By D. P. StewarP and M. F. Randolph z

Transcription

1 T-BAR PENETRATION TESTING IN SOFT CLAY By D. P. StewarP and M. F. Randolph z ABSTRACT: A t-bar penetration test for soft clay that can be performed with existing cone penetration test (CPT) equipment is described. The results of the test can be directly related to the undrained shear strength of the clay via a plasticity solution, thus eliminating some of the uncertainty involved in the interpretation of conventional penetration tests in soft clay. Results from a field trial of the device are shown to compare favorably with the results of other in-situ and laboratory strength tests. INTRODUCTION Investigations at soft clay sites will often involve the use of conventional cone penetration tests (CPTs) for stratigraphic determination, followed by in-situ vane shear testing or undisturbed sampling, to determine shear strength properties. The CPT is normally restricted to the role of profiling, since the measured cone resistance is very small in soft deposits and, therefore, the accuracy of the reading is poor. The in-situ shear strength must be determined via empirical relationships, such as s, = qc/nc, or s, = (qc - o-v)/ Nk, where s, is the undrained shear strength of the soil, qc is the measured cone resistance, cry is the ambient total vertical stress, and Arc and N~ are empirical factors. Often there is uncertainty regarding the appropriate values of the empirical factors, and it is generally accepted (Robertson and Campanella 1983) that the factors are dependent on the stiffness of the soil, the stress level, and stress history. Therefore, without prior knowledge of the site conditions, selection of an appropriate empirical factor is difficult. In recent years, cone penetration tests with pore pressure measurement (CPTU) have found more widespread use, and the higher resolution of pore pressure measurements (compared to the cone resistance) are more suited to correlation with the low undrained shear strength of soft clays. However, the correlations are dependent on the soil properties, and on the position of the pore pressure transducer (Campanella and Robertson 1988). In response to a desire for an improved site investigation tool for soft clay deposits in laboratory centrifuge models, Stewart and Randolph (1991) developed a small-scale t-bar penetrometer that combined the advantages of the CPT and in-situ vane shear tests. The new device enables: (1) A continuous profile of undrained shear strength to be determined (like the cone); and (2) the penetration force to be related directly to the shear strength without an empirical correlation (similar to the vane). A major advantage of the device is that the in-situ vertical stress is equilibrated across the t-bar, and thus there is no correction for the ambient stress level to be included, unlike the CPT. The small-scale device has yielded very good results in a relatively large number of laboratory tests. This technical note 1Res. Assoc., Dept. of Civ. Engrg., Univ. of Western Australia, Nedlands, WA 69, Australia. 2Prof., Dept. of Civ. Engrg., Univ. of Western Australia, Nedlands, WA 69, Australia. Note. Discussion open until May 1, To extend the closing date one month, a written request must be filed with the ASCE Manager of Journals. The manuscript for this technical note was submitted for review and possible publication on February 9, This technical note is part of the Journal of Geoteehnical Engineering, Vol. 12, No. 12, December, ISSN /94/12-223/$2. + $.25 per page. Technical Note No

