Graphical User Interface Aided Intelligent Diagnosis of Stator Faults in Induction Motors

Induction motors form the backbone of many modern industries; therefore, the performance and safety of these industries rely heavily on motor health. The stator faults account for more than one-third of all motor faults. The existing literature primarily focuses on current/vibration-based models for stator fault diagnosis, without any attention to providing a suitable Graphical User Interface as well. To alleviate these drawbacks, an end-to-end intelligent solution using vibration signal has been proposed that allows the end-users to upload the acquired motor data in the application and get step-by-step fault diagnosis results without requiring any coding knowledge. The solution has been made flexible for use on a wide variety of datasets by applying AutoKeras model with the capacity to select a suitable deep-learning model and adjust its parameters in accordance with the specific characteristics of the dataset. The proposed method performs stator fault detection and severity prediction with 99.81% accuracy. The prediction outcome is interpreted by Explainable AI to improve model transparency and provide trustworthiness. The user-friendly Graphical User Interface using Streamlit enables easy dataset upload, EDA visualization, and interaction with predictions, bridging ML methodologies with practical applications. This research will prove very beneficial in facilitating Industry 4.0.