Comparison of DMT, CPT, SPT, V S based Assessment on Treasure Isl during the Loma Prieta Earthquake Kyle Rollins Brigham Young University, Provo, Utah, USA. E-mail: rollinsk@byu.edu Sara Amoroso Istituto Nazionale di Geofisica e Vulcanologia, L'Aquila, Italy. E-mail: sara.amoroso@ingv.it Roman Hryciw University of Michigan, Ann Arbor, Michigan, USA, E-mail. romanh@umich.edu Keywords: liquefaction, dilatometer test, cone penetration test, shear wave velocity ABSTRACT: Investigations at Treasure Isl involving CPT, DMT, SPT V s measurements provide a relatively rare opportunity to compare these methods in predicting liquefaction. All the in-situ techniques were relatively consistent in predicting liquefaction within a depth range of to. m below the ground surface. The factors of safety predicted by the CPT DMT were quite consistent except at shallow depths where the DMT gave higher values owing to the higher K o in this zone. Analysis of the CSR K D data points within the depth range from to. m suggests that the liquefaction boundary curve is reasonable for s but may be somewhat unconservative for silty ss. The factor of safety predicted by the V s correlation was consistently lower than other methods owing to insensitivity to higher K o values higher silt content. INTRODUCTION Procedures for assessing the liquefaction potential of ss silty ss have been developed for a number of in-situ tests including: the Stard Penetration (SPT) test (Youd et al. ), the Cone Penetration (CPT) test (Robertson & Wride ), the shear wave velocity (V S ) test (Andrus & Stokoe ). More recently, Monaco et al. (), proposed a method for predicting liquefaction using the K D value obtained from Flat Dilatometer (DMT) testing. Research has shown that K D is more sensitive than V S to factors such as stress history, aging, cementation, structure, which greatly increase the liquefaction resistance for a given relative density (Maugeri & Monaco ). Unfortunately, the database for assessing liquefaction from DMT testing is relatively small additional results from field test sites are necessary to improve the reliability of the procedure. This is particularly important for profiles containing silty s in addition to clean s. As noted by several researchers, it is often useful to evaluate liquefaction using more than one in-situ test to confirm the potential for liquefaction (Rollins et al., Robertson & Wride, Idriss & Boulanger ). However, it is sometimes difficult to determine which method to rely upon when there are conflicts between competing methods as has been noted by Rollins et al. () as well as Liu & Mitchell (). This paper provides additional field performance data than can be added to the database for liquefaction assessment with the DMT for both clean silty ss. In addition, the results from the DMT based approach are compared with results from SPT, CPT V S methods at the same site. The test results were obtained from a site on Treasure Isl, a man-made isl in San Francisco Bay where liquefaction was pervasive during the M. Loma Prieta Earthquake in. SITE CHARACTERIZATION AT TREASURE ISLAND LIQUEFACTION TEST (TILT) Treasure Isl was constructed by building a rockfill berm on native shoal ss around the perimeter of the isl placing dredged s within the berm using hydraulic filling techniques. The test site
......... described in this investigation was located near the interior of the isl. This site was thoroughly investigated in connection with a series of lateral pile load tests conducted after inducing liquefaction using small explosive charges. This test program, known as the Treasure Isl Test (TILT), has been described in a number of papers (Ashford & Rollins, Ashford et al., Rollins et al., Weaver et al. ). Prior to lateral pile load testing, about. m of the hydraulic fill was excavated to bring the ground surface closer to the water table. Test results are referenced relative to this excavated ground surface although corrections required by the various methods are based on the ground surface elevation at the time of testing.. Cone Penetrometer (CPT) Testing A cone penetration sounding was performed using the Univ. of Michigan truck mounted CPT rig which could be anchored to the ground for increased reactive force. Readings were taken at. m intervals. The soil profile interpreted from the cone penetration testing along with profiles of cone tip resistance (q c ), sleeve friction (F s ), friction ratio (R f ), soil behavior type index (I c ) are provided in Fig.. The I c value provides a detailed indication of the variation of soil types with depth. There is a decrease in cone tip resistance from the ground surface consistent with overconsolidation, then the value oscillates between MPa within the interbedded s silty s layers typical of hydraulic fill. The friction ratio from to m is generally about.