机械故障诊断过程中的“类随机电报”问题

The Problem About Quasi-Random Telegram in the Mechanical Failure Diagnoses

针对工程实际中所存在的测试仪器本身的饱和限幅影响 ,提出所谓“类随机电报”问题 ,并对测试仪器不同饱和深度所造成的“类随机电报”过程的自相关函数进行了计算机仿真研究 ;本文以类随机电报为背景 ,成功地提取了弱周期性故障信号 ;文章最后给出结论 :当振动噪声较强而受测试仪器饱和特性一定程度的限幅时 ,仍然可以采用循环自相关分析方法提取淹没于其中的周期性微弱故障信号 。

The Quasi-random Telegram (QT) problem is presented against an effect of saturation and amplitude limit of testing instruments in engineering.Simulation by computer was carried out for auto-correlation function of QT,which was engendered by different saturated degree of testing instruments.Weak periodical fault signals can be extracted successfully from the background of QT.Finally the conclusion was given.When the vibration noises were so stronger that their amplitudes were limited to a certain extent due to saturated characteristic of testing instruments,weak periodical fault signals merged in the vibration noises can be extracted by use of cycling auto-correlation method,so as to detect the machine faults in its early stage.