Influence of Site Conditions on Seismic Design Spectra for Bridges

Transcription

1 International Journal of Structural and Civil Engineering Research Vol. 6, No., May 7 Influence of Site Conditions on Seismic Design Spectra for Bridges Muhammad Tariq A. Chaudhary Civil Engineering Department, Kuwait University, Kuwait, Kuwait mtariqch@hotmail.com, tariq.chaudhary@ku.edu.kw foundations are suitable in this site class for upper range of Vs while deep pile foundations can be used for the lower values of Vs. Site class D has Vs from 75 m/s to 6 m/s. Pile foundations are commonly used for bridges in this site class. This study focused on the most commonly occurring soil classes C and D which are suitable for both shallow as well as deep foundations and are characterized by a wide variation in Vs (75 76, PI ( 6), OCR ( ) and ( 5 kpa). It is to be noted that site classes C and D roughly corresponds to JRA soil types SC-II and SC-III [] and EC-8 soil classes B and C [] respectively. Abstract Site conditions (geotechnical soil properties and geological setting) influences the surface response of strata to seismic ground motion. This fact is recognized in all seismic design codes by assigning spectral shapes based on site conditions. Due to variability in site properties and a myriad number of their combinations, design codes made a number of limiting assumptions. This paper attempted to test the bounds of some of these parameters (strata depth, Vs, Tg, PI, ICR) and compared the results with bridge design code. A number of limitations in the code were identified and corrections factors for some specific conditions are suggested. Index Terms acceleration spectra, bridges, soil properties, shear wave velocity I. INTRODUCTION Influence of site conditions on seismic design spectra is well recognized in current codes. code [] caters for these effects by characterizing the site conditions into six classes based on shear wave velocity in the top m depth (Vs) or alternative procedures based on average SPT N values or undrained shear strength of top m strata. Design spectrum for a bridge is constructed based on the mapped PGA of the site, short period spectral acceleration, Ss and long period spectral acceleration, S along with site modification factors FPGA, Fa and Fv. Effect of variation in soil properties like Plasticity Index (PI), Over Consolidation Ratio (OCR), effective stress ( ), depth of soil strata over bedrock and variation in stiffness of the bedrock are currently not included in the code based site characterizing and design spectrum construction processes. This study examined the influence of these geotechnical and site parameters on spectral acceleration, SA, used for seismic design of bridges. II. METHODOLOGY A. Background Site conditions are classified into six categories (A to F) in the code based on Vs. Site Classes A and B are rock sites with Vs more than 76 m/s. Shallow spread footings are commonly used for bridges in these site classes. Site class C represents very hard soil or soilrock with Vs between 6 and 76 m/s. Shallow spread Manuscript received May 7, 6; revised November 8, 6. 7 Int. J. Struct. Civ. Eng. Res. doi:.878/ijscer B. Procedure Sensitivity of seismic spectral acceleration response, SA; to variation in Vs, PI, OCR,, depth of soil strata and variation in bedrock stiffness; characterized by variation in VsR based on CSIR classification for rocks [] was undertaken in this study. The seismic bedrock characteristics considered in this study are classified as rock classes I to V in which class I is a very good rock (Vs > 5 and class V a poor rock (Vs = 6. Five generic soil profiles falling within the Vs range for soil classes C and D were selected from the literature [5] as depicted in Fig.. Variations in PI, OCR,, strata depth and Impedance Contrast Ratio (ICR) between soil strata and bedrock considered in the study are summarized in Table I. Twenty actual far-field ground motions varying in PGA from.6 to.7g were selected from literature [6], [7] to perform onedimensional non-linear seismic site response analysis. Acceleration response spectra of these strong motions is depicted in Fig.. Table II lists the salient features of the selected ground motions as well as group these into Design Basis Earthquakes (DBE), Functional Evaluation Earthquakes (FEE) and maximum credible earthquake (MCE) based on a design PGA of.g. Ground motions 8 to were scaled down to match MCE for the design PGA. More than analysis were carried out using -D site response analysis software STRATA [8] for various combinations of soil profiles, soil properties, strata depth and ICR. STRATA performs a -D linear/non-linear seismic response analysis of the soil column in the time domain and incorporates strain dependent non-linear shear modulus reduction and damping curves from a number of sources. In this study, modulus reduction and damping curves developed by Darendeli [9] were used

