摘要
分析研究了汽轮发电机组振动智能故障诊断技术,将人工神经网络技术与面向对象技术相结合,建立了振动频谱、轴心轨迹、升降速特性和负荷特性等4个征兆神经网络,同时构建了具有不完全征兆输入的汽轮发电机组振动智能故障诊断神经网络系统。以机组振动频谱征兆为例,研究了频谱征兆的自动提取方法,给出了基于频谱征兆的不完全征兆综合故障诊断实例。在此基础上,采用B/S模式和Java技术,开发了汽轮发电机组远程智能故障诊断系统,介绍了系统的结构组成、功能模块以及服务器和客户端程序设计和实现方法。
Analyzed and studied are the techniques of intelligent fault diagnosis of vibration for turbogenerator sets.By combining artificial neural network technology with object-oriented one,a four symptom neural network has been established.The four symptoms are vibration frequency spectrum,axial-center trajectory,speeding-up-and-down characteristics and load characteristics.Meanwhile,constructed was an intelligent fault-diagnosis neural network for sensing vibrations of steam turbogenerator sets with incomplete symptom inputs.With the frequency spectrum symptoms of turbogenerator set vibrations serving as an example,a method for the automatic acquisition of frequency spectrum symptoms was studied and a specific case was given of comprehensive fault diagnosis with an incomplete symptom based on the frequency spectrum symptom.On this basis,by using a Browser/Server mode and Java technology,an intelligent remote fault-diagnosis system for turbogenerator sets was developed along with a description of the structure composition of the system,functional modules,servers and client-terminal program design and implementation method.
出处
《热能动力工程》
CAS
CSCD
北大核心
2006年第5期532-535,共4页
Journal of Engineering for Thermal Energy and Power
关键词
汽轮发电机组
振动
神经网络
智能故障诊断
远程诊断
turbogenerator set,vibration,neural network,intelligent fault diagnosis,remote diagnosis
