Optimizing probabilistic spinning reserve by an umbrella contingencies constrained unit commitment

Spinning reserve (SR) is an important resource to deal with the sudden load change, uncertain renewable energy generation, and component failures in power systems. Surplus SR will cause a higher operating cost while insufficient SR will deteriorate the system reliability level. In this paper, the SR is optimized by solving a loss of load probability constrained unit commitment (LCUC) problem. The loss of load probability (LOLP) is an appealing means to enforce the SR requirement in power systems since it can express reliability visually and explicitly. However, till now, the LCUC has not been addressed well in terms of solution accuracy and computation efficiency, due to the highly nonlinear characteristics of the LOLP. In this context, the characteristics of the LOLP are explicitly analyzed. It is found that the nonlinear LOLP constraint can be theoretically expressed as a series of linear constraints and most of the constraints can be relaxed. Then the original LCUC can be equivalently expressed by a new umbrella contingencies constrained unit commitment (UCCUC) model. An umbrella contingency identification process is proposed and the model is solved by the constraint generation technique. The proposed model possesses high computation efficiency with desirable solution accuracy, and thus it can significantly enhance the practicability of the LCUC based SR optimization method in real power systems. The proposed method was validated by case studies with the IEEE-RTS system and several larger systems.