2 International Journal of Structural and Civil Engineering Research Vol. 6, No., May 7 and are depicted in Fig. for the soil classes included in the study. Figure. Acceleration response spectra of used strong motions. TABLE I. VARIATION IN SOIL PARAMETERS Parameter Vs ( PI (atm) Strata depth (m) Vs bedrock ( Figure. Shear wave velocity profile for various site class. Range 6 (C_high) 75 (C_avg) 5 (D_high) 75 (D_avg) 75 (D_low), 5, 6,, 6, 76, 5, 5, 5 TABLE II. GROUND MOTIONS USED IN THE STUDY Group (MCE) Group (FEE) Group (DBE) EQ Record ID Seismic event Edgecomb, NZ Year 987 Station Maraenui Primary School Magnitude 6.6 PGA (g) Fault distance (km).6 69 Vs ( 5 Oroville- 975 Medical Center Irpinia, Italy 98 Calitri Chi-Chi, Taiwan 999 CHY Spitak- Armenia 988 Gukasian Kobe, Japan 995 Shin Osaka Kocaeli, Turkey San Fernando 97 Ambarli Castaic - Old Ridge Route Landers 99 Joshua Tree Dinar, Turkey 995 Dinar 6... Superstition Hills 987 Poe Road (temp) Tabas, Iran 978 Dayhook Hector Mine 999 Hector Imperial Valley 979 Elcentro Array # Cape Mendocino 99 Rio Dell Overpass - FF Loma Prieta 989 Gilroy Array # Northridge- 99 Beverly Hills - 5 Mulhol 8 Manjil, Iran 99 Abbar 7. 9 Duzce, Turkey 999 Duzce 7. Kobe, Japan 995 Takatori Int. J. Struct. Civ. Eng. Res. 7.5 (.).5 (.5).6 (.5)

3 International Journal of Structural and Civil Engineering Research Vol. 6, No., May 7 (a) (b) Figure. Modulus reduction and damping curves for soils. m soil strata m soil strata _998 5 _5 5_998 5_ 5_5 5_5 6_998 6_ 6_5 6_5. C_high (Vs = _7 _76 5 _5 5_7 5_76 5_ 5_5 5_5 6_7 6_76 6_ 6_5 6_ _67 5 _5 5_67 5_ 5_5 5_5 6_67 6_ 6_5 6_ _8 5 _5 5_8 5_ 5_5 5_5 6_8 6_ 6_5 6_ _679 _5 _5 5_ _5.5 5_5 6_679 6_5 6_ _89 _76 5 _5 5_89 5_76 5_ 5_5 5_5 6_89 6_76 6_ 6_5 6_ _967 5 _5 5_967 5_ 5_5 5_5 6_967 6_ 6_5 6_5.5.5 D_avg. (Vs = 75.6 D_high (Vs = 5.8. C_avg. (Vs = 75 _75 _8 _5 _5 5_75 5_8 5_5 5_5 6_75 6_8 6_5 6_ Int. J. Struct. Civ. Eng. Res

4 International Journal of Structural and Civil Engineering Research Vol. 6, No., May 7 _5 _5 _76 5 5_5 5_5 5_76 5_ 5_5 6_5 6_5 6_76 6_ 6_5 D_low (Vs = _85 _76 5 _5 5_85 5_76 5_ 5_5 5_5 6_85 6_76 6_ 6_5 6_ Figure. Median value of spectral acceleration at surface for EQ to 8 (DBE, median PGArock=. g). m soil strata m soil strata D_high (Vs = _679.5 _5 _5.5 5_679 5_5 5_5.5 6_679 6_5.5 6_ _89 _76 5 _5 5_89 5_76 5_ 5_5 5_5 6_89 6_76 6_ 6_5 6_ Int. J. Struct. Civ. Eng. Res _967 5 _5 5_967 5_ 5_5 5_5 6_967 6_ 6_5 6_ D_avg. (Vs = 75 _67 5 _5 5_67 5_ 5_5 5_5 6_67 6_ 6_5 6_5 _8 5 _5 5_8 5_ 5_5 5_5 6_8 6_ 6_5 6_ C_avg. (Vs = 75.5 _7 _76 5 _5 5_7 5_76 5_ 5_5 5_5 6_7 6_76 6_ 6_5 6_ _75 _8 _5 _5 5_75 5_8 5_5 5_5 6_75 6_8 6_5 6_5.5.5 C_high (Vs = 6 _998 5 _5 5_998 5_ 5_5 5_5 6_998 6_ 6_5 6_

