基于神经网络预测控制的主汽温优化控制  

Optimization Control of Main Steam Temperature Based on Neural Network Predictive Control

作　　者：苏烨 凌路加 段亚灿 董泽[2] SU Ye;LING Lu-jia;DUAN Ya-can;DONG Ze(State Grid Zhejiang Electric Power Co.,Ltd.Research Institute,Hangzhou Zhejiang 310014,China;Hebei Technology Innovation Center of Simulation&Optimized Control for Power Generation,North China Electric Power University,Baoding Hebei 071003,China)

机构地区：[1]国网浙江省电力有限公司电力科学研究院,浙江杭州310014 [2]华北电力大学河北省发电过程仿真与优化控制技术创新中心,河北保定071003

出　　处：《计算机仿真》2021年第9期114-118,共5页Computer Simulation

摘　　要：火电厂锅炉主汽温不易直接测量,容易受到外界干扰,主汽温控制过程容易受到电力生产工艺因素的制约,所以采用常规的控制技术难以取得良好的优化控制效果。根据我国火电厂的实际情况,同时结合智能控制的发展现状,将火电厂锅炉主蒸汽温度作为研究对象,提出采用基于神经网络的预测控制策略对锅炉主汽温进行控制。神经网络具有的自适应学习能力能够很好地适应控制环境的变化,可以通过简单的工具自动进行特征提取,产生有用的数据。采用LSTM循环神经网络建立主汽温的预测模型以预测主蒸汽温度的未来输出值,同时进行反馈校正,克服外界因素对系统扰动所造成的预测误差,得到较为精确的主汽温预测值,最后根据二次性能指标建立优化器实现对控制量的滚动优化。The main steam temperature of the boiler in a thermal power plant is not easy to be measured directly, and it is easy to be interfered by the outside world.The control process of the main steam temperature is easily restricted by the factors of electric power production technology, so it is difficult to achieve a good optimal control effect by using conventional control technology.According to the actual situation of a thermal power plant in China and the development status of intelligent control, the main steam temperature of the boiler in a thermal power plant was taken as the research object, and the predictive control strategy based on neural network was proposed to control the main steam temperature of the boiler.The adaptive learning ability of neural network could adapt to the change of control environment well.It could automatically extract features through simple tools to produce useful data.This article used the LSTM cycle neural network to establish the forecast model of main steam temperature in order to predict the future output value of main steam temperature, the feedback correction at the same time, to overcome the disturbance caused by the external factors on the system prediction error, and obtain more accurate main steam temperature forecast, according to the quadratic performance index to establish the optimizer to realize to control the amount of rolling optimization.

关 键 词：锅炉主蒸汽 预测控制 一维黄金分割 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]