← Back to Portfolio
Power System Forecasting Tool@ University of Calgary
Lead ML Engineer
A comprehensive tool that forecasts electricity generation and demand and optimizes power flow to minimize total system production cost while respecting transmission system constraints. Combines machine learning forecasts with Security Constrained Optimal Power Flow (SCOPF) optimization using Gurobi and a PyQt-based Single Line Diagram (SLD) interface.
Technologies
PythonScikit-learnXGBoostRandom ForestNeural NetworksGurobiMySQLTkinter
Highlights
- Implemented XGBoost, Random Forest, and Neural Networks achieving R² > 0.9 and MAPE ~10% on demand forecasting
- Full simulation completes in 13.6 seconds — a notable improvement over industry standards
- Seamlessly integrated ML forecasts with SCOPF optimization (Gurobi) for cost-optimal power dispatch
- Built robust data frameworks for historical load and real-time weather data integration