Broken bars detection on induction motor using MCSA and mathematical morphology: An experimental study

Currently, industry demands early failure detection on his processes, machines, production lines, etc. One of the most widely used motors in industry is the induction motor. A common induction motor failure is the broken bars. It is well know that broken bars produce spurious frequencies around the supply frequency. Moreover, the amplitude of the spurious frequencies in the sideband of the main frequency is sensitive to the number of broken bars. In this paper a real-time pre-processing methodology to enhance detectability for broken bar detection using motor current signature analysis and mathematical morphology is presented. The proposed methodology is implemented into a low cost FPGA. A statistical analysis is presented in order to demonstrate the detection improvement.