Correlation of Initial Tangent Modulus and Cone Penetration Resistance

350

Glenn J. Rix 1 and Kenneth H. Stokoe, II 2

Assistant Professor, School of Civil Engineering, Georgia Institute of Technology, Atlanta, Georgia 30332
12. Dario De Lima, "Development, Fabrication and Verification of the LSU In-situ 2 Brunswick-Abernathy Regents Professor, Department of Civil Engineering, University of Texas at Austin,
Testing Calibration Chamber (LSU /CALCHAS) ," Ph.D. Thesis, Louisiana State Austin, Texas 78712

University, 1990.
ABSTRACT
13. Schmertmann, J .H., " An Updated Correlation Between Relative density, D, and
Fugro Type Electric Cone Bearing, qc," Contract Report DACW 39-76-M6646, Calibration chamber measurements of the cone tip resistance, qc, and resonant column
Waterways Experiment Station, 1976. measurements of the initial tangent shear modulus, Gmax' for washed mortar sand show that
both quantities depend on the vertical and horizontal effective stresses and soil density.
14. Rad, N .S., "Effect of Cementation on Penetration Resistance of Sand", Final report However, Gmax and qc depend on the state of stress and density to different degrees, malting a
submitted to Fugro International, Department of Civil Engineering, Report No.
unique correlation between them impossible. Data from this study are compared to the
GE-84/01, Louisiana State University, June 1984, 222 pages correlation between Gmax and qc proposed by Baldi et al. (1989). All data show a trend of
15. Rad, N.S .and Tumay, M.T., "Factors Affecting Sand Specimen Preparation by decreasing Gmax to qc ratio with increasing normalized cone tip resistance (or relative density),
Raining," Technical Note, Geotechnical Testing Journal, American Society of Test- but values of this ratio differ for different sands. This suggests that there are factors that affect
ing and Materials, Vol.10, No.I, March 1987, pp.31-37. Gmax and/or qc that these correlations do not include. A modified correlation is presented that
indicates the uncertainty in using Gmax-qc correlations for different uncemented, quartz sands.
16. Parkin, A.K.," The Friction Cone Penetrometer: Laboratory Calibration for the
Prediction of Sand Properties," Norwegian Geotechnical Institute, Int. Rep. 52108- INTRODUCTION
5, 1986.
Geotechnical engineers often use penetration resistance from standard penetration tests
17. Bellotti, R., Ghionna, V .N. and Pedroni, S., "Compressibility and Crushability of
(SPT) or cone penetration tests (CPT) to estimate the initial tangent shear modulus, Gm,x·
Sands at Higher Stresses," (this symposium). Ohta and Goto [l] and Imai and Tonouchi [2] have developed correlations between SPT blow
count and Gm,x· Correlations between CPT tip resistance, qc, and Gmax have been developed
by Sykora and Stokoe [3], Robertson and Campanella [4], Rix [5], Baldi et al. [6], Bellotti et
al. [7], Lo Presti and Lai [8], and Baldi et al. [9]. These correlations were created using field
observations, laboratory measurements, calibration chamber measurements, and empirical
estimates of Gm,x and qc.
This paper describes calibration chamber measurements of cone penetration resistance
in washed mortar sand at low values of vertical effective stress encountered in near-surface
cohesionless soils (depths less than 13 m). Cone tip resistances and resonant column
measurements of the initial tangent shear modulus are used to study the relationship between
Gm,x and qc. A case study is also presented of a site in the Imperial Valley of Southern
California that liquefied during a magnitude 6.6 earthquake in 1979. Independent, in situ
measurements of Gmax and qc in three different sand deposits provided an opportunity to
evaluate the accuracy of correlations between cone penetration resistance and initial tangent
shear modulus.

TEST EQUIPMENT AND PROCEDURE

Washed mortar sand has been used extensively for dynamic soil property investigations
at the University of Texas [10]. It was used in the calibration chamber so that cone tip
resistances could be compared with previously measured initial tangent shear moduli for the
same material. A grain size distribution curve for washed mortar sand is shown in Fig. 1 and
index properties are summarized in Table 1. The grain size distribution was checked
periodically during the study to assure that no segregation occurred due to pluviation.
Copyright 1991 by Elsevier Science Publishing Company, Inc.
Calibration Chamber Testing
Editor: A.-B. Huang 351
352
353