2 describes the development of a larger-scale device that can be deployed very simply and cheaply with conventional CPT equipment in the field. The use of this device is described, and the results are compared with other strength determinations at a soft clay site. T-BAR PENETROMETER The t-bar penetrometer developed for use in the field is illustrated in Fig. 1. It comprises a 5-mm-diameter and 2-mm-long aluminium bar attached at right angles to the end of a conventional cone penetrometer. The cylindrical surface of the t-bar was sandblasted to create a relatively rough surface, and the ends of the bar were machined smooth. The 6 ~ cone tip was unscrewed and replaced by a newly manufactured spade-shaped tip containing identical o-ring seals. The spade shape was used to provide some stability to the bar as it was pushed through the soil, and a shear pin was included to prevent bending of the penetrometer shaft if a hard object was struck by one end of the bar. The shear pin also breaks when the penetrometer is withdrawn, leaving the bar in the ground. Alternatively, some form of hinge can be incorporated so that the bar is retrievable. A cheaper alternative design, which has not been field-tested yet, comprises a disposable hardwood t-bar with a conical recess to accept the standard 6 ~ cone. This does not require a new tip to be manufactured. The t-bar is pushed into the soil at the same rate as a conventional CPT, and the penetration resistance is measured using the cone-tip load cell. The projected area of the t-bar is 1, mm 2, 1 times the area of a standard cone. Thus, the measured penetration force is roughly 1 times that of a cone test in the same soil, dramatically improving the accuracy of force measurements. Note that errors due to area correction associated with excess pore pressures (Campanella and Robertson 1988) are relatively insignificant, since the penetration force is much higher than that of a cone test in the same soil. The results of the test are interpreted by making use of the plasticity solution for the limiting pressure acting on a cylinder (or pile) moving laterally through a purely cohesive soil (Randolph and Houlsby 1984). The analysis assumes full closure of the soil behind the cylinder, such that a gap does not occur. The solution results in a simple expression for the limiting force acting on an infinitely long cylinder EXISTING CPT FRICTION SLEEVE ~_~ CYLINDRICAL T-BAR i 1 / i f i =, I L.-I NEW SPADE*SHAPE ~ - ~ ~ l j,ornm 2 mm ELEVATION FIG. 1. Field T-Bar Penetrometer 2231 SIDE VIEW

3 P -- = s,d Nb (1) where P = force per unit length acting on the cylinder; d = diameter of the cylinder; and Nb = bar factor. The analytical value of Nb (Randolph and Houlsby 1984; Stewart and Randolph 1991) is dependent on the surface roughness of the cylinder, described by its adhesion factor, ~. The upper and lower bounds of the plasticity solution coincide at approximately 12 for a fully rough bar and diverge slightly at lower values of a, with a minimum Nb of about 9. It is unlikely that adhesion factors approaching either or 1 are achievable, despite the fact that the crossbar was sandblasted; therefore, some intermediate value of Nb must be chosen. Randolph and Houlsby (1984) recommend an intermediate value of 1.5 for general use, and this has been used to interpret the results. The possible range of Nb is relatively small, however, and the upper and lower limits correspond to errors of less than -+ 13% of the adopted value. Laboratory testing of the small-scale penetrometer (Stewart 1992) has indicated that Nb is insensitive to the stress level and stress history. In using the solution, the effect of the smooth ends of the t-bar are ignored; as is the influence of the penetrometer shaft, which occupies only 1% of the projected area of the bar. FIELD-TESTING A range of in-situ tests was performed adjacent to the Swan River on the Burswood Peninsular in Perth, Western Australia. At the site, a relatively homogeneous soft clay deposit, about 18 m thick, is overlain by several meters of partly cemented fill, mainly comprising fly ash. Two t-bar penetrometer tests were performed about 1.5 m apart, after digging a shallow pit partially through the surface crust. The results of these two tests are shown in Fig. 2 as the bearing resistance (penetration force/ projected area of bar) versus depth. The strong surface crust is evident above the soft clay, which then gives a gentle increase in bearing resistance with depth. As expected, the two tests were very consistent due to their proximity. Undrained shear strength was estimated from the t-bar bearing resistance data using Nb = 1.5 in (1). For clarity, the estimated shear strength profile from only one t-bar test is plotted in Fig. 3, along with shear-strength values determined from in-situ vane shear, self-boring pressuremeter, and isotropically consolidated undrained (CIU) triaxial tests. The pressuremeter data were corrected for membrane expansion (Lee Goh 1994). The pressuremeter data are somewhat higher near the base of the soft deposit, which may indicate a greater number of thin sand layers in this area. Cone resistances of about.2-.6 MPa were recorded in the soft stratum with a conventional CPT. Using the empirical relationship s, -- (qc - ~)/12, undrained shear strengths of about kpa were estimated. A profile of undrained shear strength derived from the CPT results is not presented here since the test results were only available in the form of chart recorder output, and the accuracy of data scaled from these plots was relatively poor for the low cone resistances encountered. The logging system was modified to digitally record the test data before t-bar testing was conducted. The t-bar has provided a very good estimate of undrained shear strength, which is consistent with a number of other methods of strength measurement. This is particularly pleasing when it is recognized that s, is not a 2232

