基于BP神经网络和均匀设计的边坡敏感性分析  

Sensitivity Analysis of Slope Stability Based on BP Neural Network and Uniform Design

作　　者：刘飞[1] 胡斌[1] 宋丹[2] 饶晨曦 刘智权 

机构地区：[1]中国地质大学(武汉)工程学院,湖北武汉430074 [2]西安科技大学地质与环境学院,陕西西安710054 [3]四川省冶金设计院,四川成都610041

出　　处：《水电能源科学》2014年第10期113-115,165,共4页Water Resources and Power

基　　金：国家自然科学基金项目(41172281);国家重点基础研究发展计划(973计划)项目(2011CB710604);中央高校基本科研业务费专项资金项目-特色学科团队(CUG090104)

摘　　要：为在边坡稳定性评价中确定影响因素的敏感性,采用BP神经网络和均匀试验设计相结合的方法对边坡进行敏感性分析。基于BP神经网络的非线性映射关系,利用均匀设计理论,使试验点分布更均匀、代表性更强,在考虑多因素交互作用的同时,大幅减少了试验次数。以某一岩体边坡为例,应用均匀设计理论建立边坡敏感性分析模型,并应用灰色关联分析方法评价其敏感性。结果表明,各因素对该边坡稳定性的关联度由大到小依次为重度、内摩擦角、粘聚力、边坡高度、孔隙水压力、边坡角。In order to determine the sensitivity of influencing factors in slope stability evaluation, combination of BP neural network and uniform experiment design is used to carry out sensitivity analysis of the slope. Based on the BP neural network nonlinear mapping relationship, it can make sites distribution more uniform and more representative by using the theory of uniform design. At the same time, the number of test was significantly reduced by considering the interaction of factors. Taking a rock mass slope as an example, the sensitivity analysis of slope model was set up with the theory of uniform design, and grey correlation analysis method was applied to evaluate slope sensitivity. The results show that various factors on the stability of the slope correlation degree ranks as follows., severe, internal friction angle, cohesive force, slope height, pore water pressure, slope angle.

关 键 词：边坡稳定性 敏感性分析 均匀试验 BP神经网络 灰色关联分析 

分 类 号：TU457[建筑科学—岩土工程]