100 sand· could be interrupted as needed by a rotating shutter plate. The gap between the rainer
~
I\
~~ and the top of the chamber was enclosed with a polyethylene curtain to control dust. During
~ pluviation, a minimum height of fall of 76 cm was always maintained. Since the capacity of
80 the rainer was insufficient to fill the entire chamber at once, the specimen was constructed in

~
.E Heber Road 15-cm lifts. The average dry density of the medium dense specimens was 16.0 kN/m 3 with a
Cl
·a; Sands standard deviation of 0.20 kN/m 3 • The average relative density was 73.3 percent.
s:
>, 60 Pluviation was also used to construct dense specimens. To increase the density further,
.c the specimen was vibrated during and after raining by a pneumatic vibrator attached to the

~
ai
C: wall of the calibration chamber. The combination of pluviation and vibration yielded an
u: average dry density of 16.7 kN/m 3 with a standard deviation of 0.13 kN/m3. The average
"E 40
\'
Q)
relative density of these specimens was 97.2 percent.
l=
Q)
Washed Mortar
a.. Sand Stress-controlled boundaries were used for all tests in this study. The vertical and
20 horizontal effective stresses were selected to yield one of three values of the effective

\ I'--... ~~ coefficient of lateral earth pressure (K = 0.5, 1.0, and 2.0). Values of vertical effective stress
ranged from 17.2 kPa to 103.4 kPa. This range of vertical effective stress was chosen to
o simulate values encountered in near-surface deposits that may be susceptible to liquefaction.
10 0.1 0.01 0.001
All confining pressures are referenced to the mid-height of the specimen.
Grain Size (mm)
Stress paths used to apply the confining pressures in this study differ from those often
used in other calibration chamber studies (e.g., [11]). In this study, vertical and horizontal
Figure 1 Grain Size Distributions for Washed M01tar Sand and Heber Road Sands
stresses were applied to yield a selected value of K directly. In many other investigations, the
sample is first normally consolidated and then the desired overconsolidation ratio is obtained
Specimens were prepared in a calibration chamber with an inside height of 107 cm and by reducing the vertical stress to simulate mechanical overconsolidation. The residual
an inside diameter of 86 cm. Loose specimens were constructed using a piece of hardware horizontal stress is used to calculate the coefficient of lateral earth pressure. We do not believe
cloth with 1/4-in. (0.64-cm) openings stretched across a 76-cm diameter frame. This frame that the different stress paths used in this study adversely affected the values of cone tip
was placed on the bottom of the calibration chamber and about 15 cm of sand was loosely resistance since the tip resistance is primarily a function of the current state of stress and not of
poured on top of it. Using an overhead crane, the frame was slowly lifted through the layer of stress history [6, 7, 9, 12, 13, 14].
sand. The gentle lifting and redeposition of the sand resulted in a uniform, loose layer. The The cone penetrometer was a standard Fugro-type electrical cone. An MTS system
entire specimen was constructed by repeating the process of placing the sand in layers and was used to push the cone at a rate of 2 emfs. Since the hydraulic cylinder had a maximum
lifting the frame slowly upward through the layer. This method of constructing loose stroke of 61 cm, each penetration test was performed in two stages. During the first stage the
cone was pushed about 46 cm. Penetration was stopped while an extension rod was inserted,
specimens yielded consistent values of density. The average dry density of the five specimens
and the cone was pushed another 46 cm.
constructed in this manner was 14.5 kN/m, which corresponds to an average relative density,
Because specimen construction was time consuming, the first and second penetration
D,, of 12.6 percent. The standard deviation of the dry density was 0.11 kN/m 3 •
stages were often conducted at different confining pressures to maximize the number of
Medium dense specimens were constructed by pluviating air-dried sand into the
penetration tests. Several tests were also performed with the same confining pressures during
calibration chamber. An 86-cm diameter stationary rainer with uniformly spaced, 0.95-cm
both stages. Comparisons between these single-stage tests and the two-stage tests showed that
diameter holes in the bottom was used to control the intensity of flow of the sand. The flow of the two-stage approach was valid [5].
Specimen dei;isities were calculated using the total weight and volume of the specimen.
Table 1 Index Properties of Washed Mortar Sand and Heber Road Sands The uniformity of pluviated specimens was periodically verified using six small containers
placed at various radial and vertical locations during pluviation. As an example, for Test No.
Index Prooertv Washed Mortar Sand Heber Road Sands 2 the six densities ranged from 15.5 kN/m3 to 16.0 kN/m3. The average value of these
Unilled Classification SP SM measurements was 15.75 kN/m3, which agrees with the value obtained by weighing the entire
Median Grain Diameter 0.35mm -0.12 mm chamber (15.83 kN/m 3). Following penetration, sand near the cone was removed and
Percent Fines <1 10 to 22 discarded to avoid reusing crushed particles.
Specific Gravitv 2.67 -
Minimum Void Ratio 0.563 - RELATIONSHIP BETWEEN GMAX AND 4c
Maximum Void Ratio 0.839 -
Grain Shape Subanirnlar to Subrounded Subrounded to Rounded Cone penetration tests performed in calibration chambers are influenced by the
boundaries of the chamber [15]. The effect of the boundaries depends on the ratio of the
354
355

