Evaluation of Seismic Parameters For An Earth-Quake Analysis

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

EVALUATION OF SEISMIC PARAMETERS FOR AN EARTHQUAKE ANALYSIS

Olli T. Ravaska1 and Panu A. Tolla2


ABSTRACT The Final Engineering Design for a cable-stayed 1347 m long bridge with 17 spans over the Cam river in Hai Phong, Vietnam, was made by a design team from the Finnish National Road Administration (Finnra) in 1997-1998. Subsoil on the site is composed of soft alluvial and flood plain deposits down to a depth of 30 metres, so that the only feasible foundations would be large diameter piles. One requirement for the bridge design was that an earthquake with a magnitude of M = 6.0, which may occur about 20 km from the bridge, should be taken into consideration. The special soil parameters needed for the earthquake analyses were the velocity of the shear wave, the dynamic shear modulus and the damping ratio. Because direct measurements on the site were not possible, seismic soil parameters for earthquake analyses were determined as based on laboratory tests on samples from the bridge site and on empirical correlations between geotechnical properties, SPT results and dynamic parameters. Laboratory tests performed in order to determine dynamic parameters were the resonant column test and the bender element test. INTRODUCTION Evaluation of seismic parameters for earthquake analysis as discussed in this paper deals with the Binh Bridge Project which includes the design and construction of a 17-span bridge about 1.3 km long over the Cam river in Hai Phong located on the coast of the Red River delta area in northern Vietnam, Figure 1. The bridge will be cable-stayed with a steel and concrete composite superstructure which is continuous from one abutment to another. The main spans are 100+260+100 m long and the approach spans on both sides of the cable spans have a length of 60 m and abutment spans of 50 m. The total length is 1347 m and the effective width of the deck is 22.5 m. The pylons are comprised of modified H-type concrete towers with inclined legs. The height of the towers is 101.6 m. The bridge will be founded on steel pipe piles the maximum penetration of which is 35-45 m. Hai Phong Province is situated at the south-west margin of the Caledonia-Katazia fold belt and next to the northern margin of the Red River Cenozoic Rift basin. The area has been affected by movements of the fault in the basin. Two seismic source zones are at a distance of 20-25 km from the city, with the potential of causing earthquakes with an intensity of 7 (MSK-64 scale) at the epicenter and intensities in the range 6-7 in the Hai Phong city area. Therefore the Institute of Geophysics at the Figure 1 : The Binh Bridge, modelled by Consulting Kortes Ltd. _______________________________________________________________________________________
1

Professor of Soil Mechanics and Foundation Engineering, Helsinki University of Technology, Dept. of Civil and Environmental Engineering, P.O. Box 2100, FIN-02015 HUT, Finland 2 Geotechnical Engineer, M.Sc. (Civ.Eng.), Finnish National Road Administration, Consulting Unit, P.O. Box 157, Helsinki, Finland

Vietnam National Centre for Natural Science and Technology required that in the earthquake analyses for the bridge design a magnitude of M = 6.0 should be used (Assessment, 1997).
The bridge was designed in 1997-98 by a design team from the Finnish National Road Administration (Finnra) based on the ground investigations carried out by the Vietnam Bridge and Road Centre (VIBROCE) in accordance with the specifications prepared by Finnra. Seismic risk and earthquake loading against structures were analyzed by Finnras design team in co-operation with the Institute of Geophysics and the University of Oulu, Finland. The bridge design was made by Finnras partner Consulting Kortes Ltd., Finland.

SUBSOIL CONDITIONS
Subsoil investigations on the site included SPTs, CPTs, pressuremeter tests, vane tests and sampling for laboratory testing. A detailed description of the soil investigations is presented by Tolla and Ravaska (1999). A longitudinal profile of the bridge site is presented in Figure 2. The subsoil on the site is predominantly soft silty soil of low to high plasticity, with a high content of organic matter. The thickness of these layers is

Figure 2 : Longitudinal profile of the bridge site. about 30-40 m. The topsoil consists of fills and muddy soil. Layer #1 consists of organic clayey and silty soils which are very soft to the depth of 6...10 metres. The water content varies in the range 50-60% and the organic content in the range 4-9%. Below the softest soil layers there is mostly a loose, continuous sandy interbed (layer #2). Below this is a complex and heterogeneous sequence of fine-grained sediments (layers #3 and #4). They are predominantly composed of silty clays and clayey silts, but particularly below the elevation -25 silt and sandy silt interbeds can be found. Water content, organic content and consistency vary over a very large range. These fine grained soil layers were found to be normally consolidated in accordance with the odometer test results. Below the fine grained layers there is a sandy layer #5 at the elevation of -32 to -54. It is generally in a dense or very dense state, but the grain size distribution and density vary randomly. The bore hole logs and the SPT records do not provide any uniform and bedding pattern for the sand formation. The bedrock surface (layer #7) was observed between the elevations of -41 and -55. The bedrock is composed of sand, silt and clay stones. Their weathering state varies, but a significant part is highly weathered and the rock is mostly very fractured. The bedrock is partly covered by clay and silt layers of low to high plasticity (layer #6) and 1 to 5 metres in thickness including coarse fragments. The layer is evidently almost completely decomposed (weathered) rock, and ranges from stiff to very stiff. The groundwater tables have been observed in four standpipes. In the layer #2, the groundwater table is near the ground level, but in the layer #5 it is at -5 metres.

