![[转载]21. <wbr>晒晒 <wbr>广州国际会议 <wbr>我的演讲稿 <wbr>（会上是脱稿的英语演讲）](http://simg.sinajs.cn/blog7style/images/common/sg_trans.gif "[转载]21. <wbr>晒晒 <wbr>广州国际会议 <wbr>我的演讲稿 <wbr>（会上是脱稿的英语演讲）")
Speech draft
Thank you, prof.Lv and prof.Guo. My name is zeng daming. I’m from Tongji university. I am very pleased to be here to join this forum. The topic of my presentation is Parameter identification and numerical analysis of shaking table test on liquefiable soil-structure interaction.
My presentation will include these five parts:
First, introduction; Second, summary of the shaking table test; Third, parameter identification; fourth, numerical simulation and the last: conclusions;
First, introduction.
Soil liquefaction has very often caused structure damage during past earthquakes. As shown in the two photos, the Niigata earthquake in 1964 and Kobe earthquake in 1995 demonstrated the damaging effects of liquefaction-induced loss of shear strength and lateral spreading within soils.
Dynamic test and analysis on soil-structure interaction in ground of liquefiable soil is one highlight of soil-structure interaction research, also is recognized difficulty.
Researchers have derived valuable conclusions from shaking table tests of soil-structure interaction models in liquefiable ground. With advances in model similitude theory and structural seismic testing technology, shaking table tests contribute more and more to soil-structure interaction research.
The parameter identification technology is also one of the important contents in dynamic soil structure interaction research. The response of soil layers in term of stress-strain behavior and damping characteristics can be obtained by the parameter identification technology based on the test records in the shaking table tests.
Also, numerical simulation has been used widely in the dynamic soil structure interaction research. And the parameter identification technology can provide optimized parameters for the numerical simulation.
The shaking table tests on soil-pile-structure systems were carried out in the State Key Laboratory of Disaster Reduction in Civil Engineering, Tongji University. (China, in June 2007.)
This paper presents the parameter identification of soil properties and the numerical simulation of the shaking table test for liquefiable soil structure interaction system.
Next, I would like to introduce the summary
As shown in the two photos, the prototype superstructure is a 12-story cast-in-place concrete frame with a single bay and a single span, which has 3 by 3 pile group foundation embedded in deep sand deposits.
All physical quantities were scaled using the similitude formulas according to Buckingham π theorem.
In shaking table test, the soil has to be confined within a box of limited size.
A flexible container is used to reduce the undesirable boundary effects.
Saturated sand overlain by a layer of silty clay is adopted as the model soil.
The superstructure and the pile foundation are made of micro-concrete and fine zinc-coated steel bar.
Properties of all materials were measured by material test before the shaking table test.
Accelerometers and strain gauges are used to measure the dynamic response of the superstructure, the foundation and the soil.
Pore pressure gauges are embedded in soil to measure the change of pore pressure.
Soil pressure gauges are used to measure the contact pressure between piles and the surrounding soil.
Ground shaking was applied as unidirectional in the x-direction of the shaking table. The records selected for the study include El Centro wave and Shanghai bedrock wave.
In the experiment, the peak acceleration is increased progressively. The peak acceleration and time interval is adjusted according to the similitude relation.
The table displays the complete loading sequences, which consisted of 5 levels of excitation. Before and after applying these acceleration levels, a white noise of small amplitude was applied to the model in order to determine the change in dynamic characteristics of the system.
The figure shows some of the test records. These are the acceleration and pore pressure ratio records obtained in the shaking table test at some of the points.
The pore pressure ratio is defined as the ratio between the water pressure generated during shaking and the initial effective vertical consolidation stress.
Next, I’d like to introduce the detailed parameter identification method to the shaking table test.
First, The identification procedure for dynamic shear stress and shear strain.
The soil deposits are assumed to deform as a shear beam model during seismic shaking.
This assumption leads to a realistic first order approximation to the low frequency modes of shear deformation that dominate in the flexible container.
The shear stress at the depth z can be solved using equation(1), in which ρis the mass density of the soil, and u double dots is the acceleration at depth ζ.
Utilizing a linear interpolation between accelerations at various depths, the shear stress at depth Zi and (Zi-1+Zi)/2 can be solved using equation (2) to (3) and their corresponding shear strain can be solved using equation (4) to (5).
This figure shows the shear stress-strain response curves identified at some of the measuring points.