diameters of the calibration chamber and the cone penetrometer and the relative density of the and Gmax should not be expected. Conversely, correlations must include either the density or
sand. Mayne and Kulhawy [16] have summarized data concerning these boundary effects and state of stress explicitly or these parameters must be included implicitly through appropriate

i
have developed the following expression: normalizations of the quantities involved [4, 5].
A correlation that includes the vertical effective stress or relative density explicitly is
B ·0.005 D, shown in Fig. 2. This figure was prepared using an assumed value of the coefficient of lateral
qc (corrected) Bc • 1 earth pressure (K = 0.5) to remove the horizontal effective stress as an independent variable.
[ (1)
qc (measured) - 70 Figure 2 contains two families of curves for different values of the vertical effective stress and
relative density. Solid lines in the figure are for different values of vertical effective stress
where Bc is the diameter of the chamber, B is the cone diameter, and D, is the relative density ranging from 20 kPa to 100 kPa; dashed lines are for different relative densities. We have
of the sand. Equation 1 was used to correct the cone tip resistance. The ratio of the diameters chosen to present the correlation in this manner to emphasize that the relationship between
of the chamber and cone used in this study was 24. A summary of the test conditions and cone tip resistance and initial tangent shear modulus depends on the state of stress and density
of the soil.
measured and corrected cone tip resistances is presented in the Appendix.
An expression commonly used for the relationship between cone tip resistance, relative A more convenient correlation for uncemented, predominantly quartz sands is shown in
density, and effective stress is [6, 17]: Fig. 3 [9]. In this correlation the effective stress and density of the soil are included by
normalizing the cone tip resistance with respect to the square root of the vertical effective
cr\ C1 [cr'ho] C2 stress. Lancellota [21] has shown that normalized cone tip resistance correlates well with
qc = Co Pa [ Pa 0)
Pa exp (C3 D,) (2) relative density. Thus, the ratio of Gmax to qc decreases with increasing density. This
relationship is also apparent in Fig. 2. The influence of the horizontal effective stress on Gmax
where cr\0 is the vertical effective stress, cr'ho is the horizontal effective stress, D, is the relative and qc is shown in Fig. 3 by the lines bounding the cross-hatched area. These lines, for
density expressed as a decimal, and Pa is atmospheric pressure in the same units as cr. C0 , C 1,
C2 and C3 are regression constants. A multiple regression analysis yielded C 0 = 33.5, C 1 = 120000 , - - - - - - - - - - - - - - - - - - - ,
0.22, C2 = 0.64, C3 = 2.49, and a coefficient of determination (r2) equal to 0.91. Dr= 100%
Resonant column tests were used to measure the initial tangent shear moduli of washed ,,•'a'v=
mortar sand [10]. Isotropic, biaxial, and triaxial confinement of solid and hollow cylindrical cu' _,,/ 100 kPa
specimens were used to evaluate the effect of state of stress on the initial tangent shear
modulus. The tests confirmed earlier research showing that the initial tangent shear modulus
is a function of the individual principal stresses instead of the mean effective stress [18, 19].
An equation similar in form to that originally proposed by Hardin [20] but modified to include
-
~ 100000
en
:::,
:i
'C
the individual principal stresses was used to express the results:
~ 80000
A [er' ) na [cr' ) nb «; ,,/ a'v = 40 kPa
Gmax = 0.3 + 0.7e2 p: p: Pa (3) Q)
.c
(f)

