Induction Motor Bearing Fault Diagnosis Based on Singular Value Decomposition of the Stator Current

Among the most widespread systems in industrial plants are automated drive systems, key and common element of which is induction motor. In view challenging operating conditions equipment, task fault detection based on analysis electrical parameters relevant. The authors propose identification patterns characterizing occurrence development bearing defect by singular method as applied to stator current signature. As a result decomposition, time series three-phase represented triples ordered decreasing contribution, reconstructed into form for subsequent using Hankelization matrices. Experimental studies with damage imitation made it possible establish relationship between changes contribution presence different levels defects. By level tracking movement specific series, became observe both appearance new components signal corresponding overall structure. verified clustering results visual assessment component matrices’ structure similarity scattergrams hierarchical clustering. reconstruction from grouping allows use these prediction faults electric motors.