Cross-market optimization for a hydro pumped storage using dynamic programming

Abstract: Following the liberalization of the electricity sector and the increase of renewables, the design of the electricity markets has evolved. Optimization techniques and models for power system generation scheduling including multiple markets are nowadays crucial. In this thesis, a solution has been developed in order to optimize the schedule of a hydro pumped storage while considering electricity markets with di˙erent time resolutions. The model is based on the German market design and uses deterministic dynamic programming techniques. Based on past data, the final model allows backtesting analyses to assess previous strategies. The usage of forecasted prices and forecasted executions in the reserve markets gives the possibility to provide planning and indicators to the traders. Unique price scenarios and executions scenarios are considered due to the deterministic characteristic of the model. The optimization can compare simultaneously di˙erent spot markets and provides the corresponding optimal schedule. Adjustments can be made to consider the tertiary reserve market. In the tertiary reserve market, each o˙er corresponds to an upward or downward regulation for a given period i.e. an increase or decrease in the injected energy in the power system. Only exclusive options are considered for a fixed volume chosen at the beginning of the simulation: either a possible upward regulation is o˙ered, or a downward for a given period and the fixed volume. In addition, a successful o˙er for the primary or secondary reserve market can be assumed as a fixed parameter at the beginning of the simulation. A demonstration of the optimization is provided for a hydro pumped storage defined in the thesis and validates the usage of dynamic programming. The simulations highlight the interest in considering multiple markets with high time resolution to increase the potential profit. Moreover, the financial interest of the reserve markets is shown for a hydro pumped storage. The model allows an assessment of the di˙erent trading possibilities in the reserve markets. Suggested improvements to the model and potential for future work can be found in the final chapter of this thesis.