You are on page 1of 11
406 Depth (m) N Soil Type aa ll l Fill + (9.14)] orthe maximum cyclic shear strai tip resistance, can be used to estimate postliquefaction volumetric strain (Figure 9.54). gration of these volumetric strains over the thickness of the liquefied layer produces ground surface settlement. Liquefaction Ch Loose sandy fill 10 Recent bay mud 30 Dense sand Figure E97 a) 8) @ ) 6) a Layer DepthRange(m) = V(Myko CSR fH) AH Am) 1 5 D Sea Diet 30. 0.030. 2 9 tia oes N24 0.043, 3 8 86 0.18028 0.042 4 JP 72 pe 01h 02 0.003 O.118 ‘The voluimetric strain can then be estimated using Figure 9.53. The settlement of each then computed as the product of the volumetric strain and the layer thickness. The actual settlements in the vicinity of this site were on the order of 12 to 15 cm. Some; actual settlements resulted from densification of the dry soil above the water table; that ment is not accounted for in this example. In an alternative approach, either the factor of safety against liquefaction [eq in, and the relative density, SPT resistance, or Sec. 9.6 Effects of Liquefaction 405 generated by the earthquake. Laboratory experiments have shown that the volumetric strain alter‘initial liquefaction varies with relative density and maximum shear strain. Tokimaten and Seed (987) used a correlation between (W;}gy and relative density and an estimate of the shear strain potential of liquefied soil from (WV, gp and cyclic stress ratio (Seed et al. 1984) © produce a chart (Figure 9.53) that allows the volumetric strain after liquefaction ina M = 7-5 earthquake to be estimated directly from the cyclic stress ratio and SPT resistance. For earthquakes of other magnitudes. an equivalent cyclic stress ratio, CSR,,, can be determined from equation (9.2) and Table 9-2. Note that the volumetric strain after liquefaction can be s high as 2 to 3% for loose to medium dense sands and higher for very loose sands: a 5-m thick (16 ft) layer of very loose sand produced settlements of 50 to 70 cm (20 to 28 in.) in Hachinohe, Japan following the Tokachioki earthquake (M = 7.9) in 1968 (Ohsaki, 1970); Settlements of 50 to 100 cm (20 to 39 in.) were observed on Port Island and Rokko Island in Kobe, Japan following the 1995 Hyogo-ken Nanbu earthquake. If sand boils are produced, Postearthquake settlements are likely to be irregular. The resulting differential movement can be damaging to structures. pavements, or pipelines on or near the ground surface. 06 Volumetric strain (%) hosaa 2 1 0s Hl os 04 75 0.3 csr. 02 Figure 9.53 Chart for estimation of volumetric strain in saturated sands from cyclic stress ratio and standard penetration resistance. (After Tokimatsu and Seed, 1987. Evaluation of settlements in sand due to earthquake shaking, Journal of Geotechnical Engineering, Vol. 113, No. 8 Reprinted by permission of ASCE.) 0.1 0 HOFER 25 Figo: <4) Example 9.7 Significant settlement was observed in the Marina District of San Francisco following the 1989 Loma Prieta earthquake. Much of this settlement was determined to have resulted from densi- Heation of hydraulic fills that were placed to reclaim the area from San Francisco Bay in the 1890s. Subsurface conditions at one location in the Marina District are shown in Figure E9.7 (O'Rourke et al., 1991), Estimate the settlement that would have been expected at this location in the Loma Prieta earthquake. A Peak acceleration of 0.20g was measured in the vi inity of the Marina District. Solution |The subsurface profile shows measured (uncorrected) SPT resistances. Assuming that the SPT resistances were measured using standard procedures (E, = 0.60), corrected SPT resistances can be computed using equation (6.30) (column 4 below). The simplified pro- cedure of Seed and Idriss (1971) can be used to estimate the cyclic stress ratio (column 5 below). Volumetric Strain (%) N o © Tokiratsu-Seed (1987) ——— Equation (2) 5 10 15 20 6 x (Neo Figure-2. Relationship between volumetric strain and (Ni)us xs —Vaide SPT Jam voli 6 teva cero cums 2063 [ MISSOURI S < . Missouri University of Science and Technology Scholars' Mine Intemational Conferences on Recent Advances 2001 - Fourth international Conference on in Geotechnical Earthquake Engineering and Recent Advances in Geotechnical Earthquake Soil Dynamics Engineering and Soil Dynamics 29 Mar 2001, 7:30 pm - 9:30 pm Assessing Probabilistic Methods for Liquefaction Potential Evaluation — An Update C. Hsein Juang Clemson University, Clemson, SC Tao Jiang Clemson University, Clemson, SC Ronald D. Andrus Clemson University, Clemson, SC Der-Her Lee National Cheng Kung University, Taiwan Follow this and additional works at: https://scholarsmine.mst.edu/icrageesd G Part of the Geotechnical Engineering Commons Recommended Citation Juang, C. Hsein; Jiang, Tao; Andrus, Ronald D.; and Lee, Der-Her, "Assessing Probabilistic Methods for Liquefaction Potential Evaluation — An Update" (2001). International Conferences on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics. 18. https://scholarsmine.mst.edu/icrageesd/04icrageesd/session04/18 This Article - Conference proceedings is brought to you for free and open access by Scholars’ Mine. It has been accepted for inclusion in international Conferences on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics by an authorized administrator of Scholars’ Mine. This work is protected by U. S. Copyright