Short-term storage signals in hydrothermal dispatch models using a linked representative periods approach

Short-term energy storage systems (STESS), e.g., batteries, are becoming one promising option to deal with flexibility require-ments in power systems due to the accommodation of renewable energy sources. Previous work using medium- and long-term planning tools has modeled the interaction between STESS and seasonal storage (e.g., hydro reservoirs). Despite these develop-ments, opportunity costs considering the impact of STESS sig-nals in stochastic modeling have not been analyzed. This paper proposes a new formulation to include STESS operational deci-sions in a stochastic hydrothermal dispatch model, which is based on Linked Representative Periods (LRP) approach that allows an analysis of both short- and long-term storage at the same time. This proposal models operating decisions of STESS with errors between 5% to 10%, while the classic Load Duration Curve (LDC) approach fails by an error greater than 100%. Moreover, the LDC model cannot determine opportunity costs on an hourly basis and underestimates the water value by 6% to 24% for seasonal hydro reservoirs. On the other hand, the pro-posed LRP model produces an error on the water value lower than 3% and can determine hourly opportunity costs for STESS using dual variables from both intra- and inter-period storage balance equations. Therefore, hourly opportunity costs in the LRP model successfully internalize long-term signals due to sea-sonality in hydro reservoirs.