where Gmax is the initial tangent shear modulus, e is the void ratio, cr'a is the effective stress in 55 60000 '
C, 1/!,' a'v = 20 kPa
the direction of wave propagation, cr'b is the effective stress in the direction of particle motion, ,',, ,
C
',
,,
and Pa is the atmospheric pressure in the same units as Gmax' cr'., and cr'b· Regression analyses {!:. ,'
yielded A = 685 and na = nb = 0.22 for washed mortar sand. Equation 3 was used to calculate
~
'
values of Gmax for each combination of density and state of stress in the 42 cone penetration 40000 Washed Mortar Sand
tests. The values of cr'a and cr'b were assumed to be the vertical and horizontal stresses, .s
respectively. Values of Gmax are included in the Appendix. K= 0.5
Equations 2 and 3 show that qc and Gmax are dependent on the vertical and horizontal
stresses and the density. Since both depend on the same parameters, it is reasonable to expect 20000 ._______.____.,__..i...._.___.._.....______.____.,__..i........
qc and Gmax to be correlated. However, it is important to note that qc and Gmax depend on the 0 5000 10000 15000 20000 25000 30000
state of stress and density to different degrees. The sum of the exponents C 1 and C2 in Eq. 2 Cone Tip Resistance (kPa)
differs from the sum of the exponents na and nb in Eq. 3. Furthermore, qc and Gmax depend on
the relative density (or void ratio) differently. This means that a unique correlation between qc Figure 2 Correlation Between Initial Tangent Shear Modulus and Cone Tip Resistance for
Washed Mortar Sand
 357
356

50
26
Gmax
I •
Gmax
qc
24 qc
• Washed Mortar Sand

22 40 ~ Baldi et al., {1989)

•
~
20
Gmax} I
qc in kPa I Gmax}
18

16
a'v
I

30
••
•
qc
a' V
in kPa

14

12

10
20
I ~
-•
8

6 10
4

2
0 .___ _...__....___ __._...._..........._____....___...__ _,
0 2000 3000
200 500 1000 100 200 300 500 1,000 2,000 3,000 5,000
qc qc
~ ~
Figure 3 Correlation Between Gmax and qc Proposed by Baldi et al. [9] Figure 4 Comparison of Washed Mortar Sand Data with Correlation Proposed by Baldi et al.
[9]
overconsolidation ratios of 1 and 10, account for the residual horizontal stresses due to
overconsolidation.
Jamiolkowski and Robertson [22] have plotted data from three cohesionless soil HEBER ROAD CASE STUDY
deposits in this normalized form. These deposits were slightly silty to silty sands and gravelly
sands. All were normally to lightly overconsolidated (OCR < 2), uncemented, and less than On October 15, 1979 an earthquake (M = 6.6) occurred along the Imperial Fault in
3000 years old (Late Holocene)[8]. The data from these three sites plot within the cross- Southern California that resulted in extensive liquefaction of near-surface sands. The sands
hatched area shown in Fig. 3 except for several points from the gravelly sand site that may that liquefied were confined to a remnant stream channel passing through the area [23]. In
January, 1981, University of Texas and US Geological Survey personnel performed crosshole
have been affected by the presence of gravel.
The data from this study are plotted in normalized form in Fig. 4. Also shown in the tests and cone penetration tests to evaluate the characteristics of sands that did and did not li l
figure is the cross-hatched area from Fig. 3. The washed mortar sand data have the same trend liquefy. Because initial tangent moduli and cone tip resistances were independently measured,
of decreasing ratio of Gmax to qc for increasing relative density. For normalized penetration this case study provides an opportunity to evaluate the accuracy of laboratory-based
resistances less than 1000, the washed mortar sand data plot above the range suggested by correlations between Gmax and qc when applied to field situations.
Baldi et al. [9]. This means that for sands with low relative densities the correlation Three sands with different depositional histories are present in the area of the remnant
underpredicts Gmax by 30 to 40 percent, depending on the value of the normalized tip stream channel. A cross section of the site is shown in Fig. 5. The point bar (Unit A 1),
lli''
resistance. channel fill (Unit Ai,), and levee ( Unit A3) deposits are Late Holocene in age. The different
depositional histories have resulted in different densities. The point bar deposit is the densest
~·w
!!~
of the three, followed by the levee deposit and the channel fill. Sand boils and a lateral spread ~\\~~
occurred in the channel fill following the 1979 earthquake. No surficial evidence of lJ?~
358 359

25
......,. WEST EAST,.__ Heber Road Sands
Gmax
0

?\~
Art iliciall Fill (Sand) qc
• Point Bar
20, ■ Channel Fill
~· Sand A i · • ~•-•' • . Sandi A2 ■

::
tO f-,(Po ,nl Bari.-'·;-'; '.••.. :. : • • .
t~\:/,::\;~~-~-~-~;-~ir(~i;;;: ~- • •
• (Cha nnel Fill) • .
... Levee

