Study On Parameter Of Tunnel Surrounding Rocks By Back-analysis Method Of Artificial Intelligence And Its Engineering Application

During the reliability analysis of surrounding rock and detail design, using appropriate model and parameters is very important to the result of calculation for it's special material. Back-analysis is a good method to get constitution model and parameter of rock. So it is of great significance to study the theories and methods of back-analysis.In practicality project, the elastic-plastic characters of rock already satisfied the needs. So in this paper, the elastic-plastic model of rock has been analysis. The main content is as follows:Using important parameter to calculate the sensibility of surrounding rock displacement, it is negligible to rock displacement impact on possion's ratio. Expatiating the concept of system identification of rock. Introducing the method of optimizing back analysis using displacement and summarizing the elastoplastic model. Changing elastoplastic model identification into stress subdued function and elastomeric parameter identification.Empoldering the improved annealing algorithm combining the GA for optimizing, and tested the validity of this method by simple functionFirst, optimizing neural networks structure. Second, testing neural networks. Thirdly, applying improved annealing algorithm combining BP neural networks to back-analysis of the elastoplastic model and parameters of rock.Orthogonal experimental design and uniform design method are adopted to constitute study examples and test examples based on practical tunnel engineering. Training and optimizing the structure to prove the forecast capability of neural networks is good enough, then identify the model and estimate the parameters of wall rock. Validating the results identified by back test. Getting the displacement results exactly in order to validate the rationality of the method in this paper.