摘要
针对火力发电厂湿法除灰系统中柱塞泵的泄露问题 ,介绍了频域和时域相结合检测故障的方法 .文中对比了声音样本的波形特点 ,利用加窗傅立叶变换分析了时域短时能量和频域短时自相关函数、功率谱函数 ,提取了声音样本的频率特征量 ,证明了利用特征频率点附近功率谱函数方差值检测故障的可行性 。
This paper introduces a real time method to diagnose fault of axial piston pump system used in power plant to pump up wet ashes. Some characters of sound samples such as waveform, short time power, autocorrelation function and power spectrum function are analyzed. Then the characteristic frequencies are extracted and the variance of power spectrum at these characteristic frequencies are used to judge the fault. Finally, an engineering implementation based on time domain and frequency domain information of sound is provided. The algorithm in this fault diagnosis system is simple so that multi channel sound signal can be monitored simultaneously.
出处
《东南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2003年第z1期130-133,共4页
Journal of Southeast University：Natural Science Edition
关键词
实时检测
方差
功率谱
故障特征
声音
real time fault diagnosis
variance
power spectrum
fault information
sound
