Analysis and Simulation of Bearing Fault Detection Using FFT Analyser for Self-Aligning Ball Bearing

Abstract

Self-aligning ball bearing is the essential component in various industrial machinaries where it is required to allow certain specific amount of rotating in the components. Defect in such bearing due to various factors will cause the machine to produce unnecessary vibration and noise which cause eventually breakdown of the machinery or loss to components.

There are numerous predictive maintenance techniques, Vibration monitoring is the most effective technique to detect mechanical defects in rotating machinery. In the present study, self-aligning ball bearing (Specification No. 1207k) has been taken for simulating and evaluating the methodology of fault detection experimenting the methodology of fault detection experimentally. The experimental setup has been designed and fabricated to perform the study of fault diagnostics. Artificial defects on the defect free bearing has been generated using wire cut EDM process on various components viz. inner race, outer race, rolling elements and cage to simulate the faulty conditions experimentally under various working parameters like bearing load and rotational speed. The fault diagnostics has been performed using vibration measurement with 4 channel FFT analyzer using uni axial accelerometer.

The vibration plots in both time and frequency domain has been obtained for both healthy and defective ball bearing. The predictive correlation between the performance of healthy and the faulty bearing has been modeled using Artificial Neural Network by using various statistical quantities from vibration plots. The obtained network that can be used to predict the fault inside the bearing for the purpose of predictive maintenance.

The results obtained has been used to derive co relation with various types of defects and the spectrum obtained for the same. The same result can be used as benchmarks for predicting unknown defects in bearings used for various machinery.