5 International Journal of Structural and Civil Engineering Research Vol. 6, No., May 7. D_low (Vs = _85 _76 5 _5 5_85 5_76 5_ 5_5 5_5 6_85 6_76 6_ 6_5 6_5.. _5 _5 _76 5 5_5 5_5 5_76 5_ 5_5 6_5 6_5 6_76 6_ 6_ Figure 5. Median value of spectral acceleration at surface for EQ 9 to (MCE, median PGArock=.8 g). The choice of PI values as listed in Table I were based on the generally accepted limit for sand, medium plastic and highly plastic soils []. OCR was kept constant (= ) for all soil profiles based on the observations of Guerreiro et al. []. Two values of confining pressure were considered: atm for sandy soils (PI=) and atm for soils with PI= 5 and 6. The choices of confining pressure were made based on the observations of Guerreiro et al. [] and to limit the number of simulation cases to a realistic number. III. acceleration increased with a decrease in Vs for soil class C while it decreased with a decrease in Vs for soil class D. This observation is contrary to the current code practice that calls for an increased spectral acceleration value for softer soils with less Vs. RESULTS & DISCUSSIONS Salient results of the numerical simulations are presented and compared with the current code provisions in this section. Fig. and Fig. 5 present the results of response spectra for the selected five site classes for m and m depths of soil strata at the strata surface. Fig. presents the results for ground motions to 8 which had a maximum PGA value of.66g and a median value of.g and were representative of DBE ground motions. Fig. 5 depicts the results for ground motions 9 to with a maximum PGA of.5g and a median value of.8g and were representative of MCE. Design spectra based on code [] is also superimposed on the computed spectra in Fig. and Fig. 5. In the legend of Fig. and Fig. 5, the first number refers to the PI value and the second to the bedrock shear wave velocity. The following observations were made regarding the computed response spectra. B. Influence of Bedrock Vs on Spectral Acceleration Shear wave velocity of the bedrock (Vs_bedrock) seems to have a pronounced effect on the spectral acceleration as depicted in Fig. 7. This is especially true for soil types with higher Vs values. The effect of bedrock shear wave velocity on the spectral response of the soil strata can be characterized through impedance contrast ratio (ICR). ICR is defined as: where is the unit weight, Vs is the shear wave velocity and subscripts R and S refers to parameters of the bedrock and the soil layer above it respectively. For the analysis cases investigated in this study, ICR varied between and for type C soils and between and 8 for type D soils for m deep strata, while it was between and 7 for type C soils and between and 6 for type D soils respectively for m soil profiles. It can be noted in Fig. 7 that the effect of ICR is very negligible for soil D_low as maximum spectral acceleration varied between.6g and.9g. The maximum impact was observed for soil type D_high with a variation between.g and.g. Referring to Fig. 5 and Fig. 6, it was observed that increase in SA was only due to the increase in ICR for soil C_high and there was almost negligible Figure 6. Variation in SA for various soil types with strata depth & PI. A. Influence of Soil Type Fig. 6 depicts the variation in surface spectral acceleration for various soil types. Maximum spectral 7 Int. J. Struct. Civ. Eng. Res. Figure 7. Variation in SA for various soil types with strata depth & Vs_bedrock.