%. The cone tip resistance decreases to around MPa in the clayey silt/silty clay layer while the friction ratio ranges from to %. The I c in the clayey silt layer is typically. or greater indicating that this layer is not liquefiable; however there are some thin interbedded layers with I c below.. Cone Tip Resistance, q c (MPa) Sleeve Friction, F s kpa) Friction Ratio, R f (%) Soil Behavior Type Index, I c.......... Clayey Silt Depth (m) Depth (m) Depth (m) Depth (m) Gravelly Clean to to y Silt Clayey Silt to Silty Clay Clay Fig.. Soil profile interpreted from Cone Penetration (CPT) testing along with profiles of cone tip resistance (q c ), sleeve friction (F s ), friction ratio (R f ) soil behavior type index (I c ).. Flat Dilatometer (DMT) Testing The flat dilatometer testing was performed with the Univ. of Michigan rig immediately adjacent to the CPT test sounding described previously. Readings were taken at. m intervals. The soil profile interpreted from the DMT test results is shown in Fig. along with profiles of the dilatometer modulus (E D ), the material index (I D ), the horizontal stress index (K D ), according to the common DMT interpretation formulae (Marchetti, Marchetti et al. ). The interpreted soil profile is generally consistent with that obtained from the CPT. The soil profile from to. m generally consists of thinly interbedded layers of silty s s, typical of hydraulic fill, which are well resolved relative to other in-situ tests. Below this layer the profile becomes somewhat more uniform with thicker
......... layers of silty s (. to. m), silty clay (. to. m), silty s (. to m). The high horizontal stress index above a depth of m suggests overconsolidation near the surface. The coefficient of earth pressure at rest (K o ) has been interpreted using the K D value from the DMT the q c value from the CPT according to the following equations proposed by Baldi et al. (): K =. +. K D -. q c / ' v () valid for "seasoned" s K =. +. K D -. q c / ' v () valid for "freshly deposited" s Eq. was used for the hydraulic fill to a depth of about. m Eq. was used for the deeper s deposits the results are plotted in Fig.. The K o value from to. m is typically between.. indicating a normally consolidated s; however, above m, higher K o values are indicating overconsolidated ss, presumably owing to desiccation. Dilatometer Modulus, E D (MPa) Material Index, I D Horiz. Stress Index, K D Horiz. Stress Ratio, K o.. Clay Silt Clayey Silt Fig.. Soil profile interpreted from flat dilatometer (DMT) testing along with profiles of dilatometer modulus, material index, horizontal stress index, earth pressure ratio.. Shear Wave Velocity (V S ) Testing Shear wave velocity measurements were made at the single pile test site located approximately m north of the CPT/DMT soundings. In two cases the wave velocity was measured at m intervals using a downhole seismic cone penetrometer approach. In the third case, the velocity was measured using downhole techniques with the receiver inside a steel test pile. The overburden corrected shear wave velocity (V S ) profiles obtained from the three tests are plotted in Fig. along with the soil profile at the test location. V s was obtained using the equation: V S = V S ( vo /P a ). () where P a = atmospheric pressure approximated by kpa vo = initial effective vertical stress in the same units as P a (Youd et al. ).
.................. Clayey Silt Shear Wave Velocity, V s (m/sec) CPT- CPT- Downhole Fig.. Overburden corrected shear wave velocity profiles at a site about m from the CPT/DMT soundings. In general, the agreement between the three tests is very good. On average the velocity within the interbedded s silty s layer is about m/s. Based on the correlation between liquefaction resistance V s developed by Andrus & Stokoe (), all of the s layers would be considered susceptible to liquefaction as the measured V s is less than m/s.. Stard Penetration (SPT) Testing Stard penetration (SPT) tests were performed in bore holes at the single pile four pile test sites located about m north south, respectively of the DMT/CPT soundings. Test borings were performed with drilling mud to stabilize the hole. A safety hammer, lifted with a rope cathead system, was used to perform the tests. Energy measurements indicated that the hammer was providing % of the theoretical free-fall energy on average. The normalized penetration resistance (N ) was computed using the equation: (N ) = C N C E N () where C N = ( vo /P a )., C E = ER/, ER is the ratio of free-fall energy (%) supplied by the hammer. Plots of the (N ) values fines contents from the two holes are provided in Fig.. The penetration resistance decreases from about at m to about at a depth of. m. In contrast, the fines content increases with depth may explain part of the decrease in (N ) with depth. Fine w/ Shells (SP) SPT Blow Count, (N ) (Blows/ mm) Fines Content (%) Fine (SP-SM) Fine Silty (SM) SPT Gray Silty Clay (CL) SPT (SM) Fig.. Soil profile from SPT borings along with profiles of (N ) fines content.
