lesser extent on the stress level at which the
loop is performed. As such, Er varies widely from one user to
another and cannot be relied upon for standard calculations
unless the strain amplitude and stress level have been selected to
match the problem at hand. In my practice, I perform an unload
reload loop at the end of the linear phase and unload until the
pressure has reached one half of the peak pressure. This has the
advantage of being consistent but does not necessarily
correspond to a consistent strain amplitude from one test to the
next. I would strongly discourage the use of the reload modulus
because it is not a standard modulus. Instead I would
recommend the use of a hyperbolic extension of the PMT curve
to find the modulus at the right strain level.
c. PMT STRESS STRAIN CURVE
7.6 The yield pressure py.
The yield pressure py is found at the end of the straight line
corresponding to the PMT modulus. Up to py, the amount of
creep is reasonably small but becomes much larger beyond that.
In geotechnical engineering it is always desirable to apply
pressures on the soil below the value of py. Typically py is 0.5
pL for clays and 0.33 pL for sands. Therefore, at working loads,
it is advisable to keep the pressure under foundations at most
equal to 0.5 pL in clays and 0.33 pL in sands to limit creep
deformations.
7.7 Correlations between PMT parameters and other soil

parameters
Correlations based on 426 PMT tests performed at 36 sites in
sand and 44 sites in clay along with other measured soil
parameters were presented by Briaud (1992). These correlations d. PMT MODULUS VS NUMBER OF CYCLES
exhibit significant scatter and should be used with caution. CURVE
Nevertheless they are very useful in preliminary calculations
and for estimate purposes. Table 3 gives the range of expected
PMT limit pressure and modulus in various soils while Tables 4
and 5 give the correlations.
Table 3. Expected values of Eo and PL in soils
CLAY
Soil
Soft Medium Stiff Very Stiff Hard
strength
p*L(kPa) 0200 200400 400800 800-1600 >1600
E0 (MPa) 0 2.5 2.5 - 5.0 5.0 - 12 12 - 25 > 25
SAND
Soil Figure 14. Cyclic response of a laterally loaded pile A and a
Loose Compact Dense Very Dense
strength PMT test (Little, Briaud, 1988).
p*L(kPa) 0 500 500 - 1500 1500-2500 > 2500
E0(MPa) 0 3.5 3.5 - 12 12 22.5 > 22.5
115
th
Table 4. Correlations for Sand (Column A = Number in Table x Table 6. Bearing capacity factors k for in situ tests
Row B)
Strength parameter Clay Sand
Column A = number in table x row B
PMT pL(kPa) 1.25 1.7
B E0 ER p*L qc fs N
A (kPa) (kPa) (kPa) (kPa) (kPa) (bpf) CPT qc(kPa) 0.3 0.2
E0 SPT N(bpf)* 60 75
1 0.125 8 1.15 57.5 383
(kPa)
* Ultimate bearing capacity pu in kPa.
ER
8 1 64 6.25 312.5 2174
(kPa)
p*L 8.2 Load settlement curve method for footings on sand
0.125 0.0156 1 0.11 5.5 47.9 The typical approach in the design of shallow foundations is to
(kPa)
qc calculate the ultimate bearing capacity pu, reduce that pressure
0.87 0.16 9 1 50 436 to a safe pressure psafe by applying a combined load and
(kPa)
fs resistance factor, use that safe pressure to calculate the
0.0174 0.0032 0.182 0.02 1 9.58 corresponding settlement, compare that settlement to the
(kPa)
N allowable settlement, and adjust the footing size until both the
0.0026 0.00046 0.021 0.0021 0.104 1
(bpf) ultimate limit state and the serviceability limit state are satisfied.
In other words the design of shallow foundations defines two
Table 5. Correlations for Clay (Column A = Number in Table x points on the load settlement curve: one for the ultimate load
Row B) and one for the service load. It would be more convenient if the
entire load settlement curve could be generated. Then the
Column A = number in table x row B engineer could decide where, on that curve, the foundation
B E0 ER p*L qc fs su N should operate. This was the incentive to develop the load
A (kPa) (kPa) (kPa) (kPa) (kPa) (kPa) (bpf) settlement curve method (Briaud, 2007).
E0
1 0.278 14 2.5 56 100 667
Five very large spread footings on sand up to 3m x 3m in
(kPa) size were loaded up to 12 MN at the Texas A&M University
ER National Geotechnical Experimentation Site (Fig. 15a).
3.6 1 50 13 260 300 2000
(kPa) Inclinometer casings were installed at the edge of the footings
p*L as part of the instrumentation. They were read at various loads
0.071 0.02 1 0.2 4 7.5 50
(kPa) during the test and indicated that the soil was deforming in a
qc barrel like shape (Fig. 15b). This is the reason why the
0.40 0.077 5 1 20 27 180
(kPa)
pressuremeter curve was thought to be a good candidate to
fs 0.003
(kPa)
0.079
8
0.25 0.05 1 1.6 10.7 generate the load settlement curve for the footing. Note that,
su 0.003 0.62
during these tests, the inclinometers never showed the type of
0.010 0.133 0.037 1 6.7 wedge failure assumed in the general bearing capacity equation.
(kPa) 3 5
N 0.001 0.000 0.005 0.09 It is reasonned that the footings were not pushed to sufficient
0.02 0.14 1 penetration to generate this type of failure mechanism.
(bpf) 5 5 6 1
The transformation required a correspondence principle
8 SHALLOW FOUNDATIONS between a point on the pressuremeter curve and a point on the
footing load settlement curve (Fig. 16). This correspondence
8.1 Ultimate bearing capacity was established on the basis of two equations: the first one
The general bearing capacity equation for a strip footing is: would satisfy average strain compatibility between the two
1 loading processes and the second one would transform the PMT
c ' Nc BN DNq
pu (39) pressure into the footing pressure for corresponding average
2 strains. These equations are:
Where pu is the ultimate bearing pressure, c the effective stress s R
cohesion intercept, the effective unit weight of the soil, Nc, N, 0.24 (41)
and Nq bearing capacity factors depending on the friction angle B Ro
. The assumptions made to develop this equation include that p f f L / B f e f f , d p p (42)
the unit weight and the friction angle of the soil are constant.
Therefore the strength profile of the soil is linearly increasing Where s if the footing settlement, B the footing width, R/Ro
with depth. For strength profiles which do not increase linearly the relative increase in cavity radius in the PMT test, pf the
with depth, this equation does not work and can severely average pressure under the footing for a settlement s, fL/B, fe, f,
overestimate the value of pu. However equations of the f,d the correction factors to take into account the shape of the
following form always take into account the proper soil footing, the eccentricity of the load, the inclination of the load,
strength: and the proximity of a slope respectively, a function of s/B,
and pp the pressuremeter pressure corresponding to R/Ro. The
pu ks D (40) function was originally obtained from the large scale footing
Where k is a bearing capacity factor, s is a strength parameter load tests on sand at Texas A&M University (Jeanjean, 1995,
for the soil, is the unit weight of the soil, and D is the depth of Briaud, 2007) and then supplemented with other load tests. This
embedment. The parameter s can be the PMT limit pressure pL, led to the data shown on Fig. 17. Using all the curves (Fig. 17a),
the CPT point resistance qc, or the SPT blow count N. Table 6 a mean and a design function were obtained (Fig. 17b). The
gives the values of k for various soils and various tests in the design function curve is the mean function curve minus one
case of a horizontal square foundation on horizontal flat ground standard deviation.
under axial vertical load. The f correction factors have been determined through a
series of numerical simulations previously calibrated against the
large scale loading tests (Hossain, 1996, Briaud, 2007). Their
expressions are as follows
116
th
a. LOAD TEST SET UP 0.15

