查询结果:   任志玲,孙雪飞.电厂废水中和过程的CPSO-RBF神经网络预测控制[J].计算机应用与软件,2018,35(10):74 - 79.
中文标题
电厂废水中和过程的CPSO-RBF神经网络预测控制
发表栏目
应用技术与研究
摘要点击数
672
英文标题
PREDICTIVE CONTROL IN THE PROCESS OF POWER PLANT WASTEWATER NEUTRALIZATION BASED ON CPSO-RBF NEURAL NETWORK
作 者
任志玲 孙雪飞 Ren Zhiling Sun Xuefei
作者单位
辽宁工程技术大学电气与控制工程学院 辽宁 葫芦岛 125105     
英文单位
Faculty of Electrical and Control Engineering, Liaoning Technical University,Huludao 125105, Liaoning, China     
关键词
废水中和 RBF神经网络 强酸当量 预测控制
Keywords
Neutralization of wastewater RBFNN Strong acid equivalent (SAE) Predictive control
基金项目
国家自然科学基金项目(51277090,51477071);辽宁省高等学校优秀人才支持计划项目(LR2013013)
作者资料
任志玲,副教授,主研领域:智能化控制。孙雪飞,硕士生。 。
文章摘要
针对电厂废水中和过程的非线性、时变性和滞后特性,为提高控制的响应速度和稳定性,提出一种基于混沌粒子群(CPSO)优化的RBF神经网络预测控制算法。以强酸当量(SAE)模型作为控制对象,设计RBF神经网络预测模型。引入灵敏度法(SA)修正网络隐层神经元,CPSO算法快速准确搜索粒子信息,实现RBF神经网络辨识模型的最优化。通过在电厂用水加药系统循环控制上的仿真测试,表明该控制策略相比PID控制和基于遗传算法优化的BP神经网络控制,在控制的平稳性和快速性上具有一定优势。
Abstract
Aiming at the nonlinear, time-varying and hysteresis characteristics in the neutralization process of power plant wastewater, we proposed a RBF neural network predictive control algorithm based on chaotic particle swarm optimization (CPSO) to improve the control response speed and stability. Taking the strong acid equivalent (SAE) model as the control object, we designed the prediction model of RBF neural network. The sensitivity approach (SA) was introduced to modify the hidden layer neurons of the network and CPSO algorithm was adopted to fast and accurately search the particle information, thus realizing the optimization of the identification model of the RBF neural network. Through the simulation test on the circulation control of water charging system in the power plant, it shows that compared with PID control and BP neural network control based on genetic algorithm, the control strategy has a certain advantage in the stability and rapidity of control.
下载PDF全文