Certain investigations on the monitoring and detection of inter turn short circuit faults in induction machines using signal processing and soft computing techniques

Abstract

In recent days, fault diagnosis and detection of Induction Motor (IM) is becoming more important in industry as a result of the requirement to improve reliability and reduce potential production losses due to machine failure. Due to environmental stress or a range of other factors, induction motors can develop a wide range of faults. These faults if left uncharted at the incipient stage may lead to acute deterioration of the machine. Many investigators have proposed various methods and approaches for fault detection and diagnosis. This thesis focuses on induction machines inter-turn short circuit (ITSC) fault detection and monitoring through online. The primary difficulties in IM fault analysis are with incipient stage fault detection. Hence with accurate feature extraction of the induction motor parameters, the goal can be accomplished. Here, a three-phase IM is used to conduct an exploratory investigation of stator inter-turn fault in the incipient stage. This thesis proposes a method to detect and monitor the fault occurring in induction motors, by making use of Fuzzy Logic Fault Detector (FLFD), Fast Fourier Transform (FFT) and Wavelet Packet Transform (WPT). The performance of the Fuzzy Logic (FL) fault detector with hybrid membership function is superior in detecting incipient inter turn winding fault in an induction motor during Short Circuit (SC) condition than the triangular and trapezoidal membership functions. To overcome the Fuzzy Logic Fault Detector (FLFD), Fast Fourier Transforms (FFT) and Wavelet Packet Transform (WPT) were proposed for extracting the stator current and vibration signatures to get an optimal result. The fast Fourier Transform (FFT) and Wavelet Packet Transform (WPT) are used in the proposed technique for authentic fault identification. newline