A rolling element bearing fault diagnosis approach based on hierarchical fuzzy entropy and support vector machine

Abstract

Aiming at the non-linear characteristics of bearing vibration signals as well as the complexity of condition-indicating information distribution in the signals, a new rolling element bearing fault diagnosis method based on hierarchical fuzzy entropy and support vector machine is proposed in this paper. By incorporating the advantages of both the concept of fuzzy sets and the hierarchical decomposition of hierarchical entropy, hierarchical fuzzy entropy is developed to extract the fault features from the bearing vibration signals, which can provide more useful information reflecting bearing working conditions than hierarchical entropy. After feature extraction with hierarchical fuzzy entropy, a multi-class support vector machine is trained and then employed to fulfill an automated bearing fault diagnosis. The experimental results demonstrate that the proposed approach can identify different bearing fault types as well as severities precisely.

Investigation of a multi-sensor data fusion technique for the fault diagnosis of gearboxes

“…Artificial intelligence (AI) techniques, which can effectively analyze a mass of data and automatically provide accurate diagnosis results, has good potential for this. [6][7][8][9] However, the gearbox vibration signals are usually non-linear, non-stationary and noisy. Some conventional AI techniques such as support vector machine (SVM) and back propagation neural network (BPNN) are shallow architectures, which contain no more than one non-linear transformation, do not easily learn the complex non-linear relationships in the data.…”

A Rolling Element Bearing Fault Diagnosis Approach Based on Multifractal Theory and Gray Relation Theory

“…One of the most common manners is to refine and extract the fault features from vibration signals by combination of a few of advanced signal processing methods (e.g., wavelet package transform (WPT) [8,10], hilbert transform(HT) [10,11], empirical mode decomposition (EMD) [11] and higher order spectra (HOS) [12]), to recognize the fault frequency used to compare with the theoretical value with involvement of expert’s empirical judgement. With the advent of artificial intelligence, the process of rolling bearing fault diagnosis is more and more treated as a procedure of fault pattern recognition, and its effectiveness and reliability significantly depend on the selection of dominant characteristic vector of the fault features [13]. Recently, Some entropy based methods, e.g., approximate entropy (ApEn) [14,15], sample entropy (SampEn) [16], fuzzy entropy (FuzzyEn) [16,17], hierarchical entropy (HE) [13,18] and hierarchical fuzzy entropy (HFE) [13], have been proposed to to extract characteristic vector of the fault features from bearing vibration signals.…”

“…With the advent of artificial intelligence, the process of rolling bearing fault diagnosis is more and more treated as a procedure of fault pattern recognition, and its effectiveness and reliability significantly depend on the selection of dominant characteristic vector of the fault features [13]. Recently, Some entropy based methods, e.g., approximate entropy (ApEn) [14,15], sample entropy (SampEn) [16], fuzzy entropy (FuzzyEn) [16,17], hierarchical entropy (HE) [13,18] and hierarchical fuzzy entropy (HFE) [13], have been proposed to to extract characteristic vector of the fault features from bearing vibration signals. In the paper, we introduce a multifractal theory based method, i.e., a generalized multifractal dimension algorithm, to extract dominant characteristic vector of the fault features from bearing vibration signals.…”

“…Usually, after feature extraction, a pattern recognition technique is needed to achieve the rolling element bearing fault diagnosis automatically [13]. A series of pattern recognition methods have been introduced for mechanical fault diagnosis nowadays, among which the most commonly used ones are artificial neural networks (ANNs) [20,21,22] and support vector machines (SVMs) [23].…”

“…However, the accuracy of a support vector machine (SVM) classifier is highly decided by the selection of optimal parameters for SVMs [24,25]. To ensure the diagnostic accuracy, an optimization algorithm [24,25] or / and complex multi-class concept [13,26] have to be used subsidiarily to improve the effective of SVMs. Here, in order to solve the issue of generality versus accuracy, an adaptive gray relation algorithm was developed to achieve accurate patter recognition based on a small number of samples.…”