Optimized dispatch of energy storage systems based on improved battery model

Abstract

Energy storage systems (ESS) are widely applied in power grids to absorb renewable energy sources, shift demands, and balance short-term electricity. However, the traditional dispatch methods ignore the battery's dynamic power limit and degradation characteristics, which leads to the mismatched power between ESS dispatch commands and the actual optimal responses, and shortened battery lifetime. This paper proposes a novel battery model to achieve an optimized dispatch of ESS. First, a model with a dynamic power limit is developed to vary the power limit with the state of charge. Second, a multi-factor degradation model is established to quantify the degradation of the battery during charging/discharging. However, the optimized model of ESS becomes nonconvex when the battery power and degradation characteristics are incorporated. In order to solve the nonconvex model with standard solvers, the proposed battery model is transformed into a nonlinear mapping function; then, a prediction–correction algorithm with a series of convex models is introduced to approximate the original nonconvex model. As shown in the case study, the proposed model achieves a more accurate allocation of ESS power and provides higher cost-effectiveness over the lifetime of ESS than the traditional one. Moreover, the introduced prediction–correction algorithm outperforms the heuristic algorithm in terms of computational time and global optimality.