Calibration of Cone Penetration Test in Sand: H - H HSU A - B Huang
Calibration of Cone Penetration Test in Sand: H - H HSU A - B Huang
Calibration of Cone Penetration Test in Sand: H - H HSU A - B Huang
579-590
A N -B IN HUANG
Department of Civil Engineering National Chiao-Tung University Hsinchu, Taiwan, R.O.C. (Received October 1, 1998; Accepted April 21, 1999) ABSTRACT Due to its simplicity, the cone penetration test (CPT) is a popular in-situ testing method. CPT is especially desirable in characterizing sand where it is difficult to obtain undisturbed samples. Because of the large strain which occurs during a cone penetration, theoretical analysis of CPT results has been difficult and, hence, limited. Accordingly, interpretation of CPT data is mostly based on empirical correlations. Some of the empirical correlations are based on CPT in the calibration chambers. An important drawback of interpreting CPT in a calibration chamber is its boundary effects. Correction factors have been proposed to account for these boundary effects. However, the validity of the use of correction factors and the mechanisms of boundary effects have not been independently verified. The authors have developed an axisymmetric field simulator in which CPT calibration tests can be conducted under substantially reduced boundary effects. A series of CPT calibration tests has been performed in the new simulator system to correlate the cone tip resistance ( q c ) with the stress state. Results show that the correlation of q c and the initial effective mean normal stress (prior to cone penetration) is clearer than the other components of the initial stress state. However, under the same initial effective mean normal stress, qc has a consistent localized relationship with the initial horizontal stress. q c is affected by the horizontal stress near the cone tip, where there exists an obvious correlation between them. This paper introduces chamber calibration tests of CPT, describes this new field simulator system, presents available CPT data obtained using the new simulator and discusses the stress state affected q c values. Key Words: cone penetration test, sand, dilatancy, stress state, calibration chamber test
I. Introduction
Because of the lack of cohesion, it is essentially impossible to obtain undisturbed samples in sand. The Cone Penetration Test (CPT) is an efficient tool used to determine the engineering properties of sand, in situ. Results of CPT in sand typically include the cone tip resistance ( q c ), sleeve friction ( f s) and friction ratio (F R= f s/ q c). The main purpose of CPT in sand usually is to determine the in situ state of stress, relative density ( D r ) or void ratio ( e ), and friction angle ( ). The combination of q c and F R can be used in soil classification (e.g., Robertson et al ., 1986), and available methods generally agree with one another. However, such agreement is less obvious when the interpretation of CPT goes beyond soil classification. Over the past few decades, many theories and experimental procedures have been proposed to interpret q c values. These theories basically treat cone penetration either as a bearing capacity failure (e.g., Janbu and Senneset, 1974; Durgunoglu and Mitchell, 1975) or a cavity expansion (e.g., Vesic, 1972; Baligh, 1976; Yu and Houlsby, 1991; Salgado, 1993; Salgado
et al ., 1997). In the first case, q c is related to the strength and stress parameters through a limiting equilibrium under an assumed bearing capacity failure mechanism. In the second case, q c is related to the cavity expansion limiting pressure (P 1), which in turn is a function of the soil strength and stress parameters. CPT is a large strain problem, and the success of these theoretical analyses has been limited. The laboratory chamber calibration test offers an experimental or empirical way to interpret results of CPT. Uniform sand specimens can be prepared in the chamber with known stress conditions and density. Results of CPT obtained in the chamber can then be compiled to derive empirical interpretation procedures. A typical method that uses q c to infer D r or the shear strength parameter (i.e., ) normally involves the state of stress. The initial (prior to cone penetration) horizontal stress ( ho ) (Houlsby and Hitchman, 1988), vertical stress ( vo ), and mean normal stress ( oo ) (e.g., Schmertmann, 1976; Villet and Mitchell, 1981; Jamiolkowski et al., 1988) have all been proposed for coupling with either Dr or as part of the interpretation for q c. No consensus has been reached, however, as
579
H.H. Hsu and A.B. Huang to which one of the above postulations is more acceptable. The main obstacle in reaching a conclusion is that the conventional calibration chamber imposes significant boundary effects on CPT. Theories and empirical methods (Baldi et al ., 1982; Mayne and Kulhawy, 1991; Salgado, 1993) have been proposed to correct for the boundary effects. These correction factors again have different views on the relationship between q c and state of stress. There is little physical evidence where CPT is performed in a controlled environment with known initial stress conditions and no (or insignificant) boundary effects to validate any of the above statements regarding the relationships among q c , D r (or ) and the state of stress. As part of a research project funded by the National Science Council of the R.O.C., the authors have developed a calibration chamber system in which CPT can be performed under simulated field conditions. A series of CPT calibration tests have been performed in clean, uniformly graded quartz sand using the new simulator system. The q c values obtained in the simulator have been compared with some of the existing postulations or interpretation methods. This paper introduces the basic concept of CPT calibration tests, describes this new field simulator system and presents available CPT data obtained under simulated field conditions.