Near 2/1 slope d (49)
f ,d 0.7 1

B
8.5 m
Settlement LVDT Jack Load cell Where B is the width of the footing, L its length, e the load
Beam
eccentricity, the load inclination in degrees, and d the
horizontal distance from the slope-side edge of the footing to
B 0.5B
the slope crest.
SAND Telltates The shape of the function indicates that at larger strain
2B levels the need to correct the PMT curve is minimal. Indeed for
Inclinometer
10.7 m casings s/B larger than 0.03, the mean value of is constant and equal
to about 1.5. For values of s/B smaller than 0.03, there is a need
B 15 m to correct the value of the PMT pressure because of a lack of
Dywidag bars curvature on the PMT curve compared to the curvature on the
only No concrete
footing load settlement curve.
Q
LENGTH = L
CLAY
SHALE d Foundation
7.6 m Drilled shaft D
(Concrete+Bars) e Sand
Steel B
plates Pressuremeter
Test
2.7 m 2.7 m Pressuremeter ?
LOAD
b. PRESSUREMETER-LIKE LATERAL PRESSURE Test
SETTLEMENT
DEFORMATION FROM INCLINOMETER on WALL
Foundation
RELATIVE
Behavior
INCREASE IN
CAVITY RADIUS
Figure 16. Transformation of the pressuremeter curve into the

footing load settlement curve
a. FUNCTION: ALL DATA
Figure 15. Analogy between the soil deformation under a

shallow foundation and around a pressuremeter expansion test
B
Shape f L
/B 0.8 0.2 (43)
b. FUNCTION: DESIGN RECOMMENDATIONS
L
e
Eccentricity f e 1 0.33 center (44)
B
0.5
e
Eccentricity fe 1 edge (45)
B
2
Inclination center (46)
f 1
90
0.5
Inclination edge (47)
f 1
360
0.1
Near 3/1 slope d (48) Figure 17. The function for the load settlement curve method
f ,d 0.8 1

B (Briaud 2013)
117
th
8.3 Load settlement curve method for footings on stiff clay a. PMT CURVE
The load settlement curve method developed for sand was
extended to stiff clay by using some footing load tests and
parallel PMT tests. ONeill and Sheikh (1985) load tested a 2.4
m diameter bored and under-reamed pile in Houston (Fig. 18a).
The pile was 2.4 m deep (relative embedment depth D/B = 1)
and the shaft friction was disabled by a casing. The soil was a
stiff clay with an undrained shear strength of about 100 kPa.
The load was increased in equal load steps and the resulting
load settlement curve is shown in Fig. 18b. At failure, the
average pressure under the footing was 680 kPa as measured by
pressure cells on the bottom of the under-ream. Briaud et al.
(1985) performed pressuremeter tests at the same site around
the same time. The PMT test was carried out at a depth of 3.6 m
or half a diameter below the bottom of the footing; this PMT
curve (Fig.19a) was used to generate the function for that stiff
clay (Fig. 19b). As can be seen, the curve for that stiff clay is
very close to the recommended mean curve for sand. Load tests
on stiff clay using a 0.76m diameter plate at a depth of 1.52m
(Tand, 2013) were also analyzed together with parallel PMT
tests (Briaud, 1985) and gave the other functions on Fig.19b. b. THE FUNCTION
These tests on stiff clay give an indication that the design
function of Fig. 17b is equally applicable to sands and stiff
clays. Note that the load settlement curve method gives the
response of the footing as measured in load tests. These load
tests are carried out in a few hours; if the loading time is very
different (one week or more or one second or less), the time
effect must be considered separately (Section 7.4).
a. LOAD TEST SET UP
Figure 19. Pressuremeter test (Briaud et al, 1985) and

