MACHINE LEARNING BASED SMART TRANSFORMER FAULT PREDICTION WITH AUTOMATED COOLING SYSTEM

: Power transformers are vital assets in electrical power systems, and their failure can lead to severe operational disruptions, economic losses, and safety hazards. Traditional fault detection methods often rely on periodic inspections or threshold-based monitoring, which may fail to capture incipient faults. To address this limitation, we present a machine learning–driven smart transformer fault prediction framework integrated with an automated cooling system utilizing Peltier modules. The proposed system employs advanced data analytics on parameters such as winding temperature, oil level, load current, and vibration signals to identify hidden fault patterns. By training supervised learning models—including decision trees, support vector machines, and neural networks—our approach achieves high predictive accuracy and early warning capability.Complementing the predictive layer, an intelligent cooling subsystem is designed, where Peltier devices dynamically regulate transformer temperature in response to predicted thermal stress. The cooling system is automated through microcontroller-based control logic, ensuring real-time response and energy efficiency. This dual-layer integration of predictive analytics and active cooling not only minimizes the risk of catastrophic failures but also extends transformer lifespan, reduces maintenance costs, and enhances grid reliability. Experimental simulations and prototype testing validate the effectiveness of the system, demonstrating improved fault detection precision and optimized thermal management compared to conventional methods.The proposed solution represents a step toward self-healing, adaptive transformer infrastructure suitable for deployment in smart grids, aligning with the broader vision of sustainable and resilient energy systems.