6 International Journal of Structural and Civil Engineering Research Vol. 6, No., May 7 contribution to increase in SA due to increase in PI of soil. Soil C_avg also exhibited a similar trend. However, in this case, PI of soil also played a little role in increasing the SA. An increase in the bedrock shear wave velocity generally resulted in increased spectral acceleration. An increase in spectral acceleration of more than.5 times was noted for a bedrock V s increase from 998 m/s to 5 m/s for m soil depth while it was.8 times for the case of m deep profile as depicted in Fig. 6 and Fig. 7. C. Influence of PI on Spectral Acceleration Referring to Fig. 6, it was observed that spectral acceleration was independent of PI for soil class C_high and moderately influenced soil C_avg and D_low for both depths of strata. SA was strongly influenced by PI for soil classes D_high and D_avg. The trend observed for soil Class C was reversed for soil class D for which dependency shifted from being strongly dependent on bedrock Vs to strongly dependent on PI with a decrease in V s. For example, soil class D_low showed a weak dependency on bedrock Vs but was strongly influenced by PI. Spectral acceleration for this soil increased with increasing PI. D. Effect of Strata Depth It was observed in Fig. 6 and Fig. 7 that maximum value of spectral acceleration was more or less the same for both strata depths for all soil types with SA value being slightly larger for m deep strata. However, for the three soil types with the largest V s values (C_high, C_avg. and D_high), the plateau of spectral acceleration was generally higher in the deeper strata for both DBE and MCE as depicted in Fig. and Fig. 5 respectively. TABLE III. THEORETICAL AND NUMERICAL VALUES OF GROUND PERIOD (TG) Soil Type V s ( V s ( H/V s H/V s [H=m] [H = m] T g (sec) Fig. ( m) Fig. (m) C_high C_avg /.5 D_high /. D_avg /.8 D_low /..5/.8 E. Effect on Natural Period of Strata Peaks of spectral acceleration occurred at different periods for the considered soil profiles. This is expected as time period of the soil column can be computed from the well- known quarter wavelength principle as: T g = H/V s and with a decrease in V s, T g is expected to be more. Comparison of theoretical and numerically computed values of ground period are listed in Table III. Numerical values of T g were found by the peak picking method from the spectra in Fig.. It can be noted that the theoretical values of T g, computed based on a strata depth of m and V s, compared well the numerically determined values for m deep strata. However, for m deep strata, travel time based shear wave velocity for the whole strata depth (V s ) gave Tg values that compared with the spectra values for all soil types except D_low. All soil types in the m deep strata exhibited multimodal response except for soil C_high. This multi-modal response was attributed to the reflections of seismic waves within the deeper strata which was absent in the shallower strata. IV. COMPARISON WITH CODE BASED DESIGN SPECTRA A. Design Acceleration Spectra Design spectra for DBE and MCE calculated using code are overlaid on the SA graphs in Fig. and Fig. 5 respectively. A large disparity was observed between the design spectra and the numerically computed spectra for various soil types. Generally, the maximum spectral acceleration occurred for soil with highest PI and highest ICR for both strata depths. Current code does not consider dependency of spectral acceleration on PI or ICR in the creation of design spectrum. Several researchers have pointed out discrepancies in the construction of design acceleration spectra in current codes [] []. However, these studies did not consider the combined influence of PI, strata depth, T g and ICR on surface SA. Salient differences in the design spectra with the current study are summarized below. ) Influence of PI Sand and clays exhibit significantly different shear stress-strain and damping behavior with increasing seismic acceleration. code assigns spectral acceleration coefficients based on V s only. Therefore, the fact that soil type (sand or clay) is not considered in the selection of these parameters despite the fact that either type of soil can have the same V s. Influence of PI is generally insignificant for soils with very high V s. However, variation in PI has significant impact on SA in other soils as discussed earlier. ) Influence of strata depth The current study included strata depths of m and m and it was found that strata depth did not significantly influenced the maximum magnitude of SA. Therefore, this effect can be neglected for SA amplitude as in the current code. However, it was observed that deeper strata depth elongates the SA plateau over a longer period corresponding to larger T g of deeper strata. Hence, a correction is required to extend the maximum SA value over a broader range of T. 7 Int. J. Struct. Civ. Eng. Res.

