Data-Driven Health Monitoring of Medium Voltage Drives Using Industrial IoT Platforms

The recent transformation of industrial production under the industry 4.0 paradigm is driven by the convergence of Industrial Internet of Things (IIoT), AI, and high-speed communication technologies. Power electronic systems composed of high-power semiconductor devices are subjected to complex degradation that arises from thermal stress and electrical loading. High-frequency monitoring of MV drives generates large volumes of data, creating a significant issue with communication bandwidth and operational costs. These problems create a strong requirement for intelligent data management strategies that are capable of preserving diagnostic relevance while minimising unnecessary data transmission. To address this problem, this project proposes a data-driven edge-cloud health monitoring framework for MV drive systems through IIoT platforms. The proposed system integrates embedded edge computing units within MV drives to perform real-time signal processing, feature extraction, and neural network-based novelty detection. Only data segments associated with abnormal or previously unseen operating conditions are transmitted to the cloud. The normal operational behaviour is summarised through compact health indicators. The cloud layer facilitates long-term learning across distributed drive fleets and assists in decision-making. Experimental results proved that the proposed framework effectively decreases data transmission requirements while maintaining a high fault detection performance. The intelligent edge-cloud coordination can significantly enhance the feasibility of data-driven condition monitoring for MV drives.