Table 1. Boundary Conditions in Conventional Calibration Chamber Tests Top & bottom boundary Lateral boundary Stress Strain Stress Strain Constant Constant 0 0 Constant Constant 0 0
Boundary conditions B1 B2 B3 B4
double-wall is used to assure rigidity in the lateral direction when zero lateral strain ( K o) conditions are imposed on the specimen. By maintaining a cavity pressure that is equal to the chamber pressure, full rigidity of the inner-wall is effectively established. A typical cavity-wall calibration chamber is capable of creating four types of boundary conditions as shown in Table 1. Been et al . (1988) indicated that boundary conditions on the top and bottom of the chamber specimen have little effect on CPT test results. Parkin (1988) stated that of the four boundary conditions, the most significant are B1 and B3. Houlsby and Hitchman (1988) stated that q c has no consistent correlation when vo is applied in the chamber under B1 conditions. The calibration chamber has been an important research tool for use in establishing interpretation procedures for CPT in sand. According to statistics obtained by Ghionna and Jamiolkowski (1991), there were 19 calibration chambers in the world in 1991. More calibration chambers have been built (e.g., Peterson and Arulmoli, 1991; Hsu and Huang, 1998) since then. The applications of calibration chambers have also been extended to other types of in situ testing methods. These applications have included the Marchetti dilatometer (Borden, 1991), pressuremeter (Huang et al ., 1991), hydraulic fracture (Been and Kosar, 1991) and calibration of pile foundations (Kulhawy, 1991; ONeill, 1991). Table 2 shows a summary of the currently available calibration chambers in the world. The National Chiao-Tung University (NCTU) operates two calibration chamber systems. One of them is a medium sized conventional calibration chamber originally built at Clarkson University in the U.S. The other is the newly developed field simulator, which will be described in detail later in this paper. Some of the most significant advantages of conducting CPT in a calibration chamber include: (1) repeatability of the test and use of the specimen, (2) uniformity of the specimen, and (3) controlled and known boundary conditions and stress history. These advantages, plus the fact that it is essentially impossible to obtain undisturbed samples in sand, make cali-
580
Calibration chamber (Owner and location) Country Roads Board, Australia University of Florida, U.S.A. Monash University, Australia Norwegian Geotechnical Institute ENEL-CRIS, Milano, Italy ISMES, Bergamo, Italy University of California, Berkeley, U.S.A. University of Texas at Austin, U.S.A. University of Houston, U.S.A. North Carolina State University, U.S.A. Louisiana State University, U.S.A. Golder Associates, Calgary, Canada Virginia Polytechnic Institute and State University, U.S.A. University of Grenoble, France Oxford University, U.K. University of Tokyo, Japan University of Sheffield, U.K. Cornell University, U.S.A. Waterways Experiment Station, U.S.A. National Chiao-Tung University, Taiwan, R.O.C. National Chiao-Tung University, Taiwan, R.O.C. Source : Ghionna and Jamiolkowski (1991)
Boundary conditions Bottom Cushion Cushion Cushion Cushion Cushion Cushion Rigid Flexible Cushion Rigid Flexible Rigid Rigid Cushion Cushion Rigid Rigid Rigid Rigid Rigid Flexible Top Rigid Rigid Rigid Rigid Rigid Rigid Rigid Flexible Cushion Rigid Rigid Cushion Rigid Cushion Rigid Rigid Flexible Rigid Rigid Rigid Flexible