DYNAMIC PARAMETERS General Because of seismic activity close to the Binh Bridge site, earthquake analyses were necessary to the structural design of the bridge. At the first stage, a SHAKE analysis was carried out in order to figure out the response of the layered subsoil to the earthquake excitation in the rock. The computer program SHAKE calculates the responses in a system of homogeneous, visco-elastic layers of infinite horizontal extent subjected to vertically travelling shear waves. As a result the program outputs the motions and response spectra for the motions in each sub-layer. The analysis requires the following soil parameters for each geotechnical layer: the unit weight and the velocity of seismic wave or dynamic shear modulus and damping ratio. After that, the structural analysis of the bridge was carried out by using the ANSYS program. Part of the parameters needed can be measured directly by means of index tests in the laboratory. Measurement of the most essential parameters, the velocity of the seismic wave, dynamic shear modulus and damping ratio all require dynamic tests and special equipment. The tests can be performed on undisturbed samples in the laboratory, or in situ by direct or indirect methods. The most common methods for measurement of the velocity of the seismic wave and dynamic shear modulus are the following: - direct measurement in a borehole using down-hole, up-hole or cross-hole measurement techniques - seismic refraction sounding in which S-wave and P-wave can be distinguished - laboratory measurements using e.g. resonant column or bender element techniques - empirical methods based on the penetration resistance; SPT and CPT borings are most often used Direct measurements are time consuming and require special equipment, and therefore empirical methods are widely used for these purposes. The indirect methods have also certain advantages. The tests are performed in situ, and therefore uncertainties such as disturbance of the samples are avoided. The parameters can also be evaluated from standard borings which are carried out for foundation engineering purposes. Because of the tight time schedule for the Binh Bridge Project, it was impossible to make direct measurements; therefore, indirect measurements and laboratory tests were used for the evaluation of the seismic parameters. Indirect methods A great number of empirical correlations between various boring methods and the velocity of the seismic wave or maximum dynamic shear modulus are available in the literature. The value of the shear modulus depends on the strain level and is, at its maximum, at a low strain level. Its maximum value is usually measured as an initial modulus in the hysteresis loop. A great deal of the correlation formulas are merely functions of an N-value measured in a SPT test, e.g. Eq. (1) (Imai et al., 1982): G max = 14.1( N 30 ) 0.68 (1)

where Gmax is the maximum shear modulus [MPa] and N30 the SPT blow count value/0.3 m penetration. According to Hardin and Drnevich (1972), the shear modulus is strongly dependent on three factors: the state of effective stress, strain amplitude and void ratio (or relative density). The maximum shear modulus can be calculated from the following formula: G max = 22.9 (2.973 e) 2 (OCR) K 0 ' 1+ e (2)

where Gmax is the maximum shear modulus [MPa], e the void ratio, OCR the overconsolidation ratio, K the factor which depends on the plasticity index and 0 the mean effective principal stress [kPa]. Based on numerous measurements of shear velocities and SPT tests, Ohta and Goto have presented a formula for seismic wave velocity (Seed et al., 1986): v s = 69 N j
0.17

D 0.2 F1 F2

(3)

where vs is the shear wave velocity [m/s], Nj the SPT N-value as measured in Japanese practice, D the depth [m] and F1, F2 factors depending on the soil type and the origin of the soil deposit: F1 = 1 for alluvial deposits and 1.3 for diluvial deposits; F2 = 1.0 for calys, 1.09-1.14 for sands and 1.15-1.45 for gravels. Converting Nj to the U.S. practice in counting the N-values makes it possible to compare the results.