In addition to stress-strain response curves, the soil stiffness and material damping as function of equivalent shear-strain amplitude can also be extracted from measured acceleration. The hysteretic stress-strain curves can be fitted to reproduce the same energy dissipation and shear stress at peak shear strain as a visco-elastic material.
The equivalent shear modulus G and damping ratio D during a shear stress-strain cycle can be evaluated by equation(6).
This figure compare shear modulus and damping ratio evaluated from shear stress-strain curves and that from material test.
The figure also show damping ratios evaluated from G/Gmax~γ curves.
We can see that the variation of shear modulus with shear strain amplitude, which was established from shaking table test, are in good agreement with the material results.
The damping ratios determined from shear stress-strain curves and G/Gmax~γ curves are noticeably higher than the damping ratio measured from material test.
Part 4, Numerical Simulation.
The soil's stress-strain relationship coincides with the laws of strain softening and can be expressed by hyperbolic mathematical model that can be expressed as the following formula: in which ar and br are fitting parameters.
The fitted G/Gmax~γ curve can be obtained from that obtained by the parameter identification technology using the formula above.
The calculation process using equivalent linear model is as follows:
when calculating, assume the dynamic shear modulus and damping ratio of each layer of the soil at first, then solve the corresponding equivalent dynamic shear strain γe ,
then find the corresponding equivalent dynamic modulus G and damping ratio D on the G/Gmax ~γcurve and D~γcurve based on the γe .
Repeat the above steps until the difference between the dynamic shear modulus and damping ratio after and before the round is with in the permitted range.
The paper adds the equivalent linear model of soil discussed above and the calculation process into the ANSYS program using the parametric design language of ANSYS. In that way the nonlinear simulation is realized.
The increment mode of pore water pressure of shanghai sand soil can be expressed as the follow equations.
In which, ∆u is the pore pressure generated within the time of ∆T; Nf is the vibration times when liquefaction occurs; ∆N is the equivalent vibration times in every time interval.
∆N can be solved by equation (14) to (15）.
Dynamic Analysis Method with Effective Stress by Iteratilon Step by Step in Every Time Interval is shown in detail as follows:
The static effective stress σx ,σy andτxy are worked out through static analysis.
The whole period of the seismic wave is divided into some equal intervals.
The initial G and the initial D are determined for each of element.
Analysis using the above mentioned dynamic constitutive model of sand soil is carried out by step by step iteration method in the first time interval.
∆N in this time interval and the accumulative value N are figured out using Eq.14 and Eq.15.
∆u* in this time interval and the accumulative value u* are figured out using Eq. 12 and Eq. 13 for each of element.
G and D considering the effect of the pore water pressure are figured out for each of element. And the calculated value is used as the initial value for the next time interval.
Using the restart function of ANSYS program, G and D are changed into the value worked out in the step (5). And the calculation of next time interval is carried out without exiting the program.”
In this way, the continuity of result can be ensured. The above step(2)~(6)should be repeated for each of time interval until the seismic wave finishes.
The above gradually iterated nonlinear method is added into ANSYS program using the parametric design language of ANSYS program. And the gradually iterated nonlinear method will be adopted in effective stress dynamic analysis latter.
The effective stress method considering the soil as equivalent linear material in divided time intervals is introduced.
ANSYS is adopted to conduct two-dimensional finite element effective stress analysis on shaking table test based on parameter identification of soil.
The soil and pile are simulated by two-dimensional plane strain element. The column and beam are simulated by two-dimensional beam element.
(1) The figure shows the comparision of accelerations obtained from calculation and test records under EL3 excitation. From the comparision of the time history of the acceleration, we can see that the calculation value coincides with experimental value well though there are certain differences in waveform and wave amplitude. In the mean time, we can see that as the distance between measuring points and the bottom increases, the peak acceleration of soil decreases gradually which shows that the liquefied sand soil mainly reduce vibration.
(2) This fig shows the comparision of pore water pressure obtained from calculation and test records under EL3 excitation.
The increasing trend of the time history of pore pressure ratio obtained by calculation and test records coincides with the amplitude basically.
This shows that adopting the calculation model and the soil constitutive model can simulate shaking table test well and the calculation model established is reasonable.
The last, conclusions.
(1)The time history of acceleration from the calculation coincides with those from test well though there are certain differences in waveform and wave amplitude. The liquefied sandy soil mainly reduces vibration.
(2)The time history of pore pressure ratio from calculation and test coincides with each other well in the amplitude which shows that adopting calculation model and the soil constitutive model can simulate shaking table test well and the calculation model established is reasonable.
That’s all. Thank you for your kind attention.