Evaluation of methods for long-term hydropower planning in electricity market models

This paper evaluates three different methods for long-term planning of hydro dominated power systems. Two established electricity market models using water values, EMPS and BID3, are compared, as well as a look-ahead dispatch which relies on the average optimal reservoir curves computed from a long-term steady-state optimization of the system. All three methods are evaluated using out-of-sample simulations using 100 sequential weather years, to compare the resulting dispatch of the system. It is found that the proposed look-ahead dispatch comes quite close to the established models in terms of average electricity costs. Of the evaluated models, BID3 gives the best results, both in terms of average electricity costs and considering that it does not require any load shedding in the evaluations, unlike EMPS and the look-ahead dispatch. However, the results from all three models can be significantly improved by calibration. While BID3 is a bit too risk averse using the initial settings, the other models have too little risk aversion, which results in significant load shedding.