LIQUEFACTION ASSESSMENT According to the simplified procedure developed by Seed & Idriss (), the factor of safety (FS) against liquefaction is given by the equation FS = (CRR. /CSR) MSF () where CSR = calculated cyclic stress ratio generated by the earthquake shaking; CRR. = cyclic resistance ratio for magnitude. earthquakes, MSF = magnitude scale factor =. /M w., M w is the magnitude of the earthquake under consideration (Youd et al. ). For this study, which considers performance during the M. Loma Prieta earthquake, the magnitude scaling factor is.. Seed & Idriss () also developed the equation for the cyclic stress ratio (CSR) as follows CSR = ( av / vo ) =.(a max /g)( vo / vo )r d () where a max = peak horizontal acceleration at the ground surface generated by the earthquake; g = acceleration of gravity; vo vo are total effective vertical overburden stresses, respectively; r d = stress reduction coefficient =. -.z for z <. m. Ground motion recordings on Treasure Isl ground response analyses conducted at a number of sites around Treasure Isl indicate that a max in the vicinity of the test site was approximately.g (Rollins et al. ) For simplicity, the cyclic resistance ratio (CRR) was generally computed based on recommendations by Youd et al. () where the resistance ratio from CPT results is determined using techniques formulated by Robertson & Wride () the resistance ratio based on V S is based on charts developed by Andrus & Stokoe (). With respect to DMT results, the CRR was determined using the equation: CRR=.K D.K D +.K D. () proposed by Monaco et al. () where K D is the horizontal stress index obtained from the DMT. A comparison of the factor of safety against liquefaction computed using the CPT, DMT, SPT V S approaches for determining CRR is provided in Fig.. Although there are some notable differences, the agreement between the various methods is generally quite good. For example all methods predict liquefaction (FS <.) from a depth of about m to. m below the ground surface. This result provides a high degree of confidence that this layer actually liquefied during the Loma Prieta earthquake. Despite the fact that the DMT based approach was primarily developed for clean s, the computed factor of safety tracks the factor of safety obtained with the CPT quite well in the depth range from m to. m where a number of silty s layers are encountered. In contrast, the factor of safety from the DMT is considerably higher near the ground surface than predicted by the CPT or SPT methods. The high DMT-based factors of safety appear to be associated with higher K o values within this depth range (see Fig. ). Although the CPT based factor of safety also increases in this range, presumably owing to the higher K o values, they are still lower than predicted by the DMT approach. The increased factor of safety predicted by the DMT is likely a result of the increased sensitivity of the DMT to K o effects relative to the CPT (Maugeri Monaco ). Because the database of field performance data relative to the DMT approach is so limited, it is not possible to determine if the increased factor of safety is justified at the present time. As noted by a number of researchers (e.g. Ishihara et al., Seed ), the liquefaction resistance of s clearly increases as the K o value increases. However, the correlations between CRR in-situ test parameters [i.e. q c, (N ), V s, K D ] have generally been assumed to be independent of K o because an increase in K o is expected to produce a comparable increase in the in-situ test parameter as suggested by Seed (). Salgado et al. () examined the effect of K o on both liquefaction resistance cone penetration resistance separately concluded that the CRR vs q c relationship was relatively unaffected by K o for q c values less than MPa slightly unconservative for higher q c values. In contrast, Harada et al. () also investigated the influence of K o on both liquefaction resistance on SPT CPT resistance found that both the CRR vs q c CRR vs (N ) curves were conservative without consideration of K o effects. Harada et al. () recommend a suite of CRR vs q c curves to properly account for K o effects. Therefore, there is some controversy in the literature regarding the effect of K o on liquefaction correlations based on in-situ tests which is deserving of additional research. This issue is particularly important in evaluating liquefaction
......... resistance after ground improvement because ground improvement typically increases both the soil density K o. The liquefaction factor of safety predicted by the V s correlation is significantly less than that predicted by the other methods throughout the depth investigated. At shallow depths where K o is high, the lower factor of safety may be attributed to V s being less sensitive to K o than the CPT q c or DMT K D as noted by Maugeri & Monaco (). At greater depths, the discrepancy is likely associated with the presence of non-plastic silt. Liu & Mitchell () noted that CRR vs V s correlations for silty ss are often overly conservative for these materials predict liquefaction when it does not occur in the field. Liu & Mitchell () recommend curves which plot to the left of the Andrus & Stokoe () curve for silty ss. These curves would lead to higher factors of safety for the silty s layers better agreement with the other methods. Typically, the boundary between liquefaction no liquefaction for a given in-situ test