bration chamber testing a rather desirable tool in establishing correlations between CPT and engineering properties for sands and other materials. Among other drawbacks of performing tests in freshly deposited sand is the finite dimension of the chamber specimen. A standard cone penetrometer has a diameter of 35.7 mm. The diameter ratio (R d) of the chamber specimen ( D ) to that of a standard cone is approximately 42 even for a relatively large 1.5mdiameter chamber specimen. Ideally, R d is infinite in the field. Previous studies on the use of chamber calibration tests have indicated that the field conditions for CPT where the soil extends laterally to infinity is expected to be between the B1 and B3 conditions (Veismanis, 1974; Parkin, 1988). The cone tip resistance, q c, under B3 confinement, continues to increase with the depth and does not reach a plateau in dense sand (Parkin and Lunne, 1982; Parkin, 1988). Parkin and Lunne (1982) compiled CPT data under different boundary conditions; D r and Rd are shown in Fig. 2. For loose sand, chamber results are relatively independent of boundary conditions, even when R d is as low as 21. For dense sand, all calibration chamber results are affected by boundary conditions, even for an R d value of 60 or greater. For tests under B1 conditions, q c is mostly a function of ho , at least up to the overconsolidation ratio (OCR) of 8 (Veismanis, 1974; Chapman and Donald, 1981; Parkin, 1988;
Fig. 2. q c under different boundary conditions, Dr and R d. [Adapted from Parkin and Lunne (1982)]
Houlsby and Hitchman, 1988). In order to account for boundary effects, Baldi et al . (1982) proposed an empirical correction factor (referred to as r ) which is a function of Rd, and increases with D r and the overconsolidation ratio. However,
581
H.H. Hsu and A.B. Huang controlled only. The vertical stress is applied through four airstroke actuators attached to the reaction frame. The lateral boundary consists of a stack of rings. This is the main difference between the conventional chamber and the new simulator. The simulator rings are lined with an inflatable silicone rubber membrane on the inside to facilitate boundary displacement measurement and stress control. Four air bellows inflated at constant pressure are placed at the bottom of the ring stack. This system, similar to the concept of a floating ring in an oedometer, reduces frictional forces between sand and rubber membranes. A sand rainer similar to that described by Rad and Tumay (1987) is used to prepare the specimen. The specimen is prepared by pluviation from a hopper through a perforated plate and two diffuser meshes. The uniformity and density can be well controlled by means of this arrangement and by controling the diameter of the holes in the perforated plate. The lateral boundary is set to be rigid, simulating Ko conditions, during sand pluviation. The membrane expansion measuring system consists of a wax lubricated, heavy duty fishing line wrapped around the membrane. The ends of the fishing line are attached to a piece of delrin chain and then to a spring loaded extensometer. The extensometer, instrumented with full bridged strain gauges, tightens the fishing line and senses the circumferential displacement of the rubber membrane. Figure 4 shows a schematic and cross sectional view of the simulator ring and its membrane. A field simulation consists of a physical cylindrical specimen and a numerically simulated soil mass that extends laterally from the physical boundary to infinity. Numerical simulation of the soil mass is
based on analysis of their database, Mayne and Kulhawy (1991) proposed that r is related to D r and R d only. Salgado (1993) stated that the major influential factors are D r, R d, the state of initial stress and some of the intrinsic parameters of sand. The validity of these correction methods has yet to be verified independently by performing CPT in sand with known density, stress conditions and under no boundary effects.
