Nonlinear Finite Element Method And Genetic Algorithms Material Parameter Back-Analysis Research In Soft Soil

The soft soil distribute extensively in our country, there are much soft soil in the coast, the downstream of river or nearby the lake. Because of the high compressibility, low permeating ,long time of transforms, the settlement of the soft soil foundation and the velocity estimate is become an important problem in the engineering design. Especially when the density of surveys is not enough, the choices of parameters of the soil body can make more obvious influence to the calculated result. It is progressively be paid attention concerning about how to confirm the influence by multi-parameters in practical project, how to carry out the dynamic state construction instruction and how to carry on the check of designing.Firstly, according to BIOT Consolidation and FEM theories, adopting the model of Duncan-Chang and considering the influence of Stress Transfer the program ZBF in FORTRAN language has been worked, which can compute the displacement of X, Y and the hole pressure of soft embankment in each time, and it can imitate various load instance to construct, easy to consider the influence of settlement by the construction grade.Secondly, from the angle of dynamic design and information construction, a displacement back-analysis method for the certain some soft soil foundation parameter is found in this paper. With the south segment of Baoan big way in Baoan District Shenzhen City for study background, with K.6+475 monitor cross sections for research object, apply the Genetic Algorithms in the soft soil foundation displacement back-analysis, using those value of back analysis into the calculation model again to predict the forecast deformation. It provides a theoretical basis to the instruction of dynamic design and the construction for the soft soil engineering.The genetic algorithms is a kind of probability optimal method, doesn't depend on problems themselves. It takes full advantage of the characteristics of the genetic algorithms such as general use, robust nature, simultaneous disposal and good fitness. It also can recessively consider the coupling function between various parameters factor. A practical multi-parameter optimized back-analysis based on the genetic algorithms mode of soft soil foundation is suggested in this paper. A nonlinear FEM back-analysis program named ZGE in FORTRAN language has been worked. Some parameters in Duncan-Chang's model are achieved by displacement back-analysis. According to these parameters value, using the ZBF to predict the tendency of soil body transforms of next stage. In the example the relative error between calculation and measurement is controlled in 5%, and the prediction in the meantime is match engineering development regulation, demonstrate the Genetic Algorithms is applicable to the soil body parameter and it can be used to predict the displacement.