The Application Of Genetic Algorithm In Initial In-Situ Stress Field Analyzing And Calculating

Initial in-situ stress field of lithosome is one of requirement data in undergroundengineering wall rock stabilization analyzing and supporting structure designing, andis one of crucial factors of high rock slope stabilization or slope excavation designingand supporting. Field actual ground stress measurement is the most straightforward way tosupply in-situ stress field. It is impossible to carry mass measurement for thelimitation of site and expenditures. The origin of in-situ stress field is complex andin-situ stress field has multi-influence-factors. The result of measurement of everysurvey point reflects local stress field to a large extent, and is affected by error, whichlead to some extent discreteness. It is an effective way to inverse the initial in-situstress field basing on actual survey ground stress data. Basing on improve the traditional regress analysis method, the tectonic stress ofinitial in-situ field is analogized by superimpose the load factor of uniformitydistribution and trigon distribution. The main influence factors are determined bystepwise regress method. At the same time, the insignificant influence factors areeliminated. The values of main influence factors are got by calculate with GeneticAlgorithm. Applied these values to Finite Element model and calculated with thismodel, the initial in-situ stress field is obtained. To describe values of initial in-situ stress field of certain position or section, theNeural Net-work method is adopted. Comparing with traditional interpolation method,the Neural Net-work method is more simple and reliable. It is very convenient todescribe the initial in-situ stress field through computer visualization technique.Therefore, it must find its way in engineering practice. The investigation results show that using the union method, which is combinethe load factor of uniformity distribution with trigon distribution, is better insimulating tectonic stress than traditional method. The optimum solution, calculatingwith Genetic Algorithm, can be got quickly, and its results can be convenientlydescribed by intelligentizing method.