582
Calibration of Cone Penetration Test in Sand obtained from the lateral compression test and the corresponding Pro ro curve. During penetration, Pro for each ring level is adjusted pneumatically and continuously updated with the change of C . A hydraulic piston equipped with a proportional valve, capable of accurate speed control, is used to push the cone penetrometer. The cone penetration rate is set at a constant value of 2.0 mm/second in tests. The slow penetration rate is necessary to allow the reaction of lateral air pressure to reach an equilibrium in all the stress control units. However, the penetration rate is not expected to influence the test results (Dayal and Allen, 1975). A series of cone penetration tests was performed in Da Nang sand, a clean uniformly graded quartz sand, using the simulator. The characteristics of Da Nang sand will be described later. Figure 6 shows the q c profiles under simulated field conditions (referred to as B5), where vo =43.7 kPa and ho =22 kPa, with Rd values of 18 and 22 and D r values of 65% and 84%. The average of the qc values at depths from 600 to 1200 mm was taken as the representative value. Results show that q c of two R d values agree within 6.9% for D r of 65% and within 0.1% for D r of 84%. The similarity of qc under two different Rd values indicates that the boundary effects were substantially reduced.
conducted based on the cylindrical cavity expansion theory. The stress-strain relationship of the sand specimen is directly measured by means of a lateral compression test on the specimen. The relationship between stress (Pro) and radial strain (ro) at the physical-simulated interface is then derived by means of integration from the physical boundary to infinity:
P ro = ho +
where
ro
0
( r) d , 2 r r
(1)
( r)= the stress strain relationship measured by means of a lateral compression test on the sand specimen; r = the strain in the radial direction.
The derived P ro ro relationship is stored in the computer. During cone penetration, the boundary displacements and stresses are measured and individually controlled at each ring level. The circumferential displacement at the boundary of each ring level, C , is converted to ro :
ro = C , D
(2)
where D = the diameter of the physical specimen. Pro in response to ro under simulated field conditions is determined in accordance with the recorded P ro ro relationship. Figure 5 shows the ( r) value
Fig. 6. q c profiles under simulated field conditions with D r values of 65% and 84%
583
H.H. Hsu and A.B. Huang shown in Fig. 7, Da Nang sand (DNS) is uniformly graded, has an average grain diameter, D 50; of 1.1 mm, and a coefficient of uniformity, Cu , of 1.74. The specific gravity G s of DNS is 2.61. The maximum dry unit weight (dmax) was 16.87 kN/m3, and the minimum dry unit weight ( dmin) was 14.13 kN/m3.According to scanning electron microscope (SEM) photographs taken of sand particles, DNS is sub-angular to angular. A series of isotropically consolidated drained triaxial (CID) tests was performed with volume change measurements to determine the strength and dilatancy characteristics of DNS. The triaxial specimens were sheared by means of axial compression. The D r and effective confining stress ( c ) applied in this series of triaxial tests are summarized in Table 3. Bolton (1986) used a saw blade model of dilatancy to describe the shearing behavior of sand. The trace of shearing development is similar to the shape of a saw blade between the contact surfaces of particles. On ) can be the inclined surface, the friction angle ( crit considered as a shearing occurred under in the critical state, where shearing continues without volume change. In order to slide upwards along the saw blade, another dilatancy angle ( ) is necessary. The friction angle ) on the sliding surface and the overriding capa(crit bility comprise the apparent angle of shearing resis is a function of the mineral content of tance ( ). crit the sand and can be obtained under the critical state. The characteristics of the dilatancy of sands are affected by sand density and confining stress. Bolton
Table 3. Variables Applied in the CID Tests D r, % 50 43.7 98.1 65 147.2 245.3 84 392.4
c , kPa
(1986) proposed a relative dilatancy index IR to present the global effect, which can be expressed as
IR =
where
Dr ) 1 , (Q ln pp 100
(3)
Q = an empirical constant that varies with the crushing strength of the grains; for quartz and feldspar, Q =10;
= the mean effective stress at peak deviator pp stress.
