基于模糊神经网络的工业废水处理预测研究  

Studies on Predicting the Effluent Treatment Process with Fuzzy Neural Network

作　　者：万金泉[1] 黄明智[1] 马邕文[1] 

机构地区：[1]华南理工大学环境科学与工程学院,广东广州510640

出　　处：《中国造纸学报》2008年第2期96-99,共4页Transactions of China Pulp and Paper

基　　金：广东省科技厅重大专项攻关项目(项目编号2003A30406);广州市科技计划项目资助(项目号2004Z3-D0271)

摘　　要：针对工业废水处理系统的时变性、非线性、复杂性和不确定性,利用废水处理监控系统取得表征废水水质的各项指标,构建基于BP算法的四层模糊神经网络模型。该网络模型仿真实际废水处理过程的结果表明,模糊神经网络具有较强的学习能力;其较BP网络对样本数据的仿真误差较小,平均相对误差仅为1.5%,为实现废水处理的自动控制提供可行途径。Concerning the characteristics of time-varying, nonlinearity, complexity and uncertainty of industrial wastewater treatment system, based on the effluent quality data from the monitoring system of wastewater treatment system , one four-layer fuzzy neural network was built by using the Back Propagation Algorithm and fuzzy logic to study the nonlinear relationships in the wastewater processing, and to anticipate the effluent treatment process. The results of simulating practical wastewater treatment process of the network model showed that the FNN network has a strong learning capability. Compared with BP network, the outcome of this method is better,the effluent treatment process is predicted with this FNN network model with the average relative error of 1.5%, which is a feasible approach to carry out automatic control in the wastewater treatment.

关 键 词：模糊神经网络 废水处理 预测模型 

分 类 号：X793[环境科学与工程—环境工程] TP29[自动化与计算机技术—检测技术与自动化装置]