7 International Journal of Structural and Civil Engineering Research Vol. 6, No., May 7 ) Influence of Tg Natural period of the strata strongly influenced the SA around the Tg resonance peaks. Current code attempts to consider this effect through the parameter T which is defined as SD/SDS. T is the period up to which the maximum SA stays constant and starts to decrease after this time period. It was observed in Fig. and Fig. 5 that the current definition of T is insufficient to cover the period range with maximum SA. ) Influence of ICR The current provisions of code do not seem to cater for values of ICR more than as evidenced from Fig. and Fig. 5. ICR influences the soil strata with relatively larger values of Vs and has a minimum influence on soils with smaller Vs. ) Modified design spectra Fig. 8 and Fig. 9 presents the modified design spectra for soil type C_avg and D_avg respectively. V. The following conclusions are drawn from this study: Acceleration response spectra for two levels of earthquakes (DBE with PGA. g and MCE with PGA between.g and.8g) were computed for two depths (m and m) of soil strata for five soil types with Vs varying between 75 m/s and 6 m/s. Acceleration response spectra exhibited dependence on soil PI, Vs, strata depth, Tg and ICR to varying degrees in different soil types. Current code does not consider influence of all variables in constructing design acceleration spectra and leads to non-conservative results for some soil types. Correction factors were proposed to the current procedure and example of modified design spectra were provided. Median.5 med.+stdev Modified.5 ACKNOWLEDGEMENT.5.5 This work was supported by Kuwait University, Research Grant No. EV/5. Figure 8. Modified spectrum for soil C_avg. for DBE. REFERENCES.5 Median [] (). LRFD Bridge Design Specifications. American Association of State Highway and Transportation Officials, Washington, DC. [] JRA (). Design Specifications for Highway Bridges, Japan Road Association, Tokyo, Japan. [] CEN (), European Committee for Standardization TC5/SC8/, Eurocode 8: Design Provisions for Earthquake Resistance of Structures, Part.: General rules, seismic actions and rules for buildings, PrEN998-. [] Z. T. Bieniawski, Geomechanics classification of rock masses and its application in tunnelling, Proc. rd International Congress on Rock Mechanics, vol., no., pp. 7-, 97. [5] J. Douglas, P. Gehl, L. F. Bonilla, O. Scotti, J. Régnier, A. M. Duval, and E. Bertrand, Making the most of available site information for empirical ground-motion prediction, Bulletin of the Seismological Society of America, vol. 99, no., pp. 55, 9. [6] ATC-6, Quantification of building seismic performance factors, ATC-6 Project Report prepared by the Applied Technology Council for the Federal Emergency Management Agency, Washington, DC, 9. [7] M. T. A. Chaudhary, Seismic response of bridges supported on shallow rock foundations considering SSI and pier column inelasticity, KSCE Journal of Civil Engineering, pp. -, 6. [8] A. R. Kottke and E. M. Rathje, Technical Manual for Strata, PEER Report 8/, University of California, Berkeley, California, 8. [9] M. B. Darendeli, Development of a new family of normalized modulus reduction and material damping curves, PhD thesis, University of Texas, Austin,, p. 9. [] D. M. Burmister, Principles and techniques of soil identification, in Proc. 9th Annual Highway Research Board Meeting, National Research Council, Washington, DC., 99, pp. -. [] P. Guerreiro, S. Kontoe, and D. Taborda, Comparative study of stiffness reduction and damping curves, in Proc. 5th World Conference on Earthquake Engineering, Lisbon, CD ROM,, pp. -. med.+stdev Modified CONCLUSIONS Figure 9. Modified spectrum for soil D_avg. for DBE. B. Proposed Changes to the Design Spectra ) Correction for PI Correction for PI is required for soil classes D_high and D_avg only. Although, soil type D_low exhibited the most dependence on PI, the current code design spectra adequately caters for this effect in this type of soil. ) Correction for Tg Correction for Tg is needed for soils with deeper strata and lower Vs values. This correction is applied by extending the limit of the flat part of the design spectra to Tg based on deeper strata. ) Correction for ICR ICR affects SA the most for soil types with higher Vs. It is proposed to increase the SA value in the flat part of the design spectra corresponding to ICR of which also coincides with the median value of computed SA for both strata depths in such soil types. 7 Int. J. Struct. Civ. Eng. Res.

8 International Journal of Structural and Civil Engineering Research Vol. 6, No., May 7 [] G. Andreotti, G. C. Lai, F. Bozzoni, and L. Scandella, New soil factors for the italian building code (NTC8) derived from D fully stochastic ground response analyses, Italian National Association of Earthquake Engineering, At Padova, Ital, [] R. P. Dhakal, S. L. Lin, A. K Loye, and S. J. Evans, Seismic design spectra for different soil classes, Bulletin of the New Zealand Society for Earthquake Engineering, vol. 6, no., pp , [] K. Pitilakis, E. Riga, and A. Anastasiadis, New code site classification, amplification factors and normalized response spectra based on a worldwide ground-motion database, Bulletin of Earthquake Engineering, vol., no., pp ,. Muhammad Tariq Chaudhary, born in Pakistan, received his BS (Hons) degree in civil engineering from University of Engineering and Technology, Lahore, Pakistan (99); MS in structural engineering from University at Buffalo, NY, USA (99) and PhD in civil engineering from the University of Tokyo, Japan (999). He is a practicing structural engineer and academic currently working in the Department of Civil Engineering at Kuwait University, Kuwait. He has previously worked as Project Manager and Senior Structural Engineer with Obayashi Corporation, NESPAK and DNCE. He has published more than forty research papers in refereed journals and international conferences. His areas of research and interest are: seismic design, structural health monitoring, sustainable design and construction, soilstructure interaction and structural condition evaluation and rehabilitation. Dr. Chaudhary is a registered Professional Engineer in USA and Canada and a LEED Accredited Professional in USA. He is a Fellow of the Institution of Engineers Pakistan and a Member of American Society of Civil Engineers. 7 Int. J. Struct. Civ. Eng. Res.