(4)
where
= the friction angle at peak deviator stress. peak crit versus p from Figure 8 shows a plot of peak p triaxial tests on DNS, where is obtained by dropping
584
oo a , kPa
a
oo ( + 2 ho =1 ) 3 vo
Table 4 summarizes the variables applied in the calibration tests under B5 conditions. A standard size cone with a cross sectional area A c of 10 cm 2 ( R d=22) was used for all of the tests. The q c profiles of different D r from CPT under the same initial boundary conditions ( oo =43.7 kPa and K =1.0) are shown in Fig. 9. The results show that a plateau of the q c value is generally reached at depths of 250 and 300 mm for Dr of 50% and 65%, respectively. For D r =84%, a stabilized q c develops at depths in excess of 300 mm. In order to analyze the test results, the average q c values were taken at depths from 400 to 1000 mm for D r of 50%, and from 600 to 1200 mm for D r of 65% and 84%. Figure 10 shows the C measurements and P ro applied to ring No. 10 (midheight of the specimen) during the tests depicted in Fig. 9. The depth in Fig. 10 is in reference to the cone tip level and is normalized with respect to the cone diameter D cone . Figure 10 clearly demonstrates that under simulated field conditions, the lateral boundary is neither constant stress ( P ro is a constant) nor rigid ( C =0). It appears that C of medium dense sand ( D r =50%) reaches a maximum near the cone tip and then maintains a constant value. The C measurements of dense to very dense sand (Dr= 65% and 84%) reach maximum values ahead of the cone tip at about 3 to 5 D cone and then decrease slightly but consistently as the cone tip passes. It should be noted that at D r of 84%, C is smaller, but P ro is much higher than that at D r of 50% and 65%. For very dense sand, the rigidity causes the applied pressure Pro to grow rapidly and to limit the dilatancy
a tangent from the origin onto a single Mohr circle of effective stress. A Q value of 10 would give a reasonable fit to the data points in Fig. 8 using Eq. (3). This would indicate that DNS is comparable to the quartz sand (e.g., Ottawa sand) reported by Bolton is 33.4 from triaxial tests on loose DNS (1986). crit specimens. Dilatancy is believed to have a marked influence on q c values, as indicated by the results of previous studies (Yu and Houlsby, 1991; Huang and Ma, 1994; Salgado et al., 1997). The boundary effects under B1 or B3 in a conventional chamber a direct reflections of the sand dilatancy, as will be discussed later.
H.H. Hsu and A.B. Huang of the sand. Figure 11 shows q c versus vo . For a given vo , the corresponding qc can vary by as much as 50% from the mean value. The test results indicate there is no unique relationship between q c and vo . Similar findings have also been reported by others (Houlsby and
Table 5. Comparison of Tests under B1 and B5 Conditions Boundary condition B5 B5 B5 B1 B1 B1 Dr % 50 65 84 50 65 84 43.7 43.7 43.7 53.5 73.1 98.5
ho
vo
kPa 43.7 43.7 43.7 43.7 43.7 43.7
hc
53.5 73.1 98.5
Hitchman, 1988). As shown in Fig. 12, the correlation between q c and ho appears to be scattered. This is different from previous studies that related q c to ho (Baldi et al ., 1986; Houlsby and Hitchman, 1988; Jamiolkowski et al ., 1988). This scattering, however, is not random. When the scale is large, there is a positive relationship between qc and ho . When Dr is larger than 50%, there exists a localized (i.e., under the same D r and oo ) negative relationship between qc and ho . Apparently, the higher value of ho prohibits dilatancy; hence, there is less lateral expansion on the physical boundary during cone penetration. In most cases, the horizontal stress at the physical boundary ( P ro in Fig. 10) remains more or less a constant after reaching a peak value. This peak P ro is chosen to represent a stabilized horizontal stress measured at the physical boundary after cone tip passage and will be referred to as hc . For all the available tests under B5 conditions, a clear and positive relationship between q c and hc was obtained as shown in Fig. 13. This result parallels the findings reported by Houlsby and Hitchman (1988), where comparisons were made between q c and ho under Bl conditions. For comparison purposes, a set of calibration tests under B5 conditions (i.e., tests B5-1, B5-2, and B53 in Table 5) were duplicated under Bl conditions (i.e., tests Bl-1, B1-2 and B1-3 in Table 5). The respective hc values recorded in tests B5-1, B5-2, and B5-3 were applied as the corresponding horizontal stress values in tests Bl-1, B1-2 and B1-3. The results show that if hc or the terminal horizontal stress expected at the physical boundary is applied, even B1 can properly simulate the field conditions. The qc values obtained under B1 and B5 conditions agreed within 7% in these tests. Two empirical equations, which relate q c to the initial stress state (i.e., oo and ho ), proposed respectively by Jamiolkowski et al . (1988) and Baldi et al . (1986), are chosen for comparison with the CPT results under B5 conditions (referred to as q c,B5). Both equa 586
q c = 492pa (
(7)
Equation (7), depicted in Fig. 16 as solid curves, has a coefficient of correlation of 0.97 with the available test data. For comparison purposes, Eq. (6) is also
tions were developed based on tests in Ticino sand, and the boundary effects were corrected. According to Baldi et al. (1986) the cone tip resistance, qc,B is related to the initial state of stress and D r as follows:
q c,B = 220pa (
where
(5)
Fig. 14. Comparison of q c,B 5 and q c,J.