Law. Unauthorized use including reproduction for redistribution requires the permission of the copyright holder. For more information, please contact scholarsmine@mst.edu. ASSESSING PROBABILISTIC METHODS FOR LIQUEFACTION POTENTIAL EVALUATION - AN UPDATE C. Hsein Juang, Tao Jiang, and Ronald D. Andrus Clemson University Clemson, South Carolina-USA-29634 ABSTRACT Der-Her Lee National Cheng Kung University ‘Tainan, TAIWAN This paper presents. an assessment of existing probabilistic methods for liquefaction potential evaluation. Emphasis is placed « comparison of probabilities of liquefaction calculated with four different methods. Two of these methods are based on SPT, and t1 other two are based on CPT. In both SPT- and CPT-based evaluations, logistic regression and Bayesian techniques are applied to mi factor of safety to probability of liquefaction. ‘The present study shows that the Bayesian approach yields more conservative resul than does the logistic regression approach, although results from the two approaches are quite comparable. Discussion of 11 procedure for risk-based liquefaction potential evaluation is also presented. INTRODUCTION Site-specific probabilistic assessment of earthquake-induced liquefaction potential of soils using in situ test data is an important task facing geotechnical engineers. Results of such probabilistic assessments may lead to better engineering decisions. Liao et al. (1988) provided a foundation for probabil assessments of liquefaction potential based on logistic regression analyses of the SPT-based field performance records. Youd and Noble (1997) extended this approach by considering earthquake magnitude in the logistic regression analysis and using an extended database. Toprak et al. (1999) conducted logistic regression analyses. of the field performance database compiled by the U.S. Geological Survey. The study by Toprak et al. (1999) also extended the approach to deal with CPT-based database. Unlike the aforementioned approach, Juang et al. (1999; 2000a) used reliability analysis and Bayes' theorem (o derive the probability of liquefaction. With their approach, a Bayesian mapping function that relates reliability index and/or factor of safety to the probability of liquefaction was first developed based on field performance data. The Bayesian mapping function can then be used to predict the probability of liquefaction for a given set of site specific information. Juang and Jiang (2000) extended carlier studies on Bayesian ‘mapping function, and found that mapping function could be developed using the distributions of the calculated factors of safety, instead of reliability indexes. They have developed Paper No. 4.23 Bayesian mapping functions for the SPT-based method | Seed et al. (1985), and for the CPT-based method | Robertson and Wride (1998). Juang and Jiang (2000) al compared probability curves developed from their Bayesi ‘mapping functions with those obtained by Toprak et al. (199 from logistic regression. Juang and Jiang (2000) adopted the magnitude scaling, fact (MSF) defined by Idriss (1999). The mapping functions th developed were compared to logistic regression equatie developed by Toprak et al. (1999) using different sets of d: than those used for the development of the mapping functiot In the present paper, the work by Juang and Jiang, (2000) refined. Here, the MSF recommended by the 1996 Natior Center for Earthquake Engineering Research (NCEE workshop (Youd and Idris, 1997) is followed. In additic the logistic regressions are performed using the same data s as used in the development of mapping functions. For the convenience of description, the CPT-based method Robertson and Wride (1998) is referred to hereinafter as CF RW method. The updated Seed and [driss (1971) SPT-ba: method, as presented in Youd and Idriss (1997), is referred hereinafter as SPT-SI method. The CPT-RW and SPT: methods are the only two deterministic methods considered the present stucy for developing mapping functions, althou other deterministic methods may be used. The Bayes mapping functions developed based on the CPT-RW and SF SI methods are referred to herein as the CPT- and SPT-bat Bayesian mapping functions, respectively. DETERMINISTIC METHODS In the liquefaction evaluation, the seismic load is generally expressed in terms of a cyclic stress ratio (CSR), which may be calculated as (modified from Seed and Idriss, 1971): can,y=062(%} [S22 ey/nsr a x & where c, is the total vertical stress at the depth in question, 6’, is the effective vertical siress at the same depth, amex is the peak horizontal ground surface acceleration, g is the acceleration due 10 gravity, MSF is the magnitude sealing factor, and ry is the stress reduction factor. The term MSF is used to adjust the calculated CSR to the reference earthquake magnitude of 7.5. Note that the convention for adjusting the effect of earthquake magnitude is to modify the cyclic resistance ratio (CRR) with MSF. However, it is more logical to include MSF in the calculation of CSR, since both are seismic load parameters, whereas CRR represents soil resistance (Juang et al., 2000b). ‘The term 14 provides an approximate correction for flexibility of the soil profile. In this study, the values of ry are calculated using the Liao et al. (1988) equation: ty = L.0-0.00765 2, forzs 9.15 m (2a) ty = 1174-00267 z, for91Sm

You might also like