基于神经网络的智能吹灰优化系统设计与应用  

DESIGN AND APPLICATION OF INTELLIGENT SOOT BLOWING OPTIMIZATION SYSTEM BASED ON NEURAL NETWORK

作　　者：郑亚锋[1] 卢海彬[1] 谢红军[1] 

机构地区：[1]国核电力规划设计研究院,北京100095

出　　处：《热力发电》2012年第10期38-40,共3页Thermal Power Generation

摘　　要：以某330MW机组锅炉为对象,求得锅炉受热面吸热净收益最大化时的最佳吹灰频率及清洁系数,并以神经网络软件为基础,借助DCS中的实时数据监测锅炉对流受热面及热辐射受热面的积灰情况,实时比较监测清洁系数与临界清洁系数,以此进行吹灰操作。应用表明,系统可实时反映炉内灰污状态,指导受热面吹灰按需进行,减少了盲目吹灰造成的能量消耗。Aiming at improving the economical efficiency of boiler soot blowing system,this paper takes a 330 MW unit boiler as the object to calculate the optimum soot blowing frequency and clean coefficient when the heat-absorbent net income of the heating surface reaches the maximization.Moreover,based on the neural network software,real-time monitoring for ash depositing situation on convective heating surface and radiant heating surface was carried out through data collected from DCS,so as to compare and monitor the clean coefficient and the critical value timely,thus to take initiative to soot blowing.Application results indicate that,the system with real-time reflecting of ash deposition condition in furnace can instruct the soot blowing be taken according to the demand.Therefore,the energy consumption caused by blind soot blowing is reduced.

关 键 词：330 MW机组 锅炉 神经网络 智能吹灰 受热面 吹灰频率 清洁系数 

分 类 号：TK323[动力工程及工程热物理—热能工程]