基于正交神经网络的焦炉立火道温度预测控制  

Prediction and Control of Coke Oven Vertical Flue Temperature Based on Orthogonal Neural Network

作　　者：邓慧君[1] Deng Huijun(Anhui Maanshan Industry School, Maanshan, Anhui 243000 Chin)

机构地区：[1]安徽省马鞍山工业学校,安徽马鞍山243000

出　　处：《冶金动力》2018年第6期69-73,共5页Metallurgical Power

摘　　要：作为冶金工业中的重要设备,焦炉在我国的钢铁行业中得到了大量的使用。在焦炉的运行中,控制变量数量多,变化呈非线性状态。作为一个大惯性时变系统,焦炉的控制方式十分复杂。而作为对焦炉生产起主要作用的因素,焦炉立火道温度能否得到有效控制对于焦炉能否达到最佳燃烧状态十分重要。文章选取正交多项式为理论基础,在此基础上建立了正交神经网络的焦炉立火道温度控制模型,以阶梯式广义预测控制为控制策略,展开相关研究工作,以期提高控制系统的响应速度和控制精度。As important equipment in metallurgical industry, coke oven has been widely used in domestic steel industry. There are a lot of control variables with nonlinearity changing during coke oven operation. As a large inertia and time-varying system, the control model of coking oven is very complicated. Whether the coke oven vertical flue temperature can be effectively controlled is extremely important for the oven to reach optimum combustion state. The paper takes orthogonal polynomials as the theoretical basis, on which an orthogonal neural network flue temperature control model is established; and takes step generalized prediction control as the control strategy to carry out related research work, in order to increase the responding speed and control accuracy of the control system.

关 键 词：焦炉 立火道温度 正交神经网络 

分 类 号：TP316[自动化与计算机技术—计算机软件与理论]