function for stiff clay
b. LOAD TEST RESULTS
9 DEEP FOUNDATIONS UNDER VERTICAL LOADS
The rules developed by the French administration (Fascicule 62,
1993) for calculating the vertical capacity of piles are based on
a very impressive database of load tests carried out by
Bustamante and Gianeselli and the Laboratoires des Ponts et
Chaussees from about 1975 to 1995. These rules were recently
updated (NF P94-262, 2012) and represent one of the most
complete and detailed axial capacity methodology in existence.
These rules should be followed closely as there is no viable
alternative for the PMT.
One area of deep foundations where the pressuremeter has
seen some expanded use is the foundation design of very tall
buildings such as the 452 m high Petronas Towers in Kuala
Lumpur, Malaysia (Baker, 2010), the 828 m high Burj Khalifa
in Dubai, UAE (Poulos 2009), the planned 1000 m high
Figure 18. Large scale footing load test in stiff clay in Houston Nakheel Tower in Dubai, UAE (Haberfield, Paul, 2010), and
(ONeill, Sheikh, 1985) the planned 1000m+ Kingdom Tower in Jeddah, Saudi Arabia
(Poeppel, 2013). It is also seeing increased use for very large
foundations such as the I10/I19 freeway interchange in Tucson,
USA (Samtani, Liu, 2005). The use of the PMT for very tall
buildings started with the work of Clyde Baker between 1965
and 1985 (Baker, 2005) for the Chicago high-rises where the
use of the pressuremeter in the glacial till allowed Clyde Baker
to increase the allowable pressure at the bottom of bored piles
from 1.4 MPa to 2.4 MPa. The 1.4 MPa value was based on
unconfined compression tests; the use of the pressuremeter
along with observations led to using the 2.4 MPa value as
confidence was gained.
118
th
In making settlement calculations for such structures, some

use the rules proposed by Menard and some use the elastic
equations often with an unload-reload modulus. Those who use
the Menard rules, use values based on local experience and
influenced by the ratio between the unload-reload modulus Er
and the first load modulus Eo. While the value of the ratio Eo/Er
varies within a range somewhat similar to the range of values,
it is not clear why one should be related to the other. The ratio
Eo/Er is influenced by the development of plastic deformation
around the probe while the value of is argued to be related to
the combination of lack of strength in tension (hoop direction as
shown in Section 7.1) and recompression process through an S
shape curve (Fig. 8). Those who use the elastic equation
together with an unload-reload modulus face the problem that
the unload reload modulus is ill defined and depends in
particular on the extent of the unloading and the stress level at
which the unloading takes place.
The case of the foundation of the tallest tower on Earth, the
Figure 21. Measured and predicted settlement of the Burj
828m high Burj Khalifa in Dubai, UAE, is studied further to
investigate the issue of the first load modulus and the reload Khalifa, Dubai, UAE (after Poulos, 2009)
modulus (Poulos, 2009). The Burj Khalifa weighs
approximately 5000MN and has a foundation imprint of about 10 DEEP FOUNDATIONS UNDER HORIZONTAL LOADS
3300m2. The foundation is a combined pile raft 3.5 m thick 10.1 Single pile behavior
founded at a depth of about 10 m below ground level on 1.5 m For vertically loaded piles, it is common to calculate the
diameter bored piles extending some 50 m below the raft. To ultimate capacity of the pile due to soil failure and then the
predict the settlement of the tower, a number of methods were settlement at working load. For horizontally loaded piles, an
used including numerical simulations. For these simulations a ultimate load due to soil failure is not usually calculated. Briaud
modulus profile was selected from all soil data available (1997) proposed an equation to calculate the ultimate horizontal
including 40 PMT tests. The PMT first load modulus profile is load due to soil failure for a horizontally loaded pile.
shown in Fig. 20 along with the selected design profile as input
for settlement calculations by numerical simulations. As can be
D v 4 lo for L 3lo
seen the design profile splits the PMT first load modulus profile
with some conservatism. The settlement of the tower was L (50)
predicted to be 77mm; it was measured during construction and 3 D v 3 for L lo
H ou pL BD v
reached 45 mm at the end of construction (Fig. 21). The 4 1/4
reasonable comparison between measured and predicted 4E I
lo p
settlement for this major case history gives an indication that it K
is appropriate to use the PMT first load modulus for settlement
K 2.3Eo
estimates.
Where Hou is the horizontal load corresponding to a horizontal
displacement equal to 0.1B, B the pile diameter, pL the PMT
limit pressure, Dv the depth corresponding to zero shear force
and maximum bending moment, lo the transfer length, L the pile
length, Ep the modulus of the pile material, I the moment of
inertia of the pile around the bending axis, K the soil stiffness,
and Eo the PMT first load modulus.
In order to expand that solution to create the entire load
displacement curve for horizontally loaded piles, it is proposed
to first use a strain compatibility equation such that the relative
displacement to reach the ultimate load on the pile (y/B = 0.1)
corresponds to the relative PMT expansion at the limit pressure
(R/Ro = 0.41).
y R (51)
0.24
B Ro
Then the load on the pile can be transformed into a pressure
within the most contributing zone as
Ho (52)
p pile
BDv
The value is the ratio of the pressure on the pile divided by
the pressure on the PMT for a corresponding set of values of
y/B and R/Ro which satisfy Eq. 51. That way and point by
Figure 20. First load PMT modulus profile and selected design point, the function can be generated as a function of y/B or
modulus values for the Burj Khalifa, Dubai, UAE (after Poulos, 0.24R/Ro. This approach is consistent with the approach taken
2009) for the load settlement curve method for shallow foundations.
This was done for 5 piles including driven and bored piles as
well as sand and clay soils. The piles are described in Briaud
(1997) and in Briaud et al. (1985). They ranged from 0.3 to 1.2
m in diameter and from 6 to 36 m in length. In each case, the
pile dimensions were known, the load displacement curve was
119
th
known and the PMT curves were measured at various depths

