Power Equipment Management System using Long Short-Term Memory and Gated Recurrent Unit

For operational dependability and safety in the age of smart grids and intelligent infrastructure, efficient power equipment management is essential. This study develops and implements a Power Equipment Management System utilizing a hybrid deep learning approach, specifically integrating Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) algorithms. Time-series sensor data is processed using these models for anomaly detection, fault diagnosis, and predictive maintenance. To predict probable equipment failures, the system combines vibration signals, temperature, voltage, and usage patterns from the past. Comparative analysis revealed that the LSTM model offered higher long-term predictive accuracy, while the GRU model showed faster convergence. the recommended method attains the highest accuracy in the Power Equipment Management System, with 98.23% of prediction accuracy, 98.2% Specificity, 98.1% Sensitivity, and 98.2% F-measure. In power distribution networks, the suggested system improves proactive decision-making and maximizes resource allocation.