|
In recent years, for the underground engineering is widely applied in differrent fields, the stability problems of surrounding rock-mass are much concerned by researchers. Especially, for those deep underground cavern projects located in complex geological and high stress conditions, people often encounter some problems that couldn’t be resolved empirically. At this time, with the development of numerical calculation method, numerical analysis methods will be the best choise. Feedback analysis has been widely applied in underground engineering, because of the existing of many uncertainties. For the request of the dynamic characteristics and aging properties, then construction-period feedback analysis method must be systematic and efficient.The purpose of the paper is to build a feedback analysis system for underground group caverns in construction-period. This system, which is built in the mixed algorithm of the lagrange modeling method and intelligent artificial neural network back analysis method, can simulate the process of construction, offer the intelligent feedback optimal calculation of rock mechanics parameters and forecast rock stability.Lagrange method is the core algorithm of the feedback analysis system, which can simulate the underground rockmass, cavity structure, initial geostress field, construction excavation and supporting measures. The back analysis and calculation of the system is mainly based on the displacement back analysis theory of considering the loose influence circle. The back analysis, to get underground rock mechanics parameters, is using the method of orthogonal test design and BP artificial neural network model.The main calculation work of feedback analysis system is controling by the core calculation documents that are writed by command language of FLAC3D software. What’s more, the system realizes the interface and visual operation function by the Visual Basic programming languages, people can exchange data with computer and control process of the feedback analysis. The system is composed of five parts: the user’s windows module, the analysis object management and data interface calculation module, the nonlinear numerical analysis, the sample formation and BP inversion artificial neural network analysis module, post-processing module and so on.In the last part of this paper described that, the system is used for the stability analysis of Jin-Ping I hydropower station underground powerhouse caverns , in order to verify the feedback analysis system is of applicability, reliability and accuracy, when it’s applied in the actual engineering project. Firstly, select the most representative of monitoring section as the object of calculation, according to analysis and evaluation for the actual monitoring data. Then, accurately simulate initial geostress field and construction process, following by the back analysis to get rock mechanics parameter. At last, analyze the results and evaluate the engineering stability of the rock.
|