:c
I- -:. -- - ~illy 1cIal'.. B _ ;
15 ~ Baldi et al., (1989)
ll. 20
w •. ··.::-_:. _.·. >·, · .-· ·--·-_'..•. ·_'._·._'.: •.. :. Sand, C . } :_ _: .: ,'·. ·· ,:-:~-::. : . •. ·.';; , ·•• :; • : - ••.• ■
a
1-.-.-.·.. .-;,.·_-:.sa□d-~~ ~-:_I -_ Clay D • Gmax}
- :. - - -: .__.· -~-_.-_·.': :-: : •: ·::~_and: <-.··: •:-\•:.:::-·:·_-.1 qc in kPa
30
10 a' V
.filX.
Horizontal
Location of CPT soundings

0 50
Scale :

100 150
1 with depth indicated
-~
~
I Stratigraphy determined
Feet I
from split-barrel samples
i
5
...
Figure 5 Cross Section of Heber Road Site Showing Point Bar, Channel Fill, and Levee
Deposits [24]
0 ....__......__ _................................__ _ _....__ __,__ _....J
200 300 500 1,000 2,000 3,000 5,000
liquefaction was observed in the point bar and levee deposits. The range of grain size
distributions and index properties of the three sands are given in Fig. 1 and Table 1, qc
respectively. The three sands are similar except for the fines content that ranged from 22 ~
percent (channel fill) to 10 percent (point bar). No evidence of cementation was observed in
any of the three sands.
Measured values of initial tangent shear modulus and cone tip resistance were used to Figure 6 Comparison of Heber Road Data with Correlation Proposed by Baldi et al. [9]
calculate the ratio of Gmax to q0 and the normalized tip resistance. Vertical effective stresses
were calculated using assumed total unit weights of 15.7 kN/m3 above the water table and 17.3 normalized cone tip resistances less than 1000. In situ seismic measurements and cone
kN/m3 below. The depth of the observed water level was 1.8 m. Data for the three Heber penetration tests at the Heber Road site provided another opportunity to evaluate the accuracy
Road sands are plotted in Fig. 6 with the correlation suggested by Baldi et al. [9] . Here the of the suggested correlation. For this site, the observed ratios of Gmax to q0 are lower than the
data plot below the correlation. This may be the result of the additional fines in these sands. predicted ratios. It appears that for an individual sand, the ratio of Gmax to q0 as a function of
If the correlation were used to estimate the initial tangent shear modulus at the Heber Road the normalized cone tip resistance is well defined, but the values of this ratio differ from sand
site, the values would be overpredicted by 30 to 40 percent. to sand. This suggests that there may be factors such as fines content, particle angularity, and
particle mineralogy that influence Gmax and/or q0 that are not included in these correlations
CONCLUSIONS between Gm./q0 and the normalized tip resistance. Until these factors can be identified, it may
be more appropriate to use a correlation between Gmax and q0 that indicates the uncertainty in
Calibration chamber studies of cone tip resistance and resonant column tests on washed
using the correlation for different sands.
mortar sand show that both Gmax and q0 depend on the current state of effective stress and the
Figure 7 shows a proposed correlation similar to that of Baldi et al. [9] that includes
density of the soil. On this basis, it seems reasonable to expect a correlation between the two
data from this study on washed mortar sand and the Heber Road sands. The average value is
quantities. It is important to note, however, that Gmax and q0 depend differently on the
given by:
effective stresses and density. This means that a unique correlation between Gmax and q0 is not
possible. Correlations must either include the effective stresses or density explicitly or account
for their effect by appropriate normalizations of Gmax or q0 •
When data from this study on washed mortar sand are plotted in the form suggested by
G;.•x = 1634 [~J o.1s (4)

Baldi et al. [9], the observed ratios of Gmax to q0 are higher than the predicted ratios for
 c•..F