within the depth Dv. An average PMT curve was created within
Dv if more than one test was available. The functions obtained
from these load tests and parallel PMT tests are shown in Fig.
22. They have a shape similar to the one for the shallow
foundations but the pile installation seems to make a difference.
The driven piles lead to one class of functions while the bored
pile leads to a lower function. More data would help refine this
first observation.
Figure 23. Plan view of a group of horizontally loaded piles.
Figure 24. Leading pile and trailing pile efficiency factors

Figure 22. The functions for transforming the PMT curve into
a horizontal load displacement curve for a pile.
10.2 Pile group behavior

The behavior of vertically loaded pile groups is often predicted
by making use of an efficiency factor of the form
Qg ev nQs (53)
Where Qg is the vertical load on the group, ev the efficiency of
the vertically loaded group, n the number of piles in the group,
and Qs the vertical load on the single pile for the same
settlement as the pile group. This approach can be extended to
the problem of horizontal loading on a pile group by writing
H g ehnHs (54)
Where Hg is the horizontal load on the group, eh the efficiency
of the horizontally loaded group, n the number of piles in the
group, and Hs the horizontal load on the single pile for the same Figure 25. Ratio of leading over trailing pile efficiency factor
horizontal movement as the pile group. Fig. 23 shows the plan
view of a group of horizontally loaded piles. Eq. 52 was developed based on ultimate load observations at
A distinction is made between the leading piles on the front large horizontal displacements. The use of the same equation for
row of the group and the trailing piles behind the front row. all range of horizontal movements was investigated by
Using data by Cox et al. (1983), Briaud (2013) proposed to comparing measured and predicted movements for two major
extend Eq. 54 to read: pile group experiments by Brown and Reese (1985) in stiff clay
elp and by Morrison and Reese (1986) in medium dense sand. The
Hg
(nlpelp ntpetp )Hs (55)
nlpelp ntp Hs plan view of the group is shown in Fig.23. The piles were
0.273m in diameter, 13.1m long steel pipe piles driven in a 3 by
Where nlp and ntp are the number of leading piles and trailing 3 group with a spacing of 3 diameter center to center. The group
piles in the group respectively, elp and etp are the efficiency was built to simulate a rigid cap condition which is most
factors for the leading pile and trailing pile respectively, and is common. The clay was a stiff clay which had an undrained
the ratio of elp over etp. Fig. 24 and 25 give the efficiency factors shear strength of about 100kPa within the top 3 m from the
as a function of the relative pile spacing based on the data by ground surface. The sand was a medium dense fine sand with a
Cox et al. (1983). CPT point resistance increasing from zero at the ground surface
to 3000 kPa at a depth of 2 m. Fig. 26 presents the result for the
test in clay and Fig. 27 for the test in sand. In each case, the
measured load-displacement curve for the single pile is
presented as well as the measured curve linking the average
load per pile in the group and the group displacement. The
120
th
efficiency in Eq. 55 was calculated as follows using Fig. 24 and yg Bg 1.91

