基于Matlab的边坡位移及稳定性预测的神经网络系统  

A Neural Network System for Slope Displacement and Stability Based on Matlab

作　　者：张洪林[1] 刘保松[1] 宋克志[2] 王本福[1] 

机构地区：[1]济宁市泗河管理处,山东济宁272100 [2]鲁东大学土木工程学院,山东烟台264025

出　　处：《西部探矿工程》2007年第8期31-34,共4页West-China Exploration Engineering

摘　　要：边坡位移的产生及发展对边坡的稳定性是极为重要的,然而导致边坡产生位移的因素是极其复杂的,采用具有非线性映射功能的神经网络理论,在Matlab环境下编程,建立了位移及稳定性预测的BP神经网络模型,运用历史位移数据训练神经网络并进行测试,将训练好的网络模型用于预测边坡位移的发展,进而预测边坡的稳定性。最后将该预测系统用于山东省莱州市仓上金矿北帮边坡工程,预测结果与已有的监测数据相比误差很小,说明效果良好。该系统对了解边坡位移的发展趋势以及研究边坡的动态稳定性具有一定的意义。As we know,displacement and it's grow are critical to stability of a slope.But the factors leading to slope displacement are complicate.However,neural network possesses strong capacity of nonlinear reflection.Based on lots of past data about displacement,we can adopt BP neural network(NNT) to simulate the relationship between slope displacements and time.The paper develops the system of forecasting slope displacement and stability,then the system is programmed by Matlab language.Finally the NNT system is used in the northern slope of Cangshang gold mine,Laizhou of Shandong province.As comparison shown,the difference between calculation value and locale observation is very minute.This study can be very significant to predict slope displacement trend and dynamic stability.

关 键 词：边坡 位移 稳定性 神经网络 MATLAB 

分 类 号：TU433[建筑科学—岩土工程]