Risk-Averse Self-Scheduling of Thermal Power Plants: Comparative Study of Deterministic and Stochastic Optimization with Parametric CVaR Analysis

This paper presents a comprehensive comparative analysis of three distinct optimization approaches for power plant self scheduling in competitive electricity markets. We examine deterministic optimization under forecasted prices and scenarios alongside a stochastic Conditional Value at Risk (CVaR) approach. The proposed model incorporates price uncertainty through scenario-based stochastic programming to optimize power generation scheduling while managing financial risk. The study evaluates thermal power plant across 24-hour operational periods, analyzing profitability, operational strategies, and risk profiles. Results demonstrate that the CVaR-based approach provides a robust scheduling solution that balances expected profit maximization with risk mitigation. The model was implemented SIPAT-I thermal generating unit in Indian grid using GAMS and solved using CPLEX, achieving optimal solutions for a 24-hour scheduling horizon under four price scenarios generated from market clearing prices of Indian Energy Exchange (IEX).