Diagnosing Faults in Rolling Bearings of an Air Compressor Set Up Using Local Mean Decomposition and Support Vector Machine Algorithm

This paper describes an air compressor fault diagnosis method based on audio signals and applied to one faulty and one healthy condition of rolling bearing considering uni-directional microphone along with a single NI 9234 Data Acquisition (DAQ) hardware unit having multiple ports, a NI 9172 USB interface, and a Lab-VIEW based Data acquisition interface.

Acquired non-stationary and non-linear bearing fault signals are processed using most recent non-traditional Local Mean Decomposition (LMD) signal processing technique. Further, six statistical indicators (Mean, Variance, RMS, RMA, AMA, and Kurtosis) have been evaluated for feature extraction. Further, four classifying techniques namely: SVM, Naïve Bayes, KNN, and Discriminant have been used for the exploration and classification of the fault features in rolling bearing of an air compressor set-up.

It has been observed that LMD along with Kurtosis and SVM machine learning approach is quite accurate for processing and monitoring in situ rolling bearing fault features in air compressor set-up.