parameter [q c, (N ), V s, K D ] is evaluated by plotting the CSR for a given event versus the in-situ parameter for the loosest layer in the profile which is assumed to be the liquefiable layer. When this procedure is employed for the Treasure Isl sites using DMT results both the loosest silty s s layers plot considerably above the CRR-K D curve. Although this result is consistent with field performance, it is not particularly helpful in defining the location of the boundary curve. In lieu of waiting for additional field performance data, the CSR K D data pairs within the depth range from. m to. m have been plotted against the CRR vs K D curve in Figs. for clean s silty ss, respectively. As noted previously, all of the in-situ tests generally predicted liquefaction within this zone. For the points identified as s based on the DMT I D, the data points in Fig. plot right up to the boundary curve but do not cross over, suggesting that the curve is appropriately positioned for this CSR value. For the data points identified as silty s based on the DMT I D, the data points in Fig. typically plot to the left of the curve but some points cross over the curve indicating that the curve may be somewhat unconservative for silty ss at this CSR value. Factor of Safety Against...... DMT CPT Fig.. Comparison of factor of safety against liquefaction computed using CRR obtained from CPT, DMT, SPT V S test results. CRR, CSR....... Clayey Silt No Fig.. Comparison of CSR vs K D values for clean s predicted to liquefy between.. m depth in comparison with the CRR vs K D curve for clean s proposed by Monaco et al. (). Vs SPT SPT Clean Monaco et al () Horizontal Stress Index, K D
CRR, CSR....... Monaco et al () No boundary curve by Monaco et al. () is reasonable for clean s but may be somewhat unconservative for sitly ss.. The liquefaction factor of safety from the Andrus & Stokoe () CRR vs V s correlation was consistently lower than predicted by other in-situ techniques. At shallow depths this appears to be due to the insensitivity of V s to increases in K o relative to other in-situ tests. At greater depths this is likely due to the higher silt content which has led to overconservatism in V s based assessments of liquefaction in silty ss at other sites (Liu & Mitchell ). Horizontal Stress Index, K D Fig.. Comparison of CSR vs K D values for silty s predicted to liquefy between.. m depth in comparison with the CRR vs K D curve for clean s proposed by Monaco et al. (). CONCLUSIONS Investigations at Treasure Isl involving CPT, DMT, SPT V s measurements provide a relatively rare opportunity to compare contrast the performance of these methods in predicting liquefaction. Based on the results of the field testing analysis described previously the follow conclusions are presented:. All the in-situ techniques were relatively consistent in predicting liquefaction within a depth range of m to. m below the ground surface. This result highlights the value of obtaining consistent results from multiple methods in assessing liquefaction hazards.. The liquefaction factor of safety predicted by the CPT DMT approaches was quite consistent in the depth range from m to. m despite the higher fines content in this zone for which the DMT correlation is not well calibrated.. At depths less than m with relatively clean s, the DMT predicted higher factors of safety than the CPT, although the CPT also showed increased resistance in this zone. The higher factors of safety from the DMT are likely a result of the increased sensitivity of the DMT to the higher K o values in this zone relative to the CPT.. Analysis of the CSR K D data points within the depth range from m to. m, where all methods predict liquefaction, suggests that the liquefaction. There is a significant need for increased research to underst better the influence of K o on liquefaction resistance curves from in-situ tests. This is particularly important in assessing liquefaction resistance after ground improvement which often increases both s density the horizontal earth pressure. REFERENCES Andrus, R.D. Stokoe, K.H., II. (). resistance of soils from shear-wave velocity. J. Geotech. Geoenviron. Eng.,, ASCE, (), -. Ashford, S.A., Rollins, K.M. (). TILT: The Treasure Isl Test, Univ. of California-San Diego, Report No.UCSD/SSRP /. Ashford, S.A., Rollins, K.M., Lane, J.D. () Blast-induced liquefaction for full-scale foundation testing, J. Geotech. Geoenviron.Engrg., ASCE, (), -. Baldi, G., Bellotti, R., Ghionna, V., Jamiolkowski, M., Marchetti, S. Pasqualini, E. (). "Flat Dilatometer Tests in Calibration Chambers." Proc. In Situ ', ASCE Spec. Conf. on "Use of In Situ Tests in Geotechn. Engineering", Virginia Tech, Blacksburg, VA, June, ASCE Geotechn. Special Publ. No., -. Harada, K., Ishihara, K. Orense, R.P., Mukai, J. (). Relations between penetration resistance cyclic strength to liquefaction as affected by Kc conditions. Proc, Geotechnical Earthquake Engineering Soil Dynamics IV, Sacramento CA, Paper, pp (in CD ROM). Idriss, I.M. Boulanger, R.W. (). "Semi-empirical procedures for evaluating liquefaction potential during earthquakes." Proc. th Int. Conf. on Soil Dyn. Earthquake Engrg. d Int. Conf. on Earthquake Geotech. Engrg., Berkeley, -. Ishihara, K., Iwamoto, S., Yasuda, S., Takatsu, H., (). of aniosotropically consolidated s. Prods. Ninth Int. Conf. on Soil Mech. Found. Engrg., Japanese Society of Soil Mech. Found. Engrg., Tokyo, Japan, Vol., -.