P a = the reference pressure (1 kPa). Jamiolkowski et al . (1988) related the cone tip resistance, q c,J , to oo as follows:
q c,J = 205pa (
(6)
Comparisons of q c,J with q c,B 5 and of q c,B with q c,B5 are, respectively, shown in Fig. 14 and Fig. 15. The q c,J / q c,B 5 and q c,B / q c,B5 values range from 70 to 100%. q c,J/ q c,B5 is not a constant under the same oo , even though q c,J depends upon oo according to Eq. (6). When oo =43.7 or 98.1 kPa, q c,J/ q c,B5 increases with K (or ho ) for D r of 65 and 84%. Similar trends are also found in the comparisons between q c,B5 and q c,B (Fig. 15). According to the available CPT data under B5 conditions, a relatively consistent relationship is possible only between q c and oo as shown in Fig. 16. Based on the statistical optimization method, the empirical equation is similar to the pattern proposed by Jamiolkowski et al . (1988): 587
H.H. Hsu and A.B. Huang to be used in a conventional chamber for CPT calibration tests, it is apparent that hc , not ho , should be imposed on the lateral boundary of the chamber specimen.
Acknowledgment
This research was funded by the Sinotech Foundation for Research and Development of Engineering Sciences and Technologies, and the National Science Council of the R.O.C. under contract NSC 87-2211-E-009-034. Their support was greatly appreciated.
Nomenclature
Ac Cu Dr D 50 D D cone e FR fs Gs IR K KO NC OC OCR pa P1 P ro pp Q qc q c,B q c,B5 q c,J Rd r dmax dmin C cross sectional area of cone coefficient of uniformity relative density average grain diameter diameter of the sand specimen cone diameter void ratio sleeve friction ratio sleeve friction resistance specific gravity relative dilatancy index ratio of horizontal stress over vertical stress at rest lateral earth pressure coefficient normally consolidated over consolidated over consolidation ratio reference pressure (1kPa) limit pressure lateral stress at the physical-simulated soil mass interface mean effective stress under peak deviator stress conditions empirical constant that varies with the crushing strength of sand grains cone tip resistance qc obtained from the empirical equation proposed by Baldi et al. (1986) q c obtained under simulated field conditions (B5) q c obtained from the empirical equation proposed by Jamiolkowski et al . (1988) diameter ratio of the physical sand specimen over that of the cone chamber size correction factor maximum dry density minimum dry density circumferential displacement at physical boundary of the ring chamber radial normal strain radial normal strain at the physical-simulated soil mass interface initial or field horizontal stress effective confining stress applied in the triaxial tests peak p ro under simulated field conditions initial (prior to cone penetration) horizontal effective stress mean effective normal stress = 1 ( vo + 2 ho ) 3 vertical effective stress drained friction angle drained friction angle under critical state, where shearing continues without volume change
r ro ho c hc ho oo vo crit
588
References
