基于小波神经网络的循环冷却水腐蚀预测研究

Study of Corrosion Prediction of Circulating Cooling Water Based on Wavelet Neural Network

利用某石化企业40个月的循环冷却水实际生产数据,基于小神经网络进行了腐蚀预测研究。经过对比分析,得出了小波神经网络预测精度最高的网络模型为6-7-1结构。在相同输入参数向量下,对含有相同隐层节点个数的小波神经网络和BP神经网络进行腐蚀预测对比,小波神经网络比BP神经网络预测精度高。

Basing on circulating cooling water's production data of forty months in a petrochemical plant and the wavelet neural network,the corrosion prediction was investigated. The comparative analysis shows that a network model with 6-7-1 structure has the highest prediction accuracy; and under the same input parameters,having the wavelet neural network compared with BP neural network,both has same hidden layer nodes,shows that the wavelet neural network outperforms the BP neural network in corrosion prediction accuracy.