Evaluation of reliability index and worth analysis for on day-ahead unit commitment using LLP

Massive inclusions of Renewable energy generation on the grid, increases uncertainty in generation schedule decision. As a result, traditional reliability index such as Lost of Load Probability (LOLP) may not be reliable criteria to evaluate system reliability. Thus, a revised formulation of unit commitment (UC) is required, which not only optimize day-ahead generation schedule for system with high penetration of renewable energy, but also stay committed to the security of the system as well. In this paper, our novel approach of a revised UC formulation with system security considerations and dynamic spinning reserve (SR) is presented. With considerations of uncertainties in both demand and renewable power generation, together with random outages of generators, an analytic algorithm is presented to calculate the optimal required spinning reserved for every time interval and the reliability cost. Our approach considers the reliability of system grid and the hourly forecast of variable demand and renewable generation sources such as wind and photovoltaic, together with their uncertainties. Particle Swarm Optimization (PSO) is utilized to solve the revised UC formulation. Numerical tests are performed and the results are analyzed for an IEEE RTS bus system. Simulation results on certain 4-unit case system validates the efficacy of our methodology.