摘要
某600 MW超超临界机组参与电网一次调频和自动发电控制(AGC),机组负荷频繁变化、经常处于深度快速变负荷工况运行,且锅炉燃煤品种多变,导致机组过热汽温波动大、控制品质差。为此,将人工神经网络(ANN)与传统PID串级控制思想结合,提出一种外环采用神经网络逆控制、内环采用PID控制器的过热器喷水减温智能串级控制策略。首先利用该机组的历史运行数据对各级过热器的汽温特性进行逆建模,在此基础上基于罗克韦尔PLC控制器开发了汽温外挂优化控制系统,经调试后在现场稳定投运。应用结果表明,采用本文的智能控制策略,有效减小了一次调频和AGC深度变负荷过程中机组过热汽温的波动,大大提高了汽温的动态和稳态控制品质,减少了运行人员对汽温设定值的频繁干预,增强了机组对AGC深度调峰的适应能力。
A 600 MW ultra-supercritical(USC)power unit,which participates in power grid primary frequency regulation(PFR)and automatic generation control(AGC),frequently changes its load and often works in wide-range variable load conditions with diverse coal quality.It leads to large fluctuation of superheated steam temperature(SST)and poor control quality.By combining artificial neural network(ANN)with traditional PID cascade control idea,this paper proposes an intelligent cascade control strategy for the superheater water-spray desuperheating system.It adopts neural network inverse control in the outer loop and PID controller in the inner loop.First of all,we establish the inverse models of the steam temperature characteristics of the superheaters based on the historical operation data of the unit.Then,we develop a plug-in optimization program for SST control based on Rockwell PLC controller.After debugging,the new control system is put into on-site operation stably.It is shown that the advanced SST control scheme effectively reduces the fluctuation of SST in wide-range load-changing conditions under PFR and AGC.It also greatly improves the dynamic and steady-state SST control quality.Besides that,it markedly reduces the operating personnel’s frequent intervention to steam temperature setpoint and enhances the AGC deep peak-load adjustment adaptability of the power unit.
作者
马良玉
燕梦
王林
汪裕杰
彭文权
丘鸿
彭春雄
MA Liangyu;YAN Meng;WANG Lin;WANG Yujie;PENG Wenquan;QIU Hong;PENG Chunxiong(School of Control and Computer Engineering,North China Electric Power University,Baoding 071003,China;Xi’an Thermal Power Research Institute Co.,Ltd.,Xi’an 710054,China;Guangdong Yuedian Dapu Power Generation Co.,Ltd.,Dapu 514265,China)
出处
《华北电力大学学报（自然科学版）》
CAS
北大核心
2021年第5期106-112,120,共8页
Journal of North China Electric Power University：Natural Science Edition
基金
国家自然科学基金资助项目(61174111).
关键词
超超临界机组
过热汽温
神经网络建模
智能优化控制
工程应用
ultra-supercritical(USC)power unit
superheated steam temperature
neural network modeling
intelligent optimization control
engineering application
