新型神经网络在钢球磨煤机制粉系统控制中的应用研究  

STUDY ON APPLICATION OF NEW TYPE NEURAL NETWORK IN CONTROL OF COAL PULVERIZING SYSTEM WITH BALL MILLS

作　　者：刘久斌[1] 承红[1] 

机构地区：[1]南京工程学院,江苏南京211167

出　　处：《热力发电》2006年第12期41-44,共4页Thermal Power Generation

摘　　要：采用新型对角回归神经网络来辩识系统模型,可对PID控制器参数进行整定,实现多变量解耦控制。应用该方法对火电厂钢球磨煤机出口温度和入口负压控制系统进行设计和仿真研究,仿真结果表明,控制系统无须专门解耦就能达到解耦目的,其控制稳定性高,鲁棒性强,调节响应速度快,动态偏差小,无静态偏差。Adopting new type diagonal regression neural network to identify the system model, the parameters of PID controller have been set, and the multi -variable decoupling control being realized. The outlet temperature and inlet negative pressure control system of ball mills in thermal power plant has been designed by using said method , and simulation study being carried out. Results of simulation show that the control system can be decoupled automatically, having high stability, strong robustness, rapid response speed to regulation, small dynamic deviation, and without static deviation.

关 键 词：火电厂 钢球磨煤机 温度 负压 神经网络 解耦 控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]