Slope Soil Property Back-analysis Based On Genetic Algorithms

Nowadays, with the rapid development of freeways, the slope, act as the important constitution of the freeways, is playing an important role in research, design, construction and protection. Because construction destroys the slope, and the stability, the destruction and the protection of the slopes were not studied enough, it cause many buildings failure during the period of construction or running, that increase the investment and impact on economy. From the angle of dynamic design and information construction, a displacement back-analysis method which applies monitor information to determine soil properties and direct design and construction of slope excavation is found in this paper. On the basis of the item 'the research of stability and deformation of high slope', which studied by Fujian highway-bureau and Fuzhou University. The slope is located in Yong-Zhang freeways. A method based on the genetic algorithms has been put forward for the displacement back-analysis of slope, and some parameters can be identified, which provides for next stage calculation, treatment, ect. The genetic algorithms is a kind of probability optimal method, don't depend on problems themselves and needn't requesting assumptions such as continuance, the function is derivative and only one peak value. It takes full advantage of the characteristics of the genetic algorithms such as general use, robust nature, simultaneous disposal and good fitness, specially, in solving the problems which aim function has more than one peak value. A practical displacement back-analysis mode is suggested in this paper, based on the genetic algorithms .Exerting non-linear finite element theorem and using FORTRAN to write the GAEF program. The four parameters ,,, in Duncan-Chang's model are achieved by displacement back-analysis. A case is calculated through GAEF program to prove the stability and good astringency of GAEF program, and display the method based on the genetic algorithms has been put forward for the displacement back-analysis of slope is credible. Based on the back-analysis data of some excavation stage, through the non-linear finite element program NEF, we can predict the tendency of next excavation stage. The relative error between calculation and measurement is controlled in 25%, demonstrate the estimate prediction is successful.