A novel health indicator developed using filter-based feature selection algorithm for the identification of rotor defects

Abstract

In this work, a novel health indicator is developed for the identification of rotor defects. The indicator is developed by extracting features from vibration data acquired from horizontal and vertical directions of rotors. A total of 38 features were initially extracted from time-domain signal, frequency-domain signal, and time–frequency representation. Out of many features, six most important features were selected using filter-based feature selection process. Thereafter, important features were fused together using manifold learning to develop health indicator. The developed indicator is used to identify misalignments (angular misalignment and parallel misalignment), rub, and unbalance. The major benefit of the proposed method is that it not only indicates the presence of defect in the rotor but also indicates the severity of defect. The experimental study presented in this article justifies that the proposed method is sensitive to the increasing levels of horizontal and angular misalignment and unbalance. The developed indicator is sensitive enough to indicate the presence of rub.

An Intelligent Hybrid Feature Selection Approach for SCIM Inter-Turn Fault Classification at Minor Load Conditions Using Supervised Learning

“…In their methodology, they selected features were fed to a support vector classifier model which gave impressive accuracy in classifying various induction motor conditions. Kumar, Gandhi, Liu, Liu, Zhou, Kumar, and Xiang [22] developed a novel health indicator for misalignment rotor fault detection using a filterbased feature selection technique. They used a filterbased algorithm (relief algorithm) in their technique to select 6 features out of the 38 extracted features; these important selected features were then fused with the help of manifold learning to achieve a health indicator.…”

Optimal Statistical Feature Subset Selection for Bearing Fault Detection and Severity Estimation

“…Finally, T30 or normal negative log likelihood for single Gaussian can be computed as follows: [31]. Moreover, these techniques perform well for the identification of dominant features for both fault type and severity detection [32].…”

Coronavirus herd immunity optimizer with greedy crossover for feature selection in medical diagnosis

“…These four stages can be achieved by applying a FS method, such as a filter, wrapper, embedded or hybrid method. The wrapper method is distinct from the filter method in that it employs a learning algorithm during the evaluation phase, whereas the filter method evaluates specific features independently of the classification process by using a certain standard threshold [23,24]. The embedded method is similar to the wrapper method as a classifier is used in the selection process in the evaluation step, but the use of the classifier in the embedded method is comparatively less cost-effective than in the wrapper method [25].…”