25: 2.65 (58)
elp ys Bs 0.273
0.95
Hg
nlpelp ntp Hs
3 0.95 6 9 0.82 Hs (56)
Hs The curve linking the average load per pile in the group versus
1.25 group displacement was obtained by using the load versus
The predicted curve describing the average horizontal load per displacement curve for the single pile and, for any given
pile in the group versus the group horizontal displacement was horizontal movement, multiplying the single pile movement by
obtained by using the horizontal load versus horizontal 2.65. That predicted curve is shown on Fig. 28 and 29 along
displacement curve for the single pile and multiplying the single with the curve measured by Brown and Reese for their test in
pile load by 0.82 for any given movement. The curve predicted clay (1985) and Morrison and Reese for their test in sand (1986)
using this approach is shown on Fig. 26 (clay) and 27 (sand) respectively. The measured single pile curve is also shown for
along with the measured curves. reference.
Figure 26. Predicted by Cox efficiency factor method and Figure 28. Predicted by ONeill efficiency factor method
measured load-displacement curve for Brown-Reese group and measured load-displacement curve for Brown-Reese
test in clay (1985) group test in clay (1985)
Figure 27. Predicted by Cox efficiency factor method and Figure 29. Predicted by ONeill efficiency factor method
measured load-displacement curve for Morrison-Reese group and measured load-displacement curve for Morrison-Reese
test in sand (1986) group test in sand (1986)
ONeill (1983) suggested that the best and simplest 11 HORIZONTAL IMPACT LOADING FROM VEHICLE
efficiency factor to use for the settlement of a group of
vertically loaded piles was: In the case of road side safety, embassy defense against terrorist
trucks, ship berthing, piles are impacted horizontally. To predict
sg Bg the behavior of piles subjected to horizontal impact, it is
(57)
ss Bs possible to use 4D programs (x, y, z, t) such as LSDYNA
(2006). This is expensive and time consuming. The problem can
Where ss is the settlement of the single pile under the working be simplified by using a P-y curve approach generalized to
load Q, sg the settlement of the group under nQ, n the number of include the effect of time. In this case the governing differential
piles in the group, Bg the width of the group and Bs the width of equation is
the single pile. This efficiency factor for the Brown and Reese
pile group was (Fig. 23)
121
th
4 y 2 y y (59)
EI 4
M C Ky 0
z t 2 t
where E (N/m2) is the modulus of elasticity of the pile, I (m4)
the moment of inertia of the pile against bending around the
horizontal axis perpendicular to impact, y (m) the pile
horizontal displacement at a depth z and a time t, M (kg/m) the
mass per unit length of pile (mass of pile Mp plus mass of
associated soil Ms), C (N.s/m2) the damping of the system per
unit length of pile, and K (N/m2) the soil spring stiffness per
unit length of pile. Note that the soil horizontal resistance is
limited to pu (kN/m2). The boundary conditions are zero
moment and zero shear at the point of impact, and zero moment
and zero shear at the bottom of the pile. The initial condition is
the displacement of the impact node during the first time step;
this displacement is equal to vo x t where vo is the velocity of
the vehicle and t the time step. Other inputs include the mass 1
and velocity of the impacting vehicle, and the parameters in Eq.
59 for the soil and the pile. The differential equation is then
solved by the finite difference method and it turns out that the
parameter matrix is a diagonal matrix so that no inversion is
necessary. As a result the solution can be provided in a simple
Excel spread sheet (Mirdamadi, 2013).
Because the problem is a horizontal load problem on a pile,
the PMT is favored to obtain the soil data. The PMT in this case
is a mini PMT called the Pencel (Fig. 30) which is driven in
place or driven in a predrilled slightly smaller diameter hole if
the soil is hard. As a result of many static and impact horizontal
load tests at various scales (Lim, 2011, Mirdamadi, 2013), the
following recommendations are made for the input parameters.
PL (60)
M s 0.036 B
g 2
C N.s / m 2
240 PL kPa (61)
K 2.3Eo and pu pL (62)

Where B is the pile width, pL the PMT limit pressure, g the
acceleration due to gravity, and Eo the first load PMT modulus.
EQUIPMENT
TEST
Figure 30. Mini pressuremeter test
122
th
parameters shown in Table 7. PMT tests were performed with a

Pencel pressuremeter by first driving a slightly smaller diameter
rod in the very stiff clay and then driving the Pencel probe in
the slightly undersized hole. A comparison between the
measured and calculated behavior of the pile (movement, load,
and time) is presented in Fig. 32. The calculations were based
on the simple Excel program (TAMU-POST, Mirdamadi, 2013)
and a 4D FEM simulation using LS-DYNA (2006). The load
was obtained by measuring the deceleration of the truck by
placing an accelerometer on the bed of the truck and the
movement by using high speed cameras.
Table 7. PMT results by driven Pencel pressuremeter
DEPTH OF TEST MODULUS LIMIT PRESSURE

5
1m 45 MPa 1400 kPa
1.8 m 25 MPa 1200 kPa
a. STATIC TEST: LOAD VS. MOVEMENT

DISPLACEMENT/WIDTH
(/B)
0.00 0.04 0.08 0.12 0.16 0.20
160 12.0
H 0
= u5
0
.
7 0
5 3
x 8
1 =
3 1
0 2
0 9
x .
o.
3
x
.
7
k
N
PRESSURE (P/BDv)
120 9.0
LOAD (kN)
(kPa)
80 6.0
6 40 3.0
0 0.0
0 14 28 42 56 70
DISPLACEMENT (mm)
b. IMPACT TEST: MOVEMENT VS. TIME

1000
800
x DISPLACEMENT (mm)
600
400 Experiment
TAMU-POST (Excel)
7
LS DYNA
200
0
0.00 0.05 0.10 0.15 0.20 0.25
TIME (sec)
c. IMPACT TEST; FORCE VS.TIME

500
Experiment
TAMU-POST (Excel)
400 LS-DYNA
300
LOAD (kN)
200
8
100
Figure 31. Pick-up truck impact test
0
Fig. 31 shows a photo sequence of an impact test where a 2300
kg pick up truck impacted a pile at 97.2 km/h. The pile was a 0.00 0.05 0.10 0.15 0.20 0.25 0.30
steel pipe with a 356mm diameter and a 12.7mm wall thickness. TIME (sec)
It was embedded 2 m into a very stiff clay which gave the PMT
123
th
d. IMPACT TEST: FORCE VS. MOVEMENT b. PMT CHART BASED ON CORRELATION WITH
Static CPT (adapted from Robertson and Wride, 1998)
Experiment
500
TAMU-POST (Excel)
LS-DYNA
400
300
LOAD (kN)
200
100
0 200 400 600 800 1000

