摘要
在具有齿轮、滚动轴承的机械设备的振动故障诊断中 ,广泛使用解调方法分析诊断故障。从理论上分析了现有的解调分析方法存在有三个局限性 :将不包括调制信息 (故障信息 )的两时域相加信号 ,也以其频率之差作为解调信号而解出 ;广义检波滤波解调分析中 ,由于取绝对值或检波过程可能产生混频效应 ;几种细化解调分析新算法中 ,无法在细化分析的选抽时进行数字低通滤波 ,有可能会出现调制频率的高次谐波成分发生频率混叠而反折到低频部分的现象。无论那种局限性出现 ,都会在机械振动信号的解调谱中出现无法分析的频率成分 ,给分析带来很大的困难 ,有时甚至引起误诊断。
Demodulation analysis is widely employed to diagnose the faults of machines fixed with gears and bearings. To analyze the algorithms of current demodulation methods, we found that there are three limitations for their applications. One is that the subtraction of non modulation signal frequencies (without fault information) will be demodulated from the two time domain added signals. The other one is that aliasing phenomenon will occur as a result of getting absolute value or detection in the process of generalized demodulation analysis. The last one is that aliasing frequencies will be produced in zoom demodulation analysis because this algorithm can not employ digital low pass filtering to avert the folding frequencies of higher harmonics in the process of zoom sub sampling. These will result in wrong diagnosis and some superfluous frequency composition which will puzzle the detection of mechanical vibration.
出处
《机械科学与技术》
CSCD
北大核心
2000年第5期722-725,728,共5页
Mechanical Science and Technology for Aerospace Engineering
关键词
信号处理
解调分析
故障诊断
机械振动分析
Mechanical vibration
Signal processing
Demodulation analysis
Spectrum
Fault diagnosis
