Ball Bearing Fault by Feature Extraction and Fault Diagnosis method based on AI ML Algorithms

Abstract

The bearing is a very important part of rotating machinery because it has a very high failure rate. If the high failure rate in bearing would affect the entire performance of the machinery equipment. In this paper, we present a method for extracting ball-bearing fault features of the Ball Bearing fault. An algorithm for detecting bearing faults using Wavelet Packet Transforms (WPT). Wavelet Packet Transform is used to extract the bearing signal's time-frequency characteristic. Then the Statistical feature Extraction for rolling bearing. ML Algorithm model to recognize the healthy conditions of rotating machinery. The frequency-domaining signals are used to feed the input network. The proposed method is validated using data from Case Western Reserve University's bearing data center. This will demonstrate that both steady-state and unsteady-state situations can be successfully diagnosed by the machine learning algorithm. Instead of using traditional feature technology. The algorithm in this paper has improved defect diagnostics and feature extraction.