x DISPLACEMENT (mm)
Figure 32. Pick-up truck impact test results Figure 33. Preliminary liquefaction charts based on the
pressuremeter limit pressure
12 LIQUEFACTION CHARTS
Liquefaction charts have been proposed over the years to 13 ANALOGY BETWEEN PMT CURVE AND EARTH
predict when coarse grained soils will liquefy. In those charts PRESSURE-DEFLECTION CURVE FOR RETAINING
(Fig. 33), the vertical axis is the cyclic stress ratio CSR defined WALLS
as av / ov where av is the average shear stress generated The load settlement curve method for shallow foundations
during the design earthquake and ov is the vertical effective shows how one can use the PMT curve to predict the load
stress at the depth investigated and at the time of the in situ soil settlement curve of a shallow foundation. This load settlement
test. On the horizontal axis of the charts is the in situ test curve method was extended to the case of horizontally loaded
parameter normalized and corrected for the effective stress level piles. Can a similar idea be extended to the earth pressure versus
in the soil at the time of the test. There is a chart based on the deflection curve for retaining walls? One of the issues is that the
normalized SPT blow count N1-60 (Youd and Idriss, 1997). PMT is a passive pressure type of loading so the potential for
There is another chart based on the normalized CPT point retaining walls may be stronger on the passive side than on the
resistance qc1 (Robertson and Wride, 1998). Using the active side. Another issue is that the PMT test is a cylindrical
correlations in Table 4, it is possible to transform the SPT and expansion while the retaining wall is a plane strain problem.
CPT axes into a normalized PMT limit pressure axis as shown Fig. 34 shows the curves generated by Briaud and Kim (1998)
in Fig. 34. The normalized limit pressure pL1 is based on several anchored wall case histories. The earth
0.5
p pressure coefficient K was obtained as the mean pressure p on
pL1 pL 'a (63) the wall divided by the total vertical stress at the bottom of the
ov wall. The mean pressure p was calculated by dividing the sum
of the lock-off loads of the anchors by the tributary area of wall
Where pL is the PMT limit pressure, pa is the atmospheric
retained by the anchors. For each case history the lock off loads
pressure, and ov is the vertical effective stress at the depth of
were known and the deflection of the wall was measured. Then
the PMT test. Note that the data points on the original charts are
the data was plotted with K on the vertical axis and the
not shown on the PMT chart not to give the impression that
horizontal deflection at the top of the wall divided by the wall
measurements have been made to prove the correctness of the
height on the horizontal axis. The shape of the curve is very
chart. Some degree of confidence can be derived from the fact
similar to the shape of a PMT curve and a transformation
that the two charts give reasonably close boundary lines.
function like the function for the shallow foundation may
Nevertheless, these two charts are very preliminary in nature
exist but this work has not been done.
and must be verified by case histories.
a. PMT CHART BASED ON CORRELATION WITH

SPT (adapted from Youd and Idriss, 1997)
Figure 34. Earth pressure coefficient vs. wall deflection (after

Briaud, Kim, 1998).
124
th
14 CONCLUSIONS 6. Briaud J.-L., 1992, The Pressuremeter, Taylor and

