K—L变换在有效特征提取中的应用

在机械故障智能诊断过程中,提取反映机器状态的有效特征参量时要花费大量的时间和精力,而且这种依靠人的经验来提取特征参量的方法有一定的盲目性。为了解决如何尽可能快而有效地寻找一组特征参量,使诊断对象不同状态之间的可分性为最佳的问题,采用了K—L变换特征提取方法,研制出相应的软件,并已将该方法应用在大庆油田抽油机减速箱智能故障诊断系统中。

The Application of K-L Transition in The Mechanical Fault Diagnosis

In the process of mechanical fault intelletual diagnosis, it takes us a lot of time to find the effective feature vectors which can reflect the difference among the different conditions of one machine, at the same time it is not reasonable enough to extract feature vectors only by man' s experience. How to find a feature vector quickly and effectively, this paper brings out the method of K-L transition to do the feature extraction, the software is made and put into practice in the diagnosis system of petroleum extractor in Da Qing oilfield.