Abstract
Applied research developing ML algorithms for real-time energy demand forecasting and grid load balancing, enabling 15-20% reduction in energy waste for urban distribution networks.
Research Methodology
Collaborative research with two state electricity boards. Historical consumption data analysis, ML model training on 5 years of grid data, and live pilot deployment in Bangalore metropolitan area.
Key Findings
- Forecasting accuracy of 96.8% for 24-hour demand prediction
- Peak load reduction of 18% during pilot period
- Estimated annual savings of ₹45 crore for participating utility
Publications
- Patel, A. & Krishnan, M. (2025). ML for Grid Optimization. IEEE Transactions on Smart Grid.