Automatic detection of bearing faults

Abstract

The potential for machine Condition Monitoring (CM) to enhance system performance and forestall harmful failures has been rising considerably. It is applicable to all or any rotating machinery where preventive fault diagnosis could be a must like Engines, drive trains, gearboxes, pumps, turbines, compressors, fans etc. Detection of bearing faults is one in all the foremost challenging tasks in bearing health condition monitoring, especially when the fault is at its initial stage. The defects in bearing unless detected in time may result in malfunctioning of the machinery. Acquisition of vibration signals coming from various machine components, performing its analysis for prediction of faults and understanding the root cause for high vibrations has been an established practice. However, while performing analysis of vibration signals, it is close to impossible to separate out and focus on ‘bearing frequencies’ especially in the presence of strong masking signals from other machine components. In this paper a study and implementation of ‘advanced’ vibration analysis technique viz envelope analysis for localization of bearing frequencies from the masking signals generated by other machine components is discussed. A smart graphical user interface-based software tool has been developed for automatic detection of bearing faults. The automatic detection techniques presented in this paper can be very helpful for condition monitoring especially for early diagnosis of bearing faults. This may help industries for minimizing the downtime due to machine breakdown. With the help of vibration sensor (accelerometer), FFT analyzer, set of faulty bearings installed with rotating arrangement and a graphical user interface-based software tool designed with the help of multi-paradigm programming language, the setup for automatic detection of bearing faults has been validated.