用神经网络辨识化学强化一级处理系统  

Modelling Studies of Chemically Enhanced Primary Treatment System by PB Network

作　　者：辛刚[1] 蒋文举[1] 金燕[1] 谭钦文[1] 刘成军[1] 

机构地区：[1]四川大学环境工程系,成都610065

出　　处：《环境科学与技术》2002年第4期10-11,16,共3页Environmental Science & Technology

摘　　要：运用实验室得到的数据为样本 ,选取了CODCr和絮凝剂投加量为指标 ,采用三层结构的神经网络 ,利用Matlab的神经网络工具箱中的批处理梯度下降法对CEPT系统经行辨识。辨识结果表明 :模型的预测值与实测值的平均误差在 3 .7%左右 ,具有较高的精度。同时 ,该模型还具有适应性强 ,使用方便 ,高效的特点 ,为CEPT系统的在线实时控制提供了一条有效的途径。Identification of the chemically enhanced primary treatment (CEPT) system was carried out by means of the Batch Gradient Steepest Descent Algorithm in Neural Network Toolbox of Matlab, where data gained in laboratory were used as training samples, CODcr concentration and cost of coagulant as indexes. The modelling is considered comparatively accurate since there is an average error of ca. 3.7% between the predicted and experimental values. The model also shows the flexibility, convenience and high efficiency, providing a useful way for on line control for CEPT.

关 键 词：化学强化一级处理 水质模型 神经网络 废水处理 

分 类 号：X703.1[环境科学与工程—环境工程]