Stationary Wavelet Singular Entropy and Kernel Extreme Learning for Bearing Multi-Fault Diagnosis

The self-organizing fuzzy (SOF) logic classifier is an efficient and non-parametric classifier. Its classification process is divided into an offline training stage, an online training stage, and a testing stage. Representative samples of different categories are obtained through the first two stages, and these representative samples are called prototypes. However, in the testing stage, the classification of testing samples is completely dependent on the prototype with the maximum similarity, without considering the influence of other prototypes on the classification decision of testing samples. Aiming at the testing stage, this paper proposed a new SOF classifier based on the harmonic mean difference (HMDSOF). In the testing stage of HMDSOF, firstly, each prototype was sorted in descending order according to the similarity between each prototype in the same category and the testing sample. Secondly, multiple local mean vectors of the prototypes after sorting were calculated. Finally, the testing sample was classified into the category with the smallest harmonic mean difference. Based on the above new method, in this paper, the multiscale permutation entropy (MPE) was used to extract fault features, linear discriminant analysis (LDA) was used to reduce the dimension of fault features, and the proposed HMDSOF was further used to classify the features. At the end of this paper, the proposed fault diagnosis method was applied to the diagnosis examples of two groups of different rolling bearings. The results verify the superiority and generalization of the proposed fault diagnosis method.

Section: Classificationmentioning

confidence: 99%

“…Its calculation process is shown in Formulas (32)- (34). In addition, in order to evaluate the results of this experiment from different perspectives, F − scores was introduced [40]. Its calculation process is shown in Formulas (32)- (34).…”

Section: Classificationmentioning

confidence: 99%

A New Fuzzy Logic Classifier Based on Multiscale Permutation Entropy and Its Application in Bearing Fault Diagnosis

Guo

Wang

et al. 2019

“…The main research objectives are fault diagnosis and prognostics of bearing. Fu et al [123] fine-sorted dispersion entropy + mutation sine cosine algorithm + particle swarm optimization optimized support vector machine 9 Rodriguez et al [124] wavelet packet Fourier entropy + kernel extreme learning…”

Section: Other Typical Entropy Theories Application On Bearingmentioning

confidence: 99%

“…Fine-sorted dispersion entropy combined with mutation sine cosine algorithm and particle swarm optimization optimized support vector machine is presented to diagnose faults of different sizes, locations and motor loads [123]. In [124], stationary wavelet packet Fourier entropy is used to extract fault features and kernel extreme learning is applied to dealing with these fault features, which can achieve better accuracy results than stationary wavelet packet permutation entropy and stationary wavelet packet dispersion entropy.…”

Section: Other Typical Entropy Theories Application On Bearingmentioning

confidence: 99%

Related Entropy Theories Application in Condition Monitoring of Rotating Machineries

Liu

Zhi

Zhang

et al. 2019

Rotating machinery plays an important role in various kinds of industrial engineering. How to assess their conditions is a key problem for operating safety and condition-based maintenance. The potential anomaly, fault and failure information can be obtained by analyzing the collected condition monitoring data of the previously deployed sensors in rotating machinery. Among the available methods of analyzing sensors data, entropy and its variants can provide quantitative information contained in these sensing data. For implementing fault detection, diagnosis, and prognostics, this information can be utilized for feature extraction and selecting appropriate training data for machine learning methods. This article aims to review the related entropy theories which have been applied for condition monitoring of rotating machinery. This review consists of typical entropy theories presentation, application, summary, and discussion.

“…Due to the advantages, WT has shown its tremendous usefulness in fault diagnosis of rotating machinery. In general, WT can be categorized as continuous wavelet transform, discrete wavelet transform and wavelet packet transform [7,19]. Consider that WT has a stronger ability of local frequency domain analysis for signals, another improved method, denoted as the empirical wavelet transform (EWT), was developed by Gilles [20].…”

mentioning

confidence: 99%

Extreme Interval Entropy Based on Symbolic Analysis and a Self-Adaptive Method

Shi

Zhao

et al. 2019

Self-adaptive methods are recognized as important tools in signal process and analysis. A signal can be decomposed into a serious of new components with these mentioned methods, thus the amount of information is also increased. In order to use these components effectively, a feature set is used to describe them. With the development of pattern recognition, the analysis of self-adaptive components is becoming more intelligent and depend on feature sets. Thus, a new feature is proposed to express the signal based on the hidden property between extreme values. In this investigation, the components are first simplified through a symbolization method. The entropy analysis is incorporated into the establishment of the characteristics to describe those self-adaptive decomposition components according to the relationship between extreme values. Subsequently, Extreme Interval Entropy is proposed and used to realize the pattern recognition, with two typical self-adaptive methods, based on both Empirical Mode Decomposition (EMD) and Empirical Wavelet Transform (EWT). Later, extreme interval entropy is applied in two fault diagnosis experiments. One experiment is the fault diagnosis for rolling bearings with both different faults and damage degrees, the other experiment is about rolling bearing in a printing press. The effectiveness of the proposed method is evaluated in both experiments with K-means cluster. The accuracy rate of the fault diagnosis in rolling bearing is in the range of 75% through 100% using EMD, 95% through 100% using EWT. In the printing press experiment, the proposed method can reach 100% using EWT to distinguish the normal bearing (but cannot distinguish normal samples at different speeds), with fault bearing in 4 r/s and in 8 r/s. The fault samples are identified only according to a single proposed feature with EMD and EWT. Therefore, the extreme interval entropy is proved to be a reliable and effective tool for fault diagnosis and other similar applications.