Developing an Edge based predictive maintenance system for transformer using low-cost sensors

Transformers are essential for ensuring the reliable distribution and transmission of electrical power. Traditional maintenance procedures, which are generally reactive and schedule-driven, can result in costly breakdowns and unanticipated expenses. This research offers an edge-based predictive maintenance system for transformers that uses powerful machine learning techniques and five temperature sensors strategically distributed throughout the transformer in non-invasive manner. The system uses edge computing to handle real-time temperature data locally, allowing for more precise and accurate anomaly detection and predictive analytics. This strategy enables proactive maintenance, timely interventions, and less operational disturbances. The proposed methodology optimises maintenance schedules, lowers downtime, and increases grid reliability, providing a viable method for managing essential infrastructure and improving overall grid dependability.