Baldi, G., R. Bellotti, V. Ghionna, M. Jamiokowski, and E. Pasqualini (1982) Design parameters for sands from CPT. Proceedings of the Second European Symposium on Penetration Testing , Amsterdam, Holland. Baldi, G., R. Bellotti, V. Ghionna, M. Jamiokowski, and E. Pasqualini, (1986) Interpretation of CPTs and CPTUspart II: drained penetration in sands. Proceedings of the Fourth International Geotechnical Seminar on Field Instrumentation and in Situ Measurements, Nanyang Technological Institute, Singapore. Baligh, M. M. (1976) Cavity expansion in sands with curved envelopes. Journal of Geotechnical Engineering Division, ASCE, 102 (11),1131-1146. Been, K. and K. M. Kosar (1991) Hydraulic fracture simulations in a calibration chamber. Proceedings of the First International Symposium on Calibration Chamber Testing, Potsdam, NY, U.S.A. Been, K., J. H. A. Crooks, and L. Rothenburg, (1988) A critical appraisal of CPT calibration chamber tests. Proceedings of the First International Symposium on Penetration Testing (ISOPT1) , Orlando, FL, U.S.A. Bolton, M. D. (1986) The strength and dilatancy of sands. Geotechnique , 36 (1), 65-78. Borden, R. H. (1991) Boundary displacement induced by DMT penetration. Proceedings of the First International Symposium on Calibration Chamber Tests , Potsdam, NY, U.S.A. Chapman, G. A. and I. B. Donald (1981) Interpretation of static penetration tests in sand. Proceedings of the 10th International Conference on Soil Mechanics and Foundation Engineering , Stockholm, Sweden. Dayal, U. and J. H. Allen (1975) The effect of penetration rate on the strength of remolded clay and sand samples. Canadian Geotechnical Journal , 3 , 336-348. Durgunoglu, H. T. and J. K. Mitchell (1975) Static penetration resistance of soils. I. Analysis. II. Evaluation of theory and implications for practice. Proceedings of Specialty Conference on In-Situ Measurement of Soil Properties, ASCE, Raleigh, NC, U.S.A. Ghionna, V. N. and M. Jamiolkowski (1991) A critical appraisal of calibration chamber testing of sands. Proceedings of the First International Symposium on Calibration Chamber Testing , Potsdam, NY, U.S.A. Holden, J. C. (1991) History of the first six CRB calibration chambers. Proceedings of the First International Symposium on Calibration Chamber Testing , Potsdam, NY, U.S.A. Houlsby, G. T. and R. Hitchman (1988) Calibration chamber tests of a cone penetrometer in sand. Geotechnique , 38 (1), 39-44. Hsu, H. H. and A. B. Huang (1998) Development of an axisymmetric field simulator for cone penetration tests in sand. ASTM Geotechnical Testing Journal , 21 (4), 348-355. Huang, A. B., R. D. Holtz, and J. L. Chameau (1991) A laboratory study of pressuremeter tests in clays. Journal of Geotechnical Engineering Division , ASCE , 117 , 1549-1567.
589
!"#$%&'()*
!"#$
!"#$%&'#(
!"#$%&'()*+,-./012"#3456%789:;<=>?0@>ABC* !" !"#$%&'()*+,-./012 34567"89:;<=>?@"'ABC="DEF GH: !"#$%&'()*+(,-./012#3456789:;<=>?'>:;<=78?@+A;BC !"# !" $%&'() *+,-#./0123456 789:;*+,-<=(345678 !"#$%&'()*+,-./0123456789:;<=>?@#$ABCD0EFGHIJ KL !"#$%&'()*+,-./0123456789/:; <=>? @/A%BC:;DE!D !"#$%&'()*+,-./0123-456789()*:;/01<=>?@A4B#CDE+, !"#$%&'()* +,-$./012(345673489:;< !=>(?1@)AB+CD$ !"#$%&'()*+,-./01234)*5-6789:;<=>?@A&BCDEFGHIJK
590