4 2 T-bar bearing resistance (MPa) I I I I ~ : : E v._1 n' FIG. 2. Results of Two T-Bar Penetration Tests unique parameter and depends on the type of test, rate of strain, and orientation of the failure planes (Wroth 1984). COMMENT A field t-bar test can be completed in a considerably shorter period of time (same rate as a CPT) than an in-situ vane shear test hole to the same depth. The test described does not measure the residual strength of the clay; although if the t-bar remained attached during removal from the ground and the extraction force was monitored, an indication of sensitivity might be obtained. This was done in the laboratory-scale equipment, and it was found that the t-bar yielded a sensitivity of about 2 in overconsolidated beds of kaolin, compared to values of 4 to 5 from a small vane apparatus. This difference was expected due to the extreme disturbance generated by remolding from the vane. Two particular applications of the t-bar penetrometer are for the design of laterally loaded piles and subsea pipelines. The test measures the ultimate soil pressure acting on the bar; thus, a direct comparison to ultimate lateral pile capacity or ultimate soil resistance to pipeline movement may be made. A device similar to the t-bar penetrometer, the Iskymeter, was developed 2233

5 Undrained shear strength (kpa) O 6 2 i i i i I o -2-4 E.--I n O I9 C' OCt. J, O <3-18 C---- LEGEND t-bar o vane 9 triaxial. SBPM FIG. 3. Undrained Shear Strengths Derived from Field Testing by the Swedish Geotechnical Institute circa 194 (Kallstenius 1961). The Iskymeter has two folding arms that are pulled out horizontally as the instrument is withdrawn from the ground, thus force measurement is conducted during extraction. The arms are approximately rectangular in cross section. The extraction force is related to the undrained shear strength of the soil, with an empirical factor of about 1, very close to the value of 1.5 adopted for the t-bar. Kallstenius (1961) also describes an empirical correction for sensitivity and depth. Although the Iskymeter does not appear to have been used extensively, the t-bar penetrometer holds a major advantage in that it can be deployed simply with existing CPT equipment. CONCLUSIONS The t-bar penetrometer provides a relatively cheap and simple alternative to other more conventional methods of in-situ shear strength determination, 2234

6 and can be deployed with the existing CPT equipment. On the basis of the single field trial, and a number of laboratory tests, it was shown that the technique yields strength estimates that are consistent with other methods and not dependent on stress level or stress history. The test is not proposed as a replacement of the CPT or vane test, but merely as another tool that may be utilized in soft clay site investigations. ACKNOWLEDGMENTS Field-testing of the t-bar penetrometer was performed using the CPT equipment owned and operated by the Water Authority of Western Australia, who provided the equipment and personnel at no charge. The assistance of Chris Potulski of the Water Authority is gratefully acknowledged. APPENDIX. REFERENCES Campanella, R. G., and Robertson, P. K. (1988). "Current status of the piezocone test." Proc., 1st Int. Syrup.: Penetration Testing, 1, J. De Ruiter, ed., A. A. Balkema, Rotterdam, The Netherlands, Kallstenius, T. (1961). "Development of two modern continuous sounding methods." Proc., 5th International Conf. on Soil Mech. and Found. Engrg., (ICSMFE), Paris, France, Vol. 1, Lee Goh, W. O. (1994). "A study of measuring the coefficient of consolidation of soft clay using cavity expansion methods," PhD thesis, Univ. of Western Australia, Nedlands, Australia. Randolph, M. G., and Houlsby, G. T. (1984). "The limiting pressure on a circular pile loaded laterally in cohesive soil." Geotechnique, London, England, 34(4), Robertson, P. K., and Campanella, R. G. (1983). "Interpretation of cone penetration tests. Part II: Clay." Can. Geotech. J., Vol. 2, Stewart, D. P. (1992). "Lateral loading of piled bridge abutments due to embankment construction," PhD thesis, University of Western Australia, Nedlands, Australia. Stewart, D. P., and Randolph, M. F. (1991). "A new site investigation tool for the centrifuge." Proc., Int. Conf. Centrifuge 1991, H. Y. Ko, ed., A. A. Balkema, Rotterdam, The Netherlands, Wroth, C. P. (1984). "The interpretation of insitu soil tests." Geotechnique, London, England, 34(4),