A Model-based Approach to Detect an Under-Lubricated Condition in a Ball Bearing

Ranjith-Kumar Sreenilayam-Raveendran, Michael H. Azarian, and Michael Pecht
Submission Type: 
Full Paper
phmce_14_045.pdf309.47 KBJune 6, 2014 - 1:46pm

Bearings with an insufficient amount of lubricant can lead to early field failures, especially in applications which fail due to lubricant degradation, such as cooling fans used for thermal management of electronics. A reduced amount of lubricant can accelerate the wear process in the bearing, since there is not enough lubricant film thickness to support the operating load on the bearing. Qualification of bearings in cooling fans is carried out by time-truncated tests, where cooling fans have to operate without failure for a pre-determined period of time. Under-lubricated bearings can survive without failure in these tests leading to the usage of these bearings in the field resulting in field returns and warranty claims.

A non-linear dynamic model of a ball bearing is developed to simulate the transfer of load from the inner race to the outer race of the bearing as well as the acceleration signal as a function of time. An under-lubricated bearing condition is simulated in this model by changing the load transmitted to the outer race due to the reduced amount of lubricant. The simulated acceleration signal of the under-lubricated bearing condition is compared with that from the normal bearing to develop a fault-characteristic feature. The changes observed in the fault-characteristic feature from the simulation is validated by comparing with that obtained from experiments conducted on bearings with varying amounts of grease, ranging from none to the nominal amount specified by the manufacturer. The vibration level of these bearings was monitored at various operating speeds during the experiment. The changes observed in the fault-characteristic feature from the experiment due to a reduction of the lubricant in the bearing were similar to that observed in the simulations. This study resulted in the development of an experimental methodology and a fault-characteristic feature which can be used as a method for rapid acceptance testing of bearings. The dynamic model developed in this study can be used to determine the fault-characteristic feature for any bearing design.

Publication Year: 
Publication Volume: 
Publication Control Number: 
Page Count: 
Submission Keywords: 
bearing fault detection
vibration-based diagnostics
dynamic model
Submission Topic Areas: 
Model-based methods for fault detection, diagnostics, and prognosis
Submitted by: 

follow us

PHM Society on Facebook Follow PHM Society on Twitter PHM Society on LinkedIn PHM Society RSS News Feed