360 361

50 r---------------------, 5. Rix, G.J., (1984), "Correlation of elastic moduli and cone penetration resistance,"
Thesis submitted to the University of Texas at Austin in partial fulfillment of the
Uncemented Quartz Sands
Gmax requirements for the degree of Master of Science.
qc
Gmax] =1634 [,Jk) -0.
75 6. Baldi, G., Bellotti, R., Ghionna, V., Jamiolkowski, M., and Pasqualini, E., (1986),
"Interpretation of CPTs and CPTUs - Part Il: Drained penetration in sands,"
40 [ Proceedings, Fourth International Geotechnical Seminar on Field Instrumentation and
qc ave
In Situ Measurements, Singapore.
7. Bellotti, R., Ghionna, V., Jamiolkowski, M., Lancellota, R., and Manfredini, G.,
Range = Average :t: Average (1986), "Deformation characteristics of cohesionless soils from in situ tests," Use of In
2 Situ Tests in Geotechnical Engineering, Geotechnical Special Publication No. 6, S.P.
30 Clemence, Ed., ASCE, pp. 47-73.
8. Lo Presti, D.C.F., and Lai, C., (1989), "Shear wave velocity from penetration tests,"
Gmax}
Qc in kPa Atti del Dipartirnento di Ingegneria Stutturale No. 21, Politecnico di Torino, Torino,
a' V Italy.
9. Baldi, G., Bellotti, R., Ghionna, V., Jamiolkowski, M., and Lo Presti, D.C.F., (1989),
20
"Modulus of sands from CPTs and DMTs," Proceedings, XII ICSMFE, Rio de Janeiro,
Vol. 1, pp. 165-170.
10. Ni, S.-H., (1987), "Dynamic properties of sand under true triaxial stress states from
resonant column/torsional shear tests," dissertation submitted to the University of
10 Texas at Austin in partial fulfillment of the requirements for the degree of Doctor of
Philosophy.
11. Bellotti, R., Bizzi, G., and Ghionna, V.N., (1982), "Design, construction, and use of a
calibration chamber," Proceedings, Second European Symposium on Penetration
Testing, Amsterdam.
0 L - - - . . . . 1 . - . . . . 1 . - - . L . . ......._,_"-'-----'----'---.....1 12. Lambrechts, J.R., and Leonard, G.A., (1978), "Effects of stress history on deformation
100 200 300 500 1,000 2,000 3,000 5,000 of sand," Journal of the Geotechnical Engineering Division, ASCE, Vol. 104, No. 11,
qc pp. 1371-1387.
_g 13. Clayton, C.R.I., Hababa, M.B., and Simons, N.E., (1985), "Dynamic penetration
resistance and the prediction of the compressibility of a fine sand - a laboratory study,"
Geotechnique, Vol. 24, No. 1, pp. 19-31.
14. Jamiolkowski, M., Ghionna, V., Lancellota, R., and Pasqualini, E., (1988), "New
Figure 7 Proposed Modified Correlation Between Gmax and qc for Uncemented, Quartz Sands correlations of penetration tests for design practice, " Penetration Testing 1988, I. de
Showing Average and Range in Values Ruiter, Ed., Balkema, Rotterdam.
15. Parkin, A.K., and Lunne, T., (1982), "Boundary effects in the laboratory calibration of
a cone penetrometer for sand," Proceedings, Second European Symposium on
where Gmax' qc, and cr". are in kPa. The range in the ratio of Gmax .to qc for a given value of the Penetration Testing, Amsterdam.
normalized penetration resistance is about ± 50 percent. 16. Mayne, P.W., and Kulhawy, F.H. , (1991), "Calibration chamber database and
boundary effects correction for CPT data, " Proceedings, First International Symposium
REFERENCES on Calibration Chamber Testing, Potsdam, New York.
17. Schmertmann, J.H., (1976), "An updated correlation between relative density, D,, and 111 I
1. Ohta, Y., and Goto, N., (1978), "Empirical shear wave velocity equations in terms of Fugro-type electric cone bearing, qc," US Army Waterways Experiment Station Report 1

characteristic indexes," Earthquake Engineering and Structural Dynamics, Vol. 6, pp. DACW 39-76-M6646. 111111
167-187. 18. Roessler, S.K., (1979), "Anisotropic shear modulus due to stress anisotropy," Journal I ii
2. Imai, T., and Tonouchi, K., (1982), "Correlation of N-values with S-wave velocity,"
of the Geotechnical Engineering Division, ASCE, Vol. 105, No. 7, pp. 871-880. ,11:
Proceedings, Second European Symposium on Penetration Testing, Amsterdam, Vol. 19. Stokoe, K.H., Il, Lee, S.H.H., and Knox, D.P., (1985), "Shear moduli measurements
111

