Search

Condition monitoring plays a vital role in predictive maintenance of machinery in today's world. It is very important to detect the faults in the machinery as early as possible in order to stop the propagation of faults which may lead to heavy damages. The objective of our study is to investigate the vibration analysis of bevel gearbox to detect the faults. An experimental setup was developed to carry out the vibration analysis. Various types of faults were induced in the gearbox. The vibration analysis was carried out for different types of faults with different lubrication levels in the gearbox. MATLAB toolbox was used for signal processing in which Coiflet wavelet was used for denoising the signal. An artificial neural network (ANN) and a deep neural network (DNN) were used to detect the faults in the bevel gearbox automatically. The results were promising in detecting the faults with high accuracy. [ABSTRACT FROM AUTHOR]

This study aims at developing a novel method for condition monitoring technique for detection and classification of developing faults and increase the working life of continuous variable transmission (CVT) using Daubechies Wavelet 06 (DB-06). The vibration data is collected for 4 different types of faults and healthy condition. Using a magnetic accelerometer and signal analyser, vibration data is collected from the system in the time-domain which is then used as input for a MATLAB code producing the plot of the frequency-domain signal. Maximum frequency is determined to diagnose the faults which are induced over three different belts. Collected data for large scale automotive system (CVT) is used to train the network and then it is tested based on random data points. Faults were classified using ANN with a classification rate of 90.8 %. [ABSTRACT FROM AUTHOR]