Jaiyeolacole/Edge-Driven-Predictive-Maintenance-System-for-a-Reliable-and-Sustainable-Power-Grid

This repository features the development and implementation of Artificial Intelligence (AI) and Machine Learning (ML) models for Predictive Maintenance (PdM), designed to significantly enhance grid reliability and extend the lifespan of infrastructure assets within Nigeria's power sector ⚡ AI-Driven Predictive Maintenance for Power Grid Reliability This repository contains an end-to-end AI-powered Predictive Maintenance (PdM) system built to enhance the reliability, resilience, efficiency, and sustainability of Nigeria’s power infrastructure. The project integrates Machine Learning through Edge Impulse, IoT sensor data, and a modern React dashboard to proactively detect failures and optimize maintenance strategies for critical energy assets. Problem & Context Nations continues to face long-standing challenges in its electricity sector: - Insufficient and unstable power generation - High technical and operational losses - Frequent equipment failures - Inadequate or reactive maintenance practices - Limited visibility into asset health across transmission and distribution networks Despite significant government investment, conventional approaches cannot keep up with the complexity of modern energy systems. Critical infrastructure such as transformers, turbines, breakers, feeders, and solar arrays often degrade silently — leading to unplanned outages, damaged equipment, and expensive repairs. Solution Overview: This project introduces a data-driven, AI-enhanced predictive maintenance solution designed to preempt failures, reduce downtime, and improve grid reliability. Key Technical Features: -- 1. Advanced AI/ML Algorithms The system leverages: - Machine Learning models - Deep Learning architectures (CNN, RNN/LSTM) - Time-series anomaly detection - Failure probability estimation - Remaining Useful Life (RUL) prediction -- 2. Real-Time IoT & Sensor Data Integration The model processes high-frequency data from assets such as sub-stations, electric poles etc: Datas such as the following are processed on the edge device attach to each sub station: Voltage, Current, Power, Frequency, Power Factor Temperature, Vibration, Pressure GPS-based environmental context Historical maintenance logs This multi-dimensional dataset enables the system to det …