Francis, London, pp.422.
The purpose of this contribution was to show how the use of the 7. Briaud J.-L., 1997, SALLOP: Simple Approach for Lateral
PMT can be expanded further than current practice. In a first Loads on Piles, Journal of Geotechnical and
part, it is shown that more soil testing should take place in Geoenvironmental Engineering, Vol. 123, No.10, ASCE,
geotechnical engineering to reach a reasonable target of Washington, USA.
8. Briaud J.-L., 2007, Spread Footings in Sand: Load
reliability. Then, it is theoretically demonstrated that if the lack Settlement Curve Approach, Journal of Geotechnical and
of tensile resistance of soils is taken into account, the true soil Geoenvironmental Engineering, Vol 133, Issue 8, August
modulus in compression is higher than what is obtained from 2007, ASCE, Reston, Virginia, USA.
conventional PMT data reduction. Then a procedure is 9. Briaud J.-L., 2013, Geotechnical Engineering: unsaturated
investigated to recreate by hyperbolic extension the small strain and saturated soils, John Wiley and Sons, New York,
early part of the curve lost by the decompression-recompression pp.848.
process associated with the preparation of the PMT borehole. 10. Briaud J.-L., Gibbens R., 1999, Behavior of Five Spread
The limitations of that procedure are identified. Best practice for Footings in Sand, Journal of Geotechnical and
preparing the PMT borehole, commonly expected values of Geoenvironmental Engineering, Vol. 125, No.9, pp. 787-
797, September 1999, ASCE, Reston, Virginia.
PMT parameters, and correlations with other soil parameters are 11. Briaud J.-L., Kim N.K., 1998, Beam Column Method for
given. Reasoning is presented against the general use of the Tieback Walls, Journal of Geotechnical and
PMT unload reload modulus. Geoenvironmental Engineering, Vol. 124, No. 1, ASCE,
It is shown that instead of limiting the use of the PMT test Washington, DC.
results to the modulus and the limit pressure, the entire 12. Briaud J.-L., Makarim C.A., Little R., Tucker L., 1985,
expansion curve can be used to predict the load settlement Development of a pressuremeter method for predicting the
behavior of shallow foundations and the load displacement behavior of single piles in clay subjected to cyclic lateral
curve of deep foundations under horizontal loading. Long term loading, Research Report RF5112 to Marathon Oil
creep loading and cyclic loading are addressed. A solution is Company, McClelland Engineers, Raymond International
Builders, Shell Development Company, Dpt of Civil
presented for the design of piles subjected to dynamic vehicle Engineering, Texas A&M University, College Station,
impact. It is also shown how the PMT can be very useful for the Texas, USA, pp214.
foundation design of very tall structures. Finally an attempt is 13. Brown D.A., Reese L.C., 1985, Behavior of a large scale
made to generate preliminary soil liquefaction curves base on pile group subjected to cyclic lateral loading, Report to
the normalized PMT limit pressure. MMS, FHWA, and USAE-WES, Geotechnical Engineering
Center Report GR85-12, Bureau of Engineering Research,
15 ACKNOWLEDGEMENTS Austin, Texas, USA.
14. Cox W.R., Dixon D.A., Murphy B.S., 1983, Lateral load
The author wishes to thank the following individual for tests on 25.4 mm diameter piles in very soft clay in side by
contributing to this paper: Roger Failmezger and Art Stephens side and in line groups, ASTM Special Technical
for sharing some PMT data in sand, Ken Tand for sharing some Publication no. STP 835, pp 122-140.
15. Duncan, J.M., and Chang, C.Y. (1970) Non-linear analysis
plate load test data in stiff clay, Harry Poulos for providing of stress and strain in soils, J. Soil Mech. Founds Div.,
some information on the Burj Khalifa measurements, Chris ASCE, 96(SM5) 1629-1653.
Haberfield for providing some information on the Nakheel 16. Fascicule 62, 1993, Regles techniques de conception et de
Tower design, Clyde Baker for providing some information on calcul des foundations des ouvrages de genie civil,
his experience with the PMT and highrise foundation design. Ministere de lequipement, du logement, et des transports,
Several of my PhD students at Texas A&M University also Publications Eyrolles, Paris, pp182.
contributed to this paper by making computations, preparing 17. Haberfield C.M., Paul D.R., 2010, Footing design of the
figures, formatting the manuscript, and more importantly Nakheel Tower, Dubai, UAE, Proceedings of the Deep
discussing various aspects of the new contributions in this Foundation Conference, February 2010, Dubai, UAE, Deep
Foundation Institute, 18pp.
paper. They are: Alireza Mirdamadi, Ghassan Akrouch, Inwoo 18. Hossain, K. M. 1996. Load settlement curve method for
Jung, Seokhyung Lee. footings in sand at various depths, under eccentric or
inclined loads, and near slopes. Ph.D. thesis, Texas A&M
16 REFERENCES Univ., Dept. of Civil Engineering, College Station, Tx,
USA.
1. Baguelin F., Jezequel J.-F., Shields D.H., 1978, The 19. Jeanjean, P. 1995. Load settlement curve method for
Pressuremeter and Foundation Engineering, Trans Tech spread footings on sand from the pressuremeter test. Ph.D.
Publications, Clausthal-Zellerfeld, W. Germany, 1978. dissertation, Texas A&M Univ., Dept. of Civil Engineering,
2. Baker C.N. Jr., 2010, Uncertain Geotechnical Truth and College Station, Tx, USA.
Cost Effective High-Rise Foundation Design, 2009 20. Lim S.G., 2011, Development of design guidelines for soil
Terzaghi Lecture, in Art of Foundation Engineering embedded post systems using wide flange I-beams to
Practice, Edited by Mohamad H. Hussein; J. Brian contain truck impact, PhD dissertation, Zachry Dpt. of
Anderson; William M. Camp, Geotechnical Special Civil Engineering, Texas A&M University, College Station,
Publications (GSP) 198, ASCE, Washington, USA. Texas, USA, pp394.
3. Baker C.N., 2005, The use of the Menard pressuremeter in 21. Little R.L., Briaud J.-L., 1988, Full scale cyclic lateral load
innovative foundation design from Chicago to Kuala tests on six piles in sand, Miscellaneous paper GL-88-27,
Lumpur, the 2nd Menard Lecture, Proceedings of the 5th US Army Engineer Waterways Experiment Station (now
Int. Symp. on the Pressuremeter ISP5, Paris, France, ERDC), Vicksburg, MS, USA, pp175.
Presses de lENPC. 22. LS-DYNA, 2006, Theory Manual and Users Manual
4. Baud J.-P., Gambin M., Schlosser F., 2013, Courbes version 971, Livermore Software Technology Corporation,
hyperboliques contraintedformation au pressiomtre Livermore, CA.
Mnard autofor, Proceedings of the 18th International 23. Mayne, P. W., and G. J. Rix, 1993, Gmax qc
Conference on Soil Mechanics and Geotechnical Relationships for Clays, Geotechnical Testing Journal,
Engineering, Paris 2013, Presses des Ponts et Chaussees, ASTM, Vol. 16, No. 1, pp. 54-60.
Paris, France. 24. Mirdamadi A., 2013, Deterministic and probabilistic model
5. Briaud J.-L., 1985, Pressuremeter tests at Amoco of single pile under lateral impact, PhD dissertation,
refinery, consulting report to K.E. Tand and Associates, Zachry Dpt. of civil engineering, texas A&M university,
Houston, Texas. College Station, Texas, USA.
125
th
25. Morrison C., Reese L.C., 1986, A lateral load test of a full 32. Robertson, P.K., and Wride, C.E., (1998), Evaluating
scale pile group in sand, Report to MMS, FHWA, and Cyclic Liquefaction Potential using the Cone Penetration
USAE-WES, Geotechnical Engineering Center Report Test, Canadian Geotechnical Journal, Vol. 35, pp. 442-
GR85-12, Bureau of Engineering Research, Austin, Texas, 459.
USA. 33. Samtini N.C., Liu J.-L., 2005, Use of in situ tests to design
26. NF P94-262, 2012, Norme francaise, Justification des drilled shafts in dense and cemented soils, Proceedings of
ouvrages geotechnicques, norme dapplication nationale de the Geo-Institute GeoFrontiers Conference, Austin, Texas,
lEurocode 7, foundations profondes, ISSN 0335-3931, as part of the Mike ONeill Memorial Volume, ASCE,
AFNOR, pp206. Washington, DC, USA, 15pp.
27. ONeill M.W., 1983, Group action in offshore piles, 34. Seed .B., Wong R.T., Idriss I.M., Tokimatsu K., 1986,
ASCE Specialty Conference on Geotechnical Engineering Moduli and damping factors for dynamic analyses of
in Offshore Engineering, Austin, Texas, USA. cohesionless soils, Journal of Geotechnical Engineering,
28. ONeill, M. W., Sheikh, S. A., 1985, Geotechnical ASCE, Vol. 112, GT11, pp1016-1032.
Behavior of Underreams in Pleistocene Clay, Drilled Piers 35. Tand K.E., 2013, Plate load test results at Amoco
and Caissons II, ed. by C. N. Baker, Jr., ASCE, May, pp 57 refinery, Personal communication.
75. 36. Youd, T.L. and Idriss, I.M., (1997). Proceedings of the
29. Poeppel A. R., 2013, Personal Communication, April 2013, NCEER Workshop on Evaluation of Liquefaction
Langan Engineering. Resistance of Soils, Salt Lake City,UT, January 5-6, 1996,
30. Poulos H.G., 2009, Tall buildings and deep foundations Technical Report NCEER-97-0022, National Center for
Middle East challenges, Terzaghi Oration, Proceedings of Earthquake Engineering Research, University at Buffalo.
the 17th International Conference on Soil Mechanics and
Geotechnical Engineering, Alexandria, Egypt, IOS Press
publisher, 3173-3205 pp.
31. Rix, G.J. and Stokoe, K.H. (1991). Correlation of initial
tangent modulus and cone resistance, Int Symp on
Calibration Chamber Testing, Elsevier, New York, pp 351-
362.
126

