Faulty bearing vibration diagnostic & monitoring

Sri Tharan, Piravin Thiran (2022) Faulty bearing vibration diagnostic & monitoring. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Text (Full Text) Faulty bearing vibration diagnostic & monitoring.pdf - Submitted Version Download (3MB)

Abstract

Vibrations is something that is unavoidable and will occur in machinery. Over a period of time, this vibration magnitude increases and becomes detrimental to the machinery if it is not monitored or reduced to a safer level. Researchers around the world are continuously conducting studies on bearing vibrations to improve or propose a solution to problems that arise due to the bearing vibrations. Many components can cause these vibrations especially bearings where it moves in a oscillating pattern which causes the machinery to vibrate in a certain frequency and can effect the machine’s performance if it is left undetected and diagnosed. High vibration magnitudes indicate that the bearing is faulty and needs to be diagnosed and if left untreated it can increase expenditure to correct the failure and reduce the machine’s lifetime. Bearings usually experience increase in vibration magnitude as it experiences damages to it over time as bearings are exposed to constant motion throughout its operation. Defective rolling element that comes into contact with another element surface causes the generation of impact force which results in an impulsive bearing response. Due to this increase in vibration magnitude, machinery experiences poor performance. Due to this it is paramount that the bearing’s vibration condition is always monitored and if there is an increase in its vibration magnitude it should be diagnosed immediately. To remedy this, vibration signal analysis can be done as it is an effective vibration monitoring technique which this thesis has done. This thesis focuses on the roller bearing vibrations under normal and damaged conditions where an experiment at speeds of 400 rpm, 800 rpm, 1200 rpm and 1600 rpm under six different bearing conditions were conducted to measure each of the condition’s vibration levels. This bearing’s vibration was then analyzed using the MATLAB and Excel software through Vibration Statistical Analysis (VSA). Through the results, there is an increase in the transient components where it increases together with the running speed as the speed increases. Through the graphs also, it can be seen that the vibration with frequency increases in amplitude as the speed increases. Through z-freq data scattering and its coefficient, it is observable that all the dots spreads across the affix and annex frequency where through this it can be said that the scattered data shows a clear pattern as the speed increases. RMS for conditions A, B, C, D, and E can be seen that the RMS increases as the speed increases. R2 for each of the bearing conditions have a value of 0.9103, 0.9351, 0.9205, 0.9389, 0.7199 and 0.3592 respectively. It can be said that conditions A, B, C, and E has the closest accuracy in the speeds of 400 rpm, 800 rpm and 1200 rpm compared to the other type of conditions and speed. Conclusively, it can be said that time domain is not suitable for fault prediction as compared to the R-Squared method as in time domain, the graph difference between the faulty and good conditions are similar to the frequency domain graphs. This thesis presents a method for predicting the condition that produces the specific vibration through the RMS and R2.

Actions (login required)

View Item