Analysis Optimization and Comparison to Detect Failures in the Squirrel-Cage Rotor using High-Level Wavelets

Abstract

The methods and tools used for signal analysis extracted from the induction motors, such as the motor current signature analysis (MCSA) used for data collection on a non-invasive basis, the multi-resolution analysis (MRA) and discrete wavelet transform (DWT), are efficient tools for the signal analysis at different levels or resolutions, these tools have been applied together to improve detection of failures in the rotor of induction motors in condition of no-load. This work focuses on the study of rotor cage end ring, in a condition with lower-load or no-load where uncertainty predominates, this area of study is complicated to analyze correctly with conventional methods, but in these circumstances, the analysis using TDW has better performance. The article presents an alternative way of detecting failures in three phases induction motors in no-load state method with an optimized method and a comparison between results of the analysis with two different levels of the high-order Wavelet Daubechies, studying and evaluating its performance for the detection of broken ring, all this supported with a specific signal pre-processing and post-processing to improve the results of detection in incipient faults.