Ménard Lecture The Pressuremeter Test: Expanding Its Use

Transféré par

Droits d'auteur :

Formats disponibles

Ménard Lecture The Pressuremeter Test: Expanding Its Use

Transféré par

Informations du document

Description originale:

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Ménard Lecture The Pressuremeter Test: Expanding Its Use

Transféré par

Droits d'auteur :

Formats disponibles

th

Figure 1. Louis Menard (courtesy of Michel Gambin and Kenji

3 INTRODUCTION Figure 2. The student t distribution

5 WHAT CAN BE IMPROVED ABOUT THE PMT

6 MAKING A QUALITY BOREHOLE IS THE MOST r z (9)

DRILLING FOR PMT DRILLING FOR Er zr E z rz (13)

7.1 PMT Modulus and tension in the hoop direction

Figure 6. Correction of PMT modulus for low tension soils CONCEPT

Figure 7. Tensile strength from PMT test 1200

b. HYPERBOLIC CURVE FITTING

Figure 9. PMT stress strain curve and secant modulus

It can be shown in elasticity that the shear modulus is given

a. PMT CURVE STIFF CLAY e. PMT CURVE DENSE SAND

f. HYPERBOLIC CURVE FITTING

g. NORMALIZED SECANT SHEAR MODULUS VS STRAIN

h. NORMALIZED SHEAR MODULUS VS LOG STRAIN

Figure 11. Examples of hyperbolic extension of the PMT curve

(Fig. 12). A pressure holding step of 10 minutes is not very time

LOG DISPLACEMENT LOG10 (S/S1)

silty sand (Briaud, Gibbens, 1999) demonstrated that the power

d. PMT MODULUS VS TIME CURVE

Figure 13. Creep response of a 3m by 3m spread footing and a

Similarly, one can conduct cyclic loading during the PMT

7.5 PMT unload-reload modulus

7.7 Correlations between PMT parameters and other soil

Table 3. Expected values of Eo and PL in soils

a. LOAD TEST SET UP 0.15

Figure 16. Transformation of the pressuremeter curve into the

a. FUNCTION: ALL DATA

Figure 15. Analogy between the soil deformation under a

a. LOAD TEST SET UP

Figure 19. Pressuremeter test (Briaud et al, 1985) and

In making settlement calculations for such structures, some

known and the PMT curves were measured at various depths

Figure 23. Plan view of a group of horizontally loaded piles.

Figure 24. Leading pile and trailing pile efficiency factors

10.2 Pile group behavior

efficiency in Eq. 55 was calculated as follows using Fig. 24 and yg Bg 1.91

K 2.3Eo and pu pL (62)

Figure 30. Mini pressuremeter test

parameters shown in Table 7. PMT tests were performed with a

Table 7. PMT results by driven Pencel pressuremeter

DEPTH OF TEST MODULUS LIMIT PRESSURE

a. STATIC TEST: LOAD VS. MOVEMENT

b. IMPACT TEST: MOVEMENT VS. TIME

c. IMPACT TEST; FORCE VS.TIME

0 200 400 600 800 1000

a. PMT CHART BASED ON CORRELATION WITH

Figure 34. Earth pressure coefficient vs. wall deflection (after

14 CONCLUSIONS 6. Briaud J.-L., 1992, The Pressuremeter, Taylor and

Vous aimerez peut-être aussi