An algorithm to diagnose ball bearing faults in servomotors running arbitrary motion profiles

This paper describes a procedure to extend the scope of classical methods to detectball bearing faults (based on envelope analysis and fault frequencies identification)beyond their usual area of application. The objective of this procedure is to allowcondition-based monitoring of such bearings in servomotor applications, where typicallythe motor in its normal mode of operation has to follow a non-constant angularvelocity profile that may contain motion inversions. After describing and analyzingthe algorithm from a theoretical point of view, experimental results obtained on areal industrial application are presented and commented