往复压缩机的状态监测和故障的诊断方法

Research on Condition Monitoring and Fault Diagnosis Approach for Reciprocating Compressors

分析了用传统的时域和频域分析方法不能有效提取往复压缩机故障特征的原因,介绍了基于小波包分析与神经网络的往复压缩机故障诊断方法,探讨了包括往复压缩机振动信号的降噪、小波包分解与重构、故障特征提取、针对防止发生漏诊或误诊问题而提出的组合RBF网络及其训练方法和渐进学习能力等问题。还专门介绍了一种新的技术,它可以帮助我们确定一个适当的阈值,用于解释经过训练的RBF分类器的输出。

This paper analyzes the causes of the difficulties to extract the fault features of reciprocating compressors by conventional time-domain and frequency-domain methods,introduces the fault diagnosis approach based on the wavelet packet analysis and neural network,explores the methods of reducing noise on vibrational signal,decomposition and reconstruction of wavelet packet,extraction of fault features,composite RBF-network for preventing wrong or leak diagnoses and its training approach and the ability of incremental learning.Especially,this paper introduces a novel technique which may be used to determine an appropriate threshold for interpreting the outputs of a trained RBF classifier.