摘要
针对水处理过程中混凝剂的准确投加,以及投药过程中的时滞、网络延迟等问题,采用基于网络学习控制的智能控制算法来改进投药控制系统。用远程专家系统和自学习BP神经网络复合算法的优点,即专家系统的前馈补偿能力解决流量、浊度突变、延迟等干扰因素;神经网络的非线性映射能力解决水处理非线性影响;滚动学习和反馈学习来解决时变、时滞等问题。该算法较好地解决了网络延迟造成的系统性能下降问题,加快了神经网络的训练速度。将该算法应用于水厂自动化系统,可以实现最佳投药量控制,水质符合标准,并取得良好的经济效益。
To the problem of adding coagulation accurately and the time-delay of the process of coagulant dosage,the intelligent control algorithm based on networked learning control is used to improve the dosing control system.Remote expert system（ES）and BP-NN is combined together.The feedforward compensation ability of ES is used to resolve the interfering factors of flow,turbidity mutation and time-delay.The neural network nonlinear mapping ability is used to solve the nonlinear effects of water treatment.Roll learning and feedback learning are used to solve the time-varying and time-delay problems.The proposed algorithm could solve the problem of network delay and speed training of BP-NN.It can achieve the optimal dosage,and the quality of the water accord with the standard.The result shows it can get well economic benefits.
出处
《控制工程》
CSCD
北大核心
2010年第6期751-754,共4页
Control Engineering of China
基金
湖南省科学技术与科技计划(2006GK3130)
湖南省自然科学基金奖励项目(05JJ30121)
关键词
智能控制器
神经网络
网络学习控制
加药控制器
intelligent controller
BP neural network
networked learning control
dosing controller
