模糊神经网络控制器在CFBB床层温度控制中的应用  

Application of Fuzzy Neural Networks Controller in Controlling Bed Temperature of Cycling Fluidized Bed Boiler

作　　者：高夫燕[1] 张丰田[2] 金超[3] 

机构地区：[1]浙江大学宁波理工学院,浙江宁波315100 [2]东北电力大学,吉林吉林132012 [3]浙江北仑第一发电有限责任公司,浙江宁波315800

出　　处：《锅炉技术》2006年第B04期42-45,76,共5页Boiler Technology

摘　　要：由于循环流化床锅炉(CFBB)的燃烧过程是一个多参数、非线性、时变以及变量强耦合的过程,很难建立被控对象的准确数学模型,根据CFBB床层温度被控对象的动态特性,设计了一个模糊神经网络自适应控制系统,引用模糊高斯基函数神经网络结构,并采用基于变尺度优化学习算法的改进型学习算法(MD-FP),其学习信号由神经网络辨识器(NNI)提供。理论分析和仿真结果表明该控制系统具有良好的控制性能,其稳定性、鲁棒性等控制品质明显优于常规模糊控制。The process of combustion of CFBB is a multi parameter, nonlinear, time varying and strongly coupling course, so the exact mathematical models of the controlled objects are hardly constructed. According to dynamic characteristic of the bed temperature of CFBB, a fuzzy neural-networks self-adaptive control system is designed, using fuzzy neural networks controller and adopting the learning arithmetic of modified diversified foot pace. The learning sign^il is offered by the neural network idenlification. Theoretic analysis and the results of simulation have indicated that this control system has satisfying controlling capability and is better than the normal fuzzy control system in stability, robust, etc.

关 键 词：CFBB 模糊神经网络 自适应 MDFP NNI 

分 类 号：TK223[动力工程及工程热物理—动力机械及工程]