为解决岩土工程中边坡稳定性预测采用BP神经网络求取概率积分法预计参数出现的局部最优解、收敛速度慢等问题,文中结合麻雀搜索算法(SSA 算法)建立 SSA-BP 神经网络模型对边坡进行预测.选取容重、边坡坡脚、边坡高度、孔隙压力比、黏聚力,以及内摩擦角等 6 个影响因子作为网络模型结构的输入变量;利用麻雀搜索算法 SSA 对 BP 神经网络优化,得到最优的权重值和偏置项;最后对比分析 SSA-BP 神经网络和 BP 神经网络对边坡稳定性的预测效果.结果表明,SSA-BP算法的预测误差要小于BP算法,优化后的网络模型预测值更接近于实际值,麻雀搜索算法 SSA对BP神经网络的优化是有效的.
Slope Stability Prediction Based on SSA Optimization BP Neural Network
In order to solve the problem that BP neural network is difficult to predict slope stability in geotechnical engineering due to the local optimal solution and slow convergence of the predicted pa-rameters of probability integral method,the SSA-BP neural network model was established based on sparrow search algorithm(SSA algorithm)for slope prediction.Firstly,six factors such as bulk den-sity,edge slope foot,slope height,pore pressure ratio,cohesion force and internal friction angle were selected as input variables of the network structure.Secondly,the sparrow search algorithm SSA was used to optimize the BP neural network,and the optimal weight value and bias term were obtained.Finally,the prediction effect of SSA-BP neural network and BP neural network on slope stability was compared and analyzed.The results show that the prediction error of SSA-BP is smaller than that of BP,the predicted value of the optimized network model is closer to the actual value,and the sparrow search algorithm SSA is effective for the optimization of BP neural network.
NETL
NSTL
万方数据
