Performance degradation evaluation method based on FCM-HMM rolling bearing

Abstract

The invention discloses a performance degradation evaluation method based on an FCM-HMM rolling bearing. The method comprises the steps that an autoregressive model (AR) is used to extract the characteristic of the vibration signal of an early fault-free rolling bearing and the failure characteristic of a failure rolling bearing of the same type and position (bearing of the same kind in short for the sake of description); a fuzzy C-mean value (FCM) performance degradation evaluation model is established through an early fault-free sample and a failure sample; an HMM model is established through the fault-free sample; in view of the advantages of FCM and HMM, degradation indexes acquired through FCM and HMM is input into the FCM performance degradation model as input characteristics, so as to acquire normal and failure clustering centers; and an early fault threshold is set at the same time. Experimental analysis shows that the evaluation indexes acquired through the performance degradation method provided by the invention can monitor the performance degradation trend of the rolling bearing in real time, and early faults can be detected in time.

Description

technical field [0001] The invention relates to an FCM-HMM-based evaluation method for performance degradation of rolling bearings, belonging to the technical field of mechanical product quality reliability evaluation and fault diagnosis. Background technique [0002] Most of the failures of rotating machinery are caused by rolling bearing failures. The performance of rolling bearings will be degraded to varying degrees in a long-term wear environment. The performance degradation assessment of bearings proposed in recent years is an extension of fault diagnosis. The essence of performance degradation assessment is to evaluate the corresponding performance of mechanical equipment by analyzing the vibration data of the equipment. The biggest difference between it and the traditional bearing fault diagnosis mode is that the performance degradation assessment is a proactive maintenance method, while the traditional bearing fault diagnosis mode Diagnosis focuses on discovering fa...

AI Technical Summary

Problems solved by technology

Wu Jun et al. used the data after 500 samples in the whole life cycle as the sample data to input into the fuzzy C-means to obtain the cluster centers of normal and fault states, and then evaluated the performance degradation trend of rolling bearings, but the consistency between the degradation curve and the fault was not good.

However, GMM tends to reach the upper limit of the value when the bearing does not fail, and cannot accurately determine the failure state of the bearing; HMM often has high computational complexity and consumes a lot of time, which is difficult to meet the needs of online real-time evaluation

Images