Liu, N. Mitchell, J. (). Influence of Nonplastic Fines on Shear Wave Velocity-Based Assessment of. J. Geotech. Geoenviron. Eng., (),. Marchetti, S. (). "In Situ Tests by Flat Dilatometer." J. Geotech. Engrg. Div., ASCE,, No.GT, -. Marchetti, S., Monaco, P., Totani, G. Calabrese, M. (). "The Flat Dilatometer Test (DMT) in Soil Investigations A Report by the ISSMGE Committee TC." Proc. Int. Conf. on Insitu Measurement of Soil Properties Case Histories, Bali,, official version reprinted in Flat Dilatometer Testing, Proc. nd Int. Conf. on the Flat Dilatometer, Washington D.C., April -,, -, R.A.Failmezger, J.B.Anderson (eds). Maugeri, M. Monaco, P. (). Potential Evaluation by SDMT. Procs.Second InternationalConference on the Flat Dilatometer, Washington D.C. p. -. Monaco, P., Marchetti, S., Totani, G. Calabrese, M. (). liquefiability assessment by Flat Dilatometer Test (DMT). Proc. XVI ICSMGE, Osaka,, -. Robertson, P.K. Wride, C.E. (). Evaluating cyclic liquefaction potential using te cone penetration test. Canadian Geotech. J.,, (), -. Rollins, K.M., Diehl, N.B. Weaver, T.J. (). Implications of V S -BPT(N ) correlations for liquefaction assessment in gravels. Geotechnical Earthquake Engineering Soil Dynamics III, Geotech.Special Pub.No., P. Dakoulas, M. Yegian& B. Holtz, eds., ASCE, Vol.I, -. Rollins, K.M., Hryciw, R.D., McHood, M.D., Homolka, M. Shewbridge, S.E. () Ground Response on Treasure Isl. The Loma Prieta, California, Earthquake of October, - Strong Ground Motion, USGS Professional Paper -A, Ed., Borcherdt, R.D., USGS, A-A Rollins, K.M., Gerber, T.M., Lane, J.D. Ashford. S.A. (). Lateral resistance of a full-scale pile group in liquefied s, J. Geotech. Geoenviron. Engrg., ASCE, (), -. Salgado, R., Boulanger, R., Mitchell, J. (). Lateral Stress Effects on CPT Resistance Correlations. J. Geotech. Geoenviron. Eng., (),. Seed, H.B., Idriss, I.M. (). Simplified procedure for evaluatingsoil liquefaction potential. J. Geotech. Engrg. Div., ASCE, (),. Seed, H.B. (). Soil Cyclic Mobility Evaluation for Level Ground during Earthquakes. J. Geotech. Eng. Div., ASCE, Vol. (), - Weaver, T.J., Ashford, S.A. Rollins, K.M. () Lateral resistance of a. m drilled shaft in liquefied s, J. Geotech.Geoenviron.Engrg., ASCE (), -. Youd, T.L., Idriss, I.M., Andrus, R.D., Arango, I., Castro, G., Christian, J.T., Dory, R., Finn, W.D.L., Harder, L.F., Hynes, M.E., Ishihara, K., Koester, J.P., Liao, S.S.C., Marcuson, W.F., Martin, G.R., Mitchell, J.K., Moriwaki, Y., Power, M.S., Robertson, P.K., Seed, R.B., Stokoe, K.H. (). Resistance of Soils: Summary Report from the NCEER NCEER/NSF Workshops on Evaluation of Resistance of Soils. J Geotech. Geoenviron. Eng., ASCE, (), -.