2, pp. 67-72. under true triaxial stresses," Advances in the Art of Testing Soils under Cyclic
3. Sykora, D.W., and Stakoe, K.H., 11, (1983), "Correlations of in situ measurements in Conditions, ASCE, pp. 166-185.
sands with shear wave velocity," Geotechnical Engineering Report GR83-33, 20. Hardin, B.O., (1978), "The nature of stress-strain behavior for soils," Earthquake
University of Texas at Austin, Austin, Texas. Engineering and Soil Dynamics, ASCE, Vol. 1, pp. 3-90.
4. Robertson, P.K., and Campanella, R.G., (1983), Interpretation of cone penetration tests 21. Lancellota, R., (1983), "Analisi di affidibilita in ingegneria geotecnica," Atti del
- Part I: Sands," Canadian Geotechnical Journal, Vol. 20, No. 4, pp. 718-733. Instituto Scienza Construzioni No. 625, Politecnico di Torino, Torino, Italy.
362
CHARACTERIZATION OF SOIL IN CALIBRATION CHAMBERS WITH
SEISMIC WAVES

22. Jamiolkowski, M., and Robertson, P.K., (1988), "Future trends for penetration testing,"
Proceedings, Penetration Testing in the UK, Birmingham, pp. 321-342. KENNETHH. STOKOE, 11, 1 JAMES NGAR-KOKLEE, 2
23. Youd, T.L., and Bennett, M.J., (1983), "Liquefaction sites, Imperial Valley, AND SHANNON HSIEN-HENG LEE 3
'Brunswick-Abernathy Regen!S Prof., Univ. of Texas, Dept. of Civil Engrg., Austin, TX 78712-1076
California," Journal of the Geotechnical Engineering Division, ASCE, Vol. 109, No. 3.
2Graduate Research Asst., Univ. of Texas, Dept. of Civil Engrg., Austin, Texas 78712-1076
24. Sykora, D.W., and Stokoe, K.H., II, (1982), "Seismic investigation of three Heber 3
Associate Prof., Natl. Taiwan Inst. of Technology, Dept. of Construction Engrg., Taipei 10772, Taiwan, ROC
Road sites after October 15, 1979 Imperial Valley earthquake," Geotechnical
Engineering Report GR82-24, University of Texas at Austin, Austin, Texas.
ABSTRACT
APPENDIX - SUMMARY OF TESTS
Seismic measurements can be performed in soil confined in calibration chambers to
evaluate the small-strain shear and constrained moduli of the soil and the variation of moduli
Test Vertical Horizontal Coeff. of Void Relative Uncorrected Corrected Initial
Lateral Ratio Cone Tip Tangent
with direction. Geophones placed in the soil during sample construction can be used as seismic
No, Effective Effective Density Cone Tip
Stress Stress Earth (%) Resistance Resistance Shear sources and receivers. Typical measurements in dry sand confined in a 2.1-m cubical chamber
(JcPa) (JcPa) Pressure (JcPa) (JcPa) Modulus are used to illustrate this approach. Inherent and stress-induced anisotropy in the sand are
Mal clearly identified with these measurements. A cross-anisotropic model is used to represent the
I 75,8 68.9 0.9 0.748 33.0 6953 8353 86497 sand under both isotropic loading and biaxial loading with equal principal horizontal stresses. In
2 24.1 17.2 0.7 0.642 71.4 2805 4172 58246 both loading cases, the horizontal plane is a plane of isotropy. The seismic measurements
3 34.S 34.5 1.0 0.612 82.2 6335 10012 768 16 required to evaluate the five independent elastic constants in a cross-anisotropic model are
4 17.2 17.2 1.0 0.612 82.2 1893 299 1 56624 discussed.
s 68.9 34.S 0.5 0.652 67.8 4893 7135 84 171
6 68.9 34.S 0.5 0.631 75.4 741 4 11276 86914
0.5 72.1 63193
INTRODUCTION
7 34.S 17.2 0.640 2255 3369
8 68,9 68.9 1.0 0.582 93.1 11062 18572 109074
9 103.4 51.7 0.5 0.637 73.2 8022 12055 102942 Many calibration chambers (CC) are in use today. The chambers are generally used to
10 34.5 17.2 0.5 0.596 88.0 7629 12453 67578 load carefully constructed samples of soil with known boundary conditions. Models or
II 68.9 34.5 0.5 0.547 105.8 16396 29542 98732 prototype systems are then inserted into the soil to evaluate their performance in a known
12 17.2 17.2 1.0 0.548 105.4 5413 9734 62391 geotechnical environment. Evaluation of the performance of in situ devices are typical uses of
