考虑时延因素的神经网络主汽温预测  

Neural Network Main Steam Temperature Prediction Considering Time Delay Factors

作　　者：周飞燃 ZHOU Fei-ran(North China Electric Power University,Beijing 102206,China)

机构地区：[1]华北电力大学,北京102206

出　　处：《价值工程》2024年第17期101-103,共3页Value Engineering

摘　　要：火电机组在灵活调峰调频时,其负荷的大范围变动导致锅炉的部分运行参数也随之频繁变动,进而影响主汽温的稳定性。为了便于对主汽温进行超前控制,采用LSTM神经网络预测模型,对主汽温的未来变化进行预测,并且针对主汽温的主要影响因子存在的迟延问题,提出了互信息法,解决参数时序对齐问题;采用发电机组实际运行数据进行仿真,结果表明此神经网络模型能够有效对主汽温进行预测。During flexible peak shaving and frequency regulation of thermal power units,the wide range of load changes leads to frequent changes in some operating parameters of the boiler,which in turn affects the stability of the main steam temperature.In order to facilitate advanced control of the main steam temperature,an LSTM neural network prediction model is adopted to predict the future changes of the main steam temperature.In response to the delay problem of the main influencing factors of the main steam temperature,the mutual information method is proposed to solve the problem of parameter timing alignment;The simulation was conducted using actual operating data of the generator set,and the results showed that this neural network model can effectively predict the main steam temperature.

关 键 词：主蒸汽温度 神经网络 迟延特性 

分 类 号：TM621.2[电气工程—电力系统及自动化]