基于模糊神经网络的混凝沉淀投药控制  

Dosing Control of Coagulating Sedimentation Process Based on Fuzzy-Neural Networks

作　　者：史步海[1] 朱学峰[1] 

机构地区：[1]华南理工大学自动化科学与工程学院,广东广州510640

出　　处：《控制工程》2009年第6期723-726,共4页Control Engineering of China

基　　金：国家自然科学基金资助项目(6207433);广东省科技攻关基金资助项目(2005B10201005)

摘　　要：根据已有的资料分析了自来水混凝沉淀过程多变量影响因素和特点,找出了影响混凝沉淀过程的主要变量是进水流量、进水浊度、原水pH值、进水温度。根据高木-关野模糊系统和神经网络非线性逼近的特点,提出了建立混凝沉淀加药控制模型的模糊适用度加神经网络的集成模糊神经网络控制方法。还结合某自来水厂的具体现场数据进行了仿真模型研究和现场的实验,实验表明控制精度满足要求,运行正常。Based on the existing materials, the characteristics and affectiveness of multi-variables of coagulating sedimentation process of the waterworks arc analyzed. The main infection variables are found out, which are the influent flow rate, the influent flow turbidity, the influent flow temperature and the pH value of the raw water. Base on the T-S fuzzy model and neural networks, a model of coagulating sedimentation dosing control is presented. An intergraded fuzzy neural network control method consisting of the fuzzy function and neural networks is put forward. The simulation and experiment results the real data of some waterworks show that the demand of control precision is satisfied and the fuzzy neural network system works on the rails.

关 键 词：模糊控制 模糊神经网络 混凝沉淀 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]