13 34.5 34.5 1.0 0.548 105.4 10807 19432 84640 calibration chambers, with the most well known results in the U.S . probably the Gibbs and
14 34,5 17.2 0.5 0.562 100.4 10493 18342 71152 Holtz (1957) SPT relationships and the Schmertrnann (1976) CPT relationships.
IS 103.4 51.7 0.5 0.562 100.4 16838 2943 4 115378
16 17.2 17.2 1.0 0.610 83.0 5246 8325 56797
17 34.5 34.5 1.0 0.610 83.0 9610 15250 77051
Most calibrations chambers have the shape of a right circular cylinder and thu s can be
18 34.5 17.2 0.5 0.786 19.2 1177 1309 50620
used to load soil isotropically (cr1 = cr2 = cr3) or biaxially Ccr1 > cr2 = cr3 or cr1 = cr2 > cr3). There
19 68 ,9 34.5 0.5 0.786 19.2 2020 2248 68671 are a few cubical calibration chambers in use. Cubical chambers have the advantage of being
20 17.2 17.2 1.0 0.788 18.5 824 913 43330 able to load soil under true triaxial (cr1 > cr2 > <J3) states of stress as well as isotropically and
21 34.5 34.5 1.0 0.788 18.5 1647 1826 58782 biaxially. Generally, stress controlled boundaries are used, although a few devices also have
22 34.5 68,9 2.0 0.814 9.1 3285 3455 65851 strain controlled boundaries (B aldi et al, 1981; and Parkin and Lunne, 1982). The soil most
23 68.9 68.9 1.0 0.814 9.1 4305 4528 76699 often used in CC testing is a uniform sand which is placed by dry pluviation.
24 103.4 51.7 0.5 0.780 21.4 2295 2585 82827
25 103.4 103.4 1.0 0.780 21.4 4138 4661 96471 The characteristics of soil in calibration chambers are generally not evaluated in place at
26 17.2 34.5 2.0 0.787 18.8 1912 2124 50544
any time during testing. An approach based on seismic measurements of body waves,
27 51.7 103.4 2.0 0.787 18.8 4589 5097 81960
28 17.2 34.S 2.0 0.634 74,3 5266 7962 63775
compression and shear waves, is presented herein. The seismic measurements can be performed
29 34.5 68.9 2.0 0.634 74.3 10669 16131 86517
in the soil at any time during CC testing. The measurements are used to determine the initial
30 68.9 34.5 0.5 0.687 55.1 7796 10592 79791 tangent shear and constrained moduli of the soil and the variation of moduli with direction in the
31 68 ,9 68.9 1.0 0.687 55.1 9218 12524 92935 soil specimen. These measurements permit one to know the small-strain stiffness of the soil and
32 17.2 34.5 2.0 0.553 103.6 8522 15170 72124 to characterize the anisotropic nature of the soil. It is possible to identify two types of anisotropy
33 34.5 68.9 2.0 0.553 103.6 13543 24107 97844 with body waves. The fi rst is inherent (or structural) anisotropy which refers to a physical
34 17.2 17.2 1.0 0.607 84.1 3138 5010 57057 characteristic of the material that is independent of the applied stress. If the material is isotropic
35 34.S 34.5 1.0 0.607 84.1 6158 9832 77404 and under an isotropic stress condition, then the stiffness should be independent of direction. If
36 34.5 17.2 0.5 0.613 81.9 4050 63 88 65851 the material is anisotropic and under an isotropic stress condition, then the stiffness should
37 103,4 51.7 0.5 0.613 81.9 8483 13379 106782
2.0 0.636
depend on direction. Inherent anisotropy is normally associated with the depositional process
38 51.7 103.4 73.6 10493 15800 103100
39 103.4 103.4 1.0 0.636 73.6 11768 17720 120084
and grain characteristics (Arthur and Menzies, 1972; and Oda, 1972). The other type of
40 68.9 68.9 1.0 0.566 98.9 16053 27837 111749 anisotropy is due to the applied stress state and is called stress-induced anisotropy. A biaxial (cr 1
41 103.4 103.4 1.0 0.566 98.9 22437 38907 133575 = cr2 > cr3 , or cr1 > cr2 = <J3) or true triaxial (cr1 > cr2 > <J3) stress condition is associated with
42 51.7 103.4 2.0 0.578 94,6 17416 29479 11261 8 stress-induced anisotropy.
The purposes of this paper are: I. to introduce seismic testin g with body waves to
evaluate the small-strain stiffness and the anisotropic character of sand, and 2. to introduce a

Copyright I 991 by Elsevier Science Publishing Company, Inc.

Calibration Chamber Testing
Editor: A.-B. Huang 363