基于因子分析和BP神经网络的滑坡抗剪强度参数取值

滑坡抗剪强度参数取值是一直困扰滑坡防治工程设计的一大难题,目前其研究方法主要有试验法、反分析和统计分析法,但都没有考虑其他基本物理力学参数的影响。为此,提出了以某一区域工程地质条件相似的滑坡基本物理力学参数为基础,采用SPSS数学分析软件分析滑坡各基本物理力学参数与抗剪强度的相关性,筛选出对滑坡抗剪强度影响较大的因子;然后采用BP神经网络模型研究区域滑坡抗剪强度与影响因子的网络结构,并以建立的神经网络结构对该区域的滑坡抗剪强度参数进行估算;最后以万州区滑坡为例进行分析和验证。研究结果表明,采用神经网络计算的结果与滑坡试验得到的结果误差基本都在5%左右,精度较高。

Parameter Value Selection on Anti-shearing Strength Parameters of Landslides Based on Factor Analysis and BP Neural Network

Parameter value selection on anti-shearing strength parameters of landslides,which had been being a difficulty in landslide prevention and control engineering design,is presently studied mainly by test,back analysis and statistical analysis without considering the influence of other basic physical and mechanical parameters.By using a method based on basic physical and mechanical parameters of landslides,which have similar engineering geological conditions in an area,the correlation between basic landslide physical and mechanical parameters and anti-shearing strength are analyzed by adopting mathematical analysis software SPSS,factors with greater influence are screened.Then,BP neural network model is used to establish a neural network structure between the anti-shearing strength and influence factors in the research area,which is applied to estimate the landslide anti-shearing strength parameters of this region.Finally,the accuracy is analyzed and verified by taking Wanzhou landslides as an example.The results show that the error between the experiment results and the calculation result using the neural network is less than 5%,and its precision is higher.