An Endogenous Approach to Quantifying the Wind Power Reserve

Abstract

Due to inherent uncertainty and volatility, large-scale wind power integration requires additional reserves for power balancing. However, it remains an open question as to how to quantify the required reserves corresponding to different uncertainties and penetration scales of wind power. This paper addresses this problem with an improved endogenous reserve determination method. Multiple uncertainties, including wind power outputs, load fluctuations, and generator failures, especially power flow uncertainties caused by wind power fluctuations and generator faults, are comprehensively integrated into a security and network-constrained economic dispatch (S&NCED) model to jointly schedule the generation and reserve. Affinely adjustable robust optimization (AARO) is adopted to address wind power and load uncertainties. Then the model is transformed into a two-stage (normal state optimization and fault verification) S&NCED model for iterative solving with the aid of Benders decomposition, and auxiliary constraints are designed to accelerate the convergence. Results of 6-bus and IEEE118-bus systems show that the proposed method can effectively provide a more economical and robust generation-reserve dispatch scheme and can clearly specify the reserve capacity required by wind power.