A Neuro-Fuzzy Approach for Broken Rotor Bar Fault Prediction in Induction Motors via Current Spectrum Analysis

Industry 4.0 have paved way for new technological advancements especially with the commencement of smart devices. The combination of information technology and physical devices has opened up new research areas and possibilities to revolutionize industrial processes. However, accurate fault prediction, particularly for broken rotor bars in induction motors, remains a challenging problem due to the complexity of fault signatures in electrical signals. This paper presents a neuro-fuzzy model that combines the pattern recognition capabilities of neural networks with the reasoning power of fuzzy logic to enhance fault prediction accuracy. Unlike conventional methods that rely solely on time-domain analysis, this approach leverages current signal spectrum analysis, extracting meaningful fault indicators to improve prediction reliability. Experimental validation demonstrates that the proposed neuro-fuzzy network outperforms traditional neural networks, achieving a validation accuracy of 99.60 %, compared to 95.74 % for standard models. The integration of fuzzy logic enhances the model's ability to handle nonlinear and complex fault patterns, reducing false positives and improving generalization. These findings suggest that the neuro-fuzzy network is a promising tool for high-precision fault prediction in induction motors, contributing to the development of more robust and intelligent condition monitoring systems. Future research will focus on extending the model to account for varying load conditions and more intricate fault scenarios, further enhancing its adaptability in industrial applications.