Loess Nonlinear Constitutive Model Parameters

With the development of the main engineering, we need educe the parameter by displacement back-analysis. Now, we have done much research in the problem of liner elastic model, and obtain much achievement. But less research have been done in the problem about nonlinear model. The research on the back analysis of nonlinear model is in the stage of start. We have not some certain appropriable method to solve this problem. More less job has been done on the loess nonlinear model.This paper use a method to educe the parameter by coincide optimization and forward-analysis. This method is genetic algorithm-a new method to make optimize, it develops very fast lately and has been largely used in function optimizing. According to its basic ideology, I write a C program to complete soil parameter optimize. The forward-analysis method is finite element method. By some software I realize the forward -analysis calculation, and realize the calculation of Duncan-Chang model. It's feasible to execute displacement back-analysis by Genetic Algorithms by engineering example.This paper provides a process: according to the experimental data from the groundsill load experiment, we can identify the parameter of the loess nonlinear model by using the genetic algorithm before construct the building. By the parameter we can forecast the groundsill distortion.This paper advances the study of the parameter back-analysis of loess nonlinear model from a new side. It has certain reference value and theory significance for how to select the parameter of loess nonlinear model and promoting the stabilization analysis precision of soil distortion.