摘要
由于BP神经网络模型容易出现局部极小值和训练时间长等缺陷,Elman神经网络模型有搜索速度慢、易出现局部最优等不足,因此考虑采用PSO-Elman神经网络模型进行尾矿坝位移预测。利用Matlab神经网络工具箱对PSO-Elman神经网络、Elman神经网络、BP神经网络三种模型进行编程,结合工程算例,验证了PSOElman神经网络模型在尾矿坝位移监测数据预测分析中具有更好的预测效果。
Considering that the BP neural network model is prone to be local minima and time consuming, and that Elman neural network model has the shortage of slow search speed and local optimum, therefore PSO-Elman neural network model was used to predict railings dam displacement. Three models including PSO-Elman neural network, Elman neural network and BP neural network were programmed by using Matlab neural network toolbox, then combining engineering examples, finally PSO-Elman neural network model was proved to have a better effect in predicting and analyzing the displacement monitoring data of railings dam.
作者
王新岩
WANG Xinyan(BGRIMM Technology Group,Beijing 100160,China;Beijing Key Laboratory of Nonferrous Intelligent Mining Technology,Beijing 102628,China)
出处
《有色金属（矿山部分）》
2018年第5期18-21,共4页
NONFERROUS METALS(Mining Section）
基金
"十三五"国家重点研发计划项目(2017YFC0804608)
