International Conference on NDT and Material Industry and Alloys (IC-WNDT-MI'14).
Feature Extraction and SOM for Bearing Fault Diagnosis
Tawfik Thelaidjia1, Abdelkrim Moussaoui 2, Salah Chenikher 3, Amar Boutaghane1, Sami Kahla1
1 Welding and NDT Research Center (CSC) PB 64, Cheraga, Algeria .
2 Laboratory of Electrical Engineering University of Guelma, Algeria.
3 Laboratory of Electrical Engineering University of Tebessa, Algeria.
Abstract— In this paper a method for fault diagnosis of rolling bearings is presented. It consists of two parts: vibration signal feature extraction and condition classification. The aim of the first step is the extraction of the relevant parameters; the proposed technique consists of preprocessing the bearing fault vibration signal using a combination of the signal’s Kurtosis and discrete wavelet transform (DWT). The Self-organization Map (SOM) is used to accomplish the classification step and automate the fault diagnosis procedure. The results have shown feasibility and effectiveness of the proposed approach.
Index Terms— Condition monitoring, Discrete wavelet transform, Fault Diagnosis, Kurtosis, Roller Bearing, Rotating machines, Self-organization Map, Vibration measurement.