基于BP神经网络算法下的边坡安全预测

Safety prediction of slope based on BP Neural Network Algorithm

边坡的实时变形一直是岩土工程界关心的问题,由于不同工程的条件不同,影响边坡位移的因素较多,进而使其变化趋势复杂.为了得到边坡位移与稳定性的关系,采用BP神经网络算法与强度折减法综合对土质边坡安全系数进行预测.结果显示:通过强度折减法计算出边坡位移,并获取较完善的BP神经网络样本数据,当迭代次数达到足够时,完全可以忽略预测结果与实际结果的误差;通过实际工程中的边坡监测数据,然后由建立的BP神经网络能够较为准确的输出边坡的强度折减系数,进而得到相应的安全系数.

Soil slopes under different stability,the displacement will be different,while the impact of slope displacement factors are more,and then make its changing trend is complex. In order to obtain the relationship between slope displacement and stability,BP neural network algorithm and strength reduction method are used to predict the safety factor of soil slope.The results show that the slope displacement is calculated by the strength reduction method and the BP neural network sample data is obtained. When the number of iterations is enough,the error between the prediction result and the actual result can be neglected. Through the slope monitoring in the actual project data,and then by the establishment of the BP neural network can be more accurate output slope strength reduction factor,and then get the corresponding safety factor.