An Assessment on Condition Monitoring of Automobile Gearbox through Naïve Bayes Approach using Statistical Features

Abstract

The gear trains have been used in devices for power transmission. The gearbox becomes defective and produces noises that cause the unit to vibrate because of the wear and tear process. The frequency of vibration and sound also increases as wear and tear increase, making the gears unsuitable for particular applications. The state of gear trains is important to know before it is replaced by the new one. This paper uses the machine learning approach to classify the status of trains by means of classification models. The research was performed on a lab setup of Triumph Herald, which collected raw vibration signals and extracted significant statistical characteristics, and selected the significant features by means of a J48 classification. The features were chosen were classified in the classification of Naïve Bayes. For the classification of the selected feature, time complexity was 0.15s and the maximum accuracy was 96.75%.