Font Size: a A A

Displacement Back Analysis For Rock Slope Based On BP Neural Network And Its Application Research

Posted on:2011-06-17Degree:MasterType:Thesis
Country:ChinaCandidate:F W YuFull Text:PDF
GTID:2132360305960785Subject:Geotechnical engineering
Abstract/Summary:PDF Full Text Request
Mechanical parameters of rock and soil play an important role to the design, construction and analysis in geotechnical engineering, affecting the safety, economy and performance of a project. Therefore, "inaccuracy of the parameters" has become a bottleneck problem to both theoretical analysis and numerical simulation.Displacement back analysis can solve this problem. However, the existing displacement back analysis method has various disadvantages when fixing parameters, such as its diffcult to optimize, inefficiency, complex and hindering the extension of the technology. Therefore, studying and establishing a displacement back analysis method, which can be well optimized, highly efficiency and easy to use and spread, is the problem demanding prompt solution.Based on uniform design, finite difference program and BP neural network, a new intelligent displacement back analysis method is established. First, according to the scope of geotechnical parameters obtained by survey and based on uniform design requirements, mechanical parameters are designed to different levels to fix the scheme of numerical simulation. So can solve the question of optimization. Then, the numerical analysis model for test pile is established. Following the scheme, corresponding neural network training samples is obtained by the finite difference program. Third, based on MATLAB,the BP neural network back analysis model is established and trained. With back analysis of the displacement of a test pile, mechanical parameters can be obtained. Finally, with the application of the parameters in a slope numerical model including test pile and anti-slide pile, the displacement and moment can be compared with the measured ones. A conclusion can be drew that the parameters are reliable. Thus, the reliability of the numerical method is also proved. An intelligent displacement back analysis method which is easy to optimize, highly efficiency, convenient for use and extension is obtained.
Keywords/Search Tags:rock slope, numerical simulation, displacement back analysis, Uniform Design, BP neural network
PDF Full Text Request
Related items