基于动态云QNN的污水出水水质在线预测方法  

Online Prediction Method for Sewage Discharge Quality Based on Dynamic Cloud—Quantum Neural Network

作　　者：马亮[1] 杨萍萍[1] 谷学静[2] 邢玉秀[1] 

机构地区：[1]河北联合大学轻工学院,唐山063000 [2]河北联合大学电气工程学院,唐山063000

出　　处：《计算机测量与控制》2014年第3期700-702,712,共4页Computer Measurement &Control

基　　金：河北省科技支撑计划项目(13210330);河北省自然科学基金钢铁联合基金资助项目(F2012209015);河北联合大学轻工学院科学研究基金资助项目

摘　　要：针对污水处理过程所具有的多变量、非线性和大时滞的特点,将云模型与QNN(量子神经网络)以串联方式有机结合,首先利用云变换方法进行网络的结构辨识和云模型的特征提取,同时通过在输入层引入单位延时环节描述污水出水的动态过程,研究提出了基于动态云QNN的污水出水水质在线预测方法;结合在线测得的污水水质数据,通过与规则多层前向神经网络对比分析的结果表明,该方法能准确的预测污水出水水质BOD5,均方误差性能函数(MSE)值为1.0×10^(-3),单步运行时长为1.122×10^(-4),完全能够满足实时性要求。In this paper, an online prediction method for sewage discharge quality based on the dynamic cloud--Quantum Neural Net- work was presented, the cloud transform method was used to identify the network configuration and to extract the cloud features, and a unit time--delay was also introduced into the input layer to describe the dynamic characteristics of the sewage discharge, which aimed at resolving the characteristics of ultivariable, nonlinear, big--lagged in the treatment process. Results with on--line measuring sewage discharge data show that the method can predict sewage discharge quality of BOD5 accuratly, comparing with regular multi--layer forward neural network, and the method can run in real--time with the single processing time being 1. 122× 10-4 and the MSE being 1.0 × 10-3.

关 键 词：污水出水水质 在线预测 云模型 QNN 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]