基于BP神经网络模型的市政道路路基边坡稳定性研究

Research on the stability of municipal road subgrade slope based on BP neural network model

为定量分析路基边坡稳定性的主要影响因素的相互关系,并得出不同的影响因素对路基边坡稳定性的影响水平,文章采用SPSS对坡高、坡角、边坡土体的容重、粘聚力、内摩擦角和孔隙水压力进行相关性分析,采用BP神经网络模型计算出各个因素对路基边坡稳定性系数的定量影响水平.研究发现:路基边坡粘聚力与安全系数的相关性最高,系数达到0-352;容重和粘聚力对路基边坡稳定性系数的影响水平最高,分别为0.0239和0.031;采用BP神经网络算法计算出各个影响因素对路基边坡安全系数的影响水平,计算结果与原始设计的安全系数进行比较,发现两者结果具有较高的一致性,误差控制在5%以内.

In order to quantitatively analyze the interrelationships of the main influencing factors of slope stability,the influence of different influencing factors on the stability of subgrade slope is obtained.This paper uses SPSS to analyze the correlation of slope height,slope angle,slope bulk density,cohesive force,internal friction angle and pore water pressure.The BP neural network model is used to calculate the factors that affect the stability of subgrade slope.Quantitative impact levels.The study found that:the correlation between the cohesion and safety factor of roadbed slope is the highest,the coefficient reaches 0.352;the weight density and cohesion force have the highest impact on the stability coefficient of roadbed slope,which are 0.0239 and 0.031,respectively;they are calculated by BP neural network algorithm.The impact of each influencing factor on the safety coefficient of the roadbed slope was compared.The calculated results were compared with the safety factor of the original design.It was found that the two results had high consistency and the error was controlled within 5%.