Condition Monitoring of Induction Motor for Industrial Applications

Induction motors (IMs) are renowned for their high efficiency and are widely utilized as prime movers in various industrial applications. To ensure uninterrupted operation, a precise fault diagnosis system for IMs is essential. Such a system not only enhances operational safety but also mitigates the risk of unexpected economic losses. Traditional diagnostic methods often struggle to adapt to real-time conditions and fluctuating working environments. In response to these challenges, this paper introduces a novel vibro-acoustic fusion technique designed for accurate fault diagnosis under varying operational conditions. The proposed method employs a Multi Input-Convolutional Neural Network (MI-CNN) to fuse features derived from both vibration and acoustic signals. Initially, raw vibration and acoustic signals are collected at different speeds and transformed into a time-frequency spectrum using the Constant Q-Non-Stationary Gabor Transform (CQ-NSGT). Following this, the MI-CNN-based vibro-acoustic fusion technique is applied to integrate the vibration and acoustic features effectively. To evaluate the effectiveness of the proposed MI-CNN model, six distinct motor conditions are analyzed. Additionally, two supplementary datasets—bearing and gearbox datasets—are utilized to validate the robustness of the proposed approach. The experimental results demonstrate that this methodology is both accurate and reliable for diagnosing faults in induction motors and other components of rotating machinery, paving the way for improved maintenance strategies and enhanced operational efficiency.

Development of an Intelligent Condition Monitoring System (CMS) for Wind Turbine