基于CM-AFSA-BP神经网络的土石坝渗流压力预测

针对经典BP神经网络训练效率低、易陷入局部极值等缺点,利用云模型对传统人工鱼群算法(AFSA)进行改进,并采用改进后的云人工鱼群算法(CM-AFSA)对BP神经网络的权值和阈值进行优化,构建基于CM-AFSA-BP神经网络的预测模型。以某土石坝测压管水位为指标,利用CM-AFSA-BP神经网络预测模型对其渗流压力进行预测,并与同结构的经典BP神经网络预测结果进行对比分析。结果表明,CM-AFSA-BP神经网络模型在训练速度和预测精度上明显更优,在土石坝渗流压力预测和分析方面具有较好的适应性。

Seepage Pressure Prediction of Earth rock Dam Based on CM AFSA BP Neural Network

In view of the shortcomings of the classical BP neural network, such as low training efficiency and easy to fall into the local extremum, the cloud model was used to improve the traditional artificial fish swarm algorithm. And then the improved algorithm was used to optimize the weights and thresholds of the BP neural network. The prediction model based the CM AFSA BP neural network was constructed. The CM AFSA BP neural network prediction model was used to predict the seepage pressure with the piezometric level as an indicator, and the prediction results were compared with the results from classical BP neural network. It was found that the CM AFSA BP neural network model is obviously better in the training speed and prediction accuracy, and has good adaptability in the prediction and analysis of seepage pressure of earth rock dam.