Comparative Analysis of Dynamic and Linear Programming Energy Systems Models Applied to the Bolivian Power System

Abstract

Energy system models are indispensable tools for energy planning and decision making. They identify the most cost-effective way of delivering energy to the final consumer. No one tool that addresses