摘要
针对大、中型火电机组广泛采用的大旁路布置 U型管管板式高压加热器系统 ,通过详细的故障仿真试验 ,较完整地总结出工程中实用的高压给水加热器系统典型故障模糊知识库 ,并详细阐述了趋势型和语义型 2种征兆的模糊提取、合成和转化方法。在此基础上 ,采用前向多阶层神经网络诊断高加系统故障 ,提出了一种恒误差修正率控制和网络学习率自适应调整方法 ,可大大提高网络训练的收敛效率。图 6表 5参
A whole bypass U pipe high pressure feedwater heater system is taken as the subject investigated in this paper because it is widely adopted in a large capacity thermal power unit. Its typical fault fuzzy knowledge library is generalized completely with detailed fault simulation tests by putting the heater system into the whole thermal system under the advantages of our power station simulation technology. The heater system fault diagnosis is further realized with BP neural network method. Figs 6, tables 5 and refs 9.
出处
《动力工程》
CSCD
北大核心
2002年第1期1615-1621,1588,共8页
Power Engineering
基金
国家电力公司科技项目 ( SPKJ0 16 -2 2
