汽温联合控制系统改进GRU控制算法研究  

Research on Improved GRU Control Algorithm for Steam Temperature Joint Control System

作　　者：齐传杰 何同祥[1] 陈博文 黄嵘 Qi Chuanjie;He Tongxiang;Chen Bowen;Huang Rong(School of Control and Computer Engineering,North China Electric Power University,Hebei,Baoding,071003,China)

机构地区：[1]华北电力大学自动化系,河北保定071003

出　　处：《仪器仪表用户》2023年第4期85-89,共5页Instrumentation

摘　　要：针对超临界机组直流锅炉汽水系统所具有的大惯性、大迟延等特点,研究如何联合控制主汽温和中间点温度。首先,基于某600MW超临界机组运行数据,以中间点温度、主汽温度系统作为被控对象进行建模;然后,设计出一款改进的门控循环单元神经网络(GRU)的控制器,神经网络各项权值通过麻雀搜索算法优化;最后,将该控制器应用于中间点温度主汽温联合控制。结果表明,相比于传统门控循环单元神经网络控制的单独主汽温控制,本文提出的控制器取得了更好的控制效果。Aiming at the characteristics of large inertia and large delay in the steam water system of supercritical unit DC boiler,how to jointly control the temperature of main steam temperature and intermediate point is studied.First of all,based on the operation data of a 600MW supercritical unit,the intermediate point temperature and the main steam temperature system were taken as the controlled objects for modeling.Then,an improved controller of gated cyclic unit neural network(GRUs)is designed.The weights of the neural network are optimized by sparrow search algorithm.Finally,the controller is applied to the joint control of the intermediate point temperature and the main steam temperature.The results show that the controller proposed in this paper achieves better control effect than the single main steam temperature control controlled by the traditional gated cycle unit neural network.

关 键 词：主汽温控制 门控循环单元神经网络(GRU) 中间点温度 麻雀搜索算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]