Online Identification Of AC Motor Misalignment Using Current Signature Analysis and Modified K-Mean Clustering Technique

Advances in metal rolling process automation and tightening quality standards result in a growing demand being placed on fault detection and diagnostics of electrical motors. Misalignment of motor or coupled load on motor shaft is one of the common causes, which creates most of the mechanical faults and leads to motor vibration. Although different algorithms are available for motor condition monitoring, but an online identification of motor misalignment and comprehensive fault reporting to the maintenance personnel are still missing. The motor current spectrum analysis for misaligned motor is not well documented. This paper portrays a novel online fault diagnostic algorithm related to misalignment of induction motors fed by variable speed drive. The innovative approach features spectral analysis and clustering based, fault detection method. A new set of feature coefficients of the mechanical faults is extracted from the stator current by its spectral decomposition. The technique is validated experimentally for a 7.5-hp induction motor.