The relationship between Gmax and vs is obtained from the equation:


G max =

2 vs g

(4)

For the Binh Bridge project simple formulas of the type Eq. (1) were rejected and those of Hardin and Ohta & Goto were taken as design formulas. In order to have comparable results from the SPT results it is usual that the SPT blow count values are normalized to the blow energy and the overburden stress. For the equipment used, this was made according to Canadian Foundation Engineering Manual (1992) using the equations: ( N1 ) 60 N C N 1920 C N = 0.77 log10 ' v where (N1)60 is the SPT N-value normalized to energy and effective overburden stress, N the measured blow count value and v the effective overburden stress. Laboratory tests Laboratory tests for the determination of dynamic parameters to be used in the seismic analyses were performed at the Geotechnical Laboratory of the Tampere University of Technology, Finland. The test methods used were the resonant column test and the bender element test. The bender element measurement techniques are based on a travel time principle with an oscillator, a rotation-prevented stepping motor at the top of an enlarged triaxial cell, and a pair of piezoceramic bender elements mounted in both the top cap and the bottom cap (Souto et al., 1994). Six soil samples from the depths of 3-38 m were taken for the laboratory investigations. Two of the samples were disturbed sand samples and four were undisturbed clay samples. The disturbed samples were compacted to a density corresponding to that in an undisturbed state. Each specimen was tested at four different stress levels under isotropic stress conditions. The strain range in the resonant column tests was selected to be as large as possible - taking into account the limitations set by the equipment - in order to correspond to the earthquake conditions in the best possible manner. The tests resulted in shear wave velocities, dynamic shear moduli and damping ratios at varying stress and strain 120 levels for all six tested samples. Figure 3 100 presents an example of the results obtained 80 from bender element tests and resonant Bender elem 60 column tests at different confining pressures. Resonant col 40 The soil type is clay taken from a depth of 9 20 m in the upper part of the layer #3. The 0 figure shows that the shear moduli obtained 35 70 150 250 from the resonant column tests are higher Cell pressure [kPa] than those obtained as a result of the bender element tests. It also shows clearly how the dynamic shear modulus increases with increasing confining pressure. Figure 3 : Dynamic shear moduli obtained from bender element tests and resonant column tests at various cell pressures, for a clay sample taken from a depth of 9 m. Comparison of the results A great amount of boring and laboratory data was available to aid in forming the design profiles for the dynamic analyses. Proper seismic parameters had to be evaluated for each soil layer in the geotechnical profile, Figure 2. At first, a comparison of the values obtained using various evaluation methods was made. The greatest number of comparable data was available at the sampling points, where the samples for the resonant column and the bender element tests were taken. In order to have the best possible correspondence, only those laboratory test results in which the cell pressure was about the same as the natural confining
Shear modulus [MPa]

(5)

pressure and the measurement was made at a small strain level, corresponding to the maximum value, were accepted. The comparison of the test results is presented in Figure 4. The horizontal axis presents the sampling depths. The first two depths at 3.0 m and 6.0 m represent layer #1, two second depths at 9.0 m and 9.2 m layer #3 and the next one at 25.6 m layer #4. All these layers are clayey silt or silty clay layers. The last two samples at depths of 36.5 m and 38.0 m are fine sand and coarse sand respectively, representing layer #5. Figure 4a shows that the shear wave velocities according to the bender element method and Hardins method correspond very well with each other - except for fine sand at the depth of 36.5 m in which the difference is 17 %. The Ohta & Goto method affords about 2548 % higher values than the bender element method independent of the depth, and the highest difference is again in the fine sand sample. Shear moduli were determined in the laboratory by both the bender element and the resonant column tests. For clays, the values obtained by the resonant column test were 10-62 % higher than those obtained by the bender element method. The same was not found for the sand samples, but the correlation was fairly a)
Velocity of shear wave [m/s] 500 400 300 200 100 0
9 9. 2 25 .6 36 .5 38 3 6

b)
350 300 250 200 150 100 50 0
9 9. 2 25 .6 36 .5 38 3 6

Shear modulus [MPa]

Bender elem Hardin Ohta&Goto

Bender elem Resonant col Hardin Ohta&Goto

Depth [m]

Depth [m]

Figure 4 : Comparison of test results: a) Velocity of the seismic wave. b) Shear modulus. good. With reference to the empirical methods, the Ohta & Goto method provided higher values than Hardins method, independent of the depth or soil type. The highest relative difference was 89 % in clay at the depth of 9.2 m. A rough correlation with the empirical formulas and laboratory tests for clay can be obtained, in that the results from Hardins method correspond to the bender element tests and those from the Ohta & Goto method to the resonant column tests. The results from the empirical formulas for sands were clearly higher than those from the laboratory tests, the Hardin results being much closer. Nevertheless, even they were 1050 % higher than the laboratory shear modulus values. Dynamic shear moduli for the earthquake analyses were determined as mainly based on the SPT test results, since they were available from both the whole area and from each soil layer. The laboratory test results afforded, however, a basis for averaging characteristic values for each soil layer in contradictory cases. Damping ratio The main factors affecting the damping ratio in sands are the strain level and the effective confining pressure to which it is subjected (Seed et al., 1986). The increase respective to both of them also increases the damping ratio - markedly at strain levels higher than 10-4-10-3. About the same strain levels are valid also for clays as reported by many researchers, recently e.g. by Cavallaro et al. (1999). For the Binh Bridge Project, the damping ratios were measured in the laboratory by the resonant column test. They were determined using the resonant frequency method (Dmmf) and the amplitude decay method (Dlog) (Souto et al., 1994). Two examples of the damping ratios are presented in Figure 5: sandy clay at the depth of 3 m (Figure 5a) and fine sand at the depth of 36.5 m (Figure 5b). The damping ratio for the sandy clay at a strain level of 10-5 is about 5 % increasing after that at higher strains. The same for the fine sand at the depth of 36.5 m is 3-4 %. Data at higher strain levels is missing due to the limitations set by the equipment used, but, as discussed above, higher values would have been obtained at higher strain levels. The estimated damping values for the dynamic analyses varied between 4 % at great depths and 15 % at the top. The analyses were carried out with damping ratios of 5 %, 7% and 10 % for each layer in order to figure out their effect on the dynamic loading for the bridge structural design.

