Enhancing Reliability in Induction Machines: A Focus on Stator Inter-Turn Fault Diagnosis

Induction Machines (IM) are widely favored in contemporary industry due to their affordability, robustness, and economical maintenance requirements. Despite these advantages, IMs are susceptible to various faults, such as bearing, stator winding, and rotor issues, which can cause significant operational disruptions, production delays, and considerable economic losses. Addressing these concerns, the importance of early fault diagnosis cannot be overstated, as it significantly improves machine reliability and efficiency, prevents further damage, and decreases the dependency on extensive repairs post-breakdown. The study focuses on early detection of Stator Inter-turn Faults (SITF) in IMs through an online diagnostic approach, utilizing a dataset from both Healthy and Faulty IMs equipped with a 2.2kW four-pole squirrel cage. An exploratory analysis has been performed on the dataset to simplify the data structure and understand feature dynamics. The most straightforward classification method has been applied to identify SITF occurrences. The effectiveness of each diagnostic technique was evaluated by comparing their accuracy levels. The research successfully developed both non-neural and neural network models capable of detecting SITFs with minimal severity, demonstrating the potential for enhancing industrial operations through early fault diagnosis.