Processor in the Loop Based Rotor Fault Diagnosis of an Induction Machine

Abstract

Induction machines (IM) are asynchronous/electromechanical machines usually used as motors where they have many advantages as they are rugged, less expensive, and require less maintenance. These advantages make them the most used machines in the industry nowadays. However, induction machines can be found mostly in hazardous environments where they are exposed to harsh conditions resulting in failures that lead eventually to the machine downtimes, therefore, production shutdowns, financial losses, and waste of raw materials. Hence, to prevent such catastrophic consequences, online detection and diagnosis of such faults becomes of interest. In this study, the induction machine model has been developed along with faulty cases when the squirrel cage bars/end ring are cracked. Using MATLAB software, the model is simulated for healthy and faulty cases along with Fast Fourier Transform analysis. Three faulty cases were taken into consideration; where, two adjacent, two separated broken bars and the end ring cracking are simulated. Moreover, and as a preliminary to real time diagnosis, the FFT is implemented on the STM32 as a processor in the loop (PIL) to quickly detect the failure in the IM. Simulation results show that time domain analysis could only categorize whether the IM is healthy or faulty. Whereas, spectral analysis would give more insight on the failure by detecting the number of broken bars. On the other hand, the FFT implementation on the STM32 board raises promises to give initial real diagnosis failures that existed on the induction machine.