Damping ratio [%]

a)

10 8 6 4 2 0
67 07 E 1. -07 09 E 3. -06 52 E 1. -06 09 E 4. -05 43 E 1. -05 17 E 2. -04 48 E04 E-

Dmmf Dlog

Damping ratio [%]

b)

5 4 3 2 1 0
07 07 06 06 05 05 EEEEEE22 05 53 21 22 06 E05

Dmmf Dlog

98

1.

03

3.

1.

3.

1.

4.

1.

6.

Shear strain

Shear strain

Figure 5 : Damping ratio versus shear strain. a) Sandy clay, depth 3 m. b) Fine sand, depth 36.5 m. CONCLUSIONS The geotechnical engineering design of the earthquake analysis for the Binh Bridge Project, including the evaluation of seismic soil parameters, was carried out as based on the field and laboratory investigation results together with relevant information obtained from research reports by acknowledged researchers. The parameters were evaluated using four different methods, the results of these being partly contradictory. The comparison of the results obtained by different methods gave, however, a good basis for the selection of the design values. Determination of the velocity of the seismic wave as well as the dynamic shear modulus both indirectly from the empirical formulas and directly from the laboratory tests provided comparatively consistent values for the silty layers situated from the ground surface to the depth of about 30-40 m. In particular, the shear wave velocities calculated in accordance with the bender element laboratory method and Hardins empirical method corresponded very well with each other. The Ohta & Goto boring method provided higher values independently in regard to depth. The highest difference was 17 % in sand layer #5 at a depth of 36.5 m. The dynamic shear modulus could be measured by use of four methods. A relatively reliable figure respective to the parameter values for the silty layers could be formed by using all these methods. For the sand layer, conversely, the results were inconsistent; the Ohta & Goto boring method in particular provided much higher values than the other methods. This was also taken into account when determining the parameters for the dynamic analyses. The damping ratios were measured by means of the resonant column test at a strain level of 10-7-10-4. Earthquake-induced shear strains may be much higher, and therefore also higher damping values than the measured ones were used in the dynamic analyses of the bridge. REFERENCES Assessment of the earthquake loading. The Binh Bridge project (1977). National Centre for Natural Science and Technology, Institute of Geophysics, Hanoi. Canadian Foundation Engineering Manual (1992). Third Edition. Canadian Geotechnical Society, Technical Committee on Foundations: 46-47. Cavallaro, A., Maugeri, M., Lo Presti, D., Pallara, O. (1999). Characterising shear modulus and damping from in situ and laboratory tests for the seismic area of Catania. Proc. of the II Int. Symp. on Pre-Failure deformation Characteristics of Geomaterials. Torino, Italy Sept. 1999. Jamiolkowski, Lancellotta & Lo Presti (eds). Balkema, Rotterdam, Vol 1: 51-58. Hardin B.O. & Drnevich, V.P. (1972). Shear modulus and damping of soils: Design equations and Curves. ASCE, Jour. of the Soil Mech. and Found. Eng. Div., Vol. 98, No. SM7, July 1972: 667-692. Imai, T., Tonouchi, K. (1982). Correlation of N-value with S-wave velocity and shear modulus. Proc. II Eur. Symp. on Penetration Testing, Amsterdam, Vol. 1: 67-72. Seed, H.B., Wong, R.T., Idriss, I.M. (1986). Moduli and damping for dynamic analyses of cohesionless Soils. ASCE, Journal of Geotechn. Eng., Vol. 112, No. 11, Nov. 1986: 1016-1032.

 

      
   !
 " #

   
 #
$

materials by the bender element and resonant column tests. Gotechnique 44, No. 3: 519-526. Tolla, P., Ravaska, O. (1999). Evaluation of soil parameters in the Binh Bridge project. Proc. of the XII Eur. Conf. on Soil Mech. and Found. Eng., Amsterdam, Netherlands,7-10 June, Edited by F.B.J. Barends, J. Lindenberg, H.J. Luger, L. de Quelerij & A. Verruijt., Vol. I: 623-630.

6.

You might also like