Solar Scheduler – AI-Powered Smart Energy Forecasting & Device Scheduling System
Link to open source: https://github.com/Tracebycode/Solar-Scheduler
Solar Scheduler is a full-stack intelligent energy management platform that forecasts solar power generation and automatically schedules electrical devices to maximize energy efficiency. The system was built under strict constraints where external weather APIs were not allowed. Instead, it uses historical irradiance datasets and an ARIMA + Persistence ensemble model to locally predict solar generation every 15 minutes. The backend scheduler consumes these forecasts and optimally allocates battery power and device runtimes based on priority levels (Critical, Flexible, Optional). To ensure reliability during emergencies, the platform also provides a Manual Override mode that allows operators to directly control devices without affecting the stability of the scheduling algorithm. Key Features • 24-hour solar generation forecasting using ARIMA ML model • Historical CSV-based training (no external APIs) • Automatic device scheduling every 15 minutes • Manual emergency override control • Priority-based energy allocation • Battery & panel capacity simulation • Real-time dashboard with charts, timelines, and device states Architecture • ML Engine: Python (ARIMA forecasting, JSON output) • Backend: Node.js + Express (REST APIs, scheduler, state management) • Frontend: React (dashboard, devices, settings UI) This project demonstrates practical skills in machine learning integration, system design, real-time scheduling algorithms, and full-stack development for renewable energy optimization.
This build was uploaded as a hackathon project
