基于RBFNN和HDP方法的蔗糖生产澄清工段中和pH值优化控制器的研究  

Research on PH Value Optimization Controller in Cane Sugar Production and Clarification Based on RBFNN and HDP

作　　者：史长城[1] 吴晓莉[1] 

机构地区：[1]郧阳师范高等专科学校物理与电子工程系,湖北十堰442000

出　　处：《郧阳师范高等专科学校学报》2011年第3期76-81,共6页Journal of Yunyang Teachers College

基　　金：湖北省教育厅科研项目(项目编号:Q20105001)

摘　　要：澄清工段是亚硫酸法生产蔗糖的关键过程之一,而pH值又是该过程最关键的工艺指标,能否将其pH值稳定地控制在要求范围将影响整个糖厂产量和白糖质量.由于该过程是一个复杂的物理、化学过程,具有多输入、大滞后、强非线性、时变参数等特点,因此该过程pH值的稳定控制和优化问题一直没有得到很好的解决.采用径向基函数神经网络(Radial Basis Function Neural Network,RBFNN)来建立该过程的神经网络模型.然后利用RBFNN来设计基于启发式动态规划(Heuristic Dynamic Programming,HDP)方法的优化控制器.仿真结果表明该方法可以取得满意的控制效果,为解决复杂非线性动态系统的优化控制问题提供了一种新途径.Clarification process is the key part of producing cane sugar by way of sulfuric acid, while ph value is the most important technical index, which has great influence upon the production of the whole sugarhouse and the quality of the sugar. The process is a complex physical and chemical one, with more input, long delay, strongly non--linear, time --varying parameters, so the stabilizing control and optimization of ph value in the process has not been achieved. The neural network model of this process has been set up by Radial Basis Function Neural Network, and then optimization controller based on HDP has been designed by use of the former. The simulation has had a satisfied controlling result, which provides a new way to settle optimization control of complex non--linear dynamic system.

关 键 词：优化控制 启发式动态规划 自适应动态规划 径向基函数神经网络 PH值控制 

分 类 号：TM571.62[电气工程—电器]