Quantifying Cost of BESS for Providing Power System Regulation Service Considering Capacity Regeneration

Predicting capacity loss is the premise for battery energy storage system (BESS) owner to evaluate the marginal cost of providing auxiliary service and also the key to enhance the market competitiveness of BESS. Most existing studies are devoted to predicting capacity degradation with empirical models, ignoring the uncertainties during degradation such as the capacity regeneration phenomenon. This paper proposes an operation cost model that considering uncertain capacity degradation. By signal decomposition algorithm, the capacity curves are divided into trend, regeneration and random noise parts. The RBF neural network is used to predict the trend and regeneration parts. In contrast, the SVR model is employed to predict the random noise. On this basis, this paper utilizes the Maryland University dataset to train these data-driven models, which can reduce the RMSE from 4.25% To 1.56%. The regulation service market clearing simulation indicates that the proposed model enhances the profitability of BESS by an average of 6.78% and a maximum of 9.51%. This is attributed to accurately predict capacity loss and identify capacity regeneration phenomena.