The Research Of Roadbed Settlement On Soft Ground For Highway And Realization Technique Of Expert System

Settlement is the key problem in building highway on soft ground, concerted with the whole project success. Thereby, it is very important to control settlement to satisfy specification. We would like to predict exactly settlement of embankment.The principle and calculation methods for analyzing ground settlement of soft clay foundations appear frequently since Terzahi's consolidation theory was published. There are so many methods, such as semi-empirical formula, numerical analysis and so on. Each method has its advantage and disadvantage, so the engineer should choose a calculation method according to actual engineering need. This thesis analyzes and compares the applicabilities and using effects among these existing methods. At the same time, we makes use of the non-linear mapping and learning abilities of ANN, and suggests a new method of Wavelet Neural Network based on PSO to calculate the last settlement of soil clay of roadbed. This thesis uses different methods to forecast the some representative sections in the background of FenGuan Highway, and compares the computing results, discusses the applicabilities and precision of different forecast models.Combined with the existing methods and new research production, directed by modern software engineering theory, we developed successfully the expert system software for prediction of soft ground, which based on .net environment. Through the relevant engineering data, the conclusion is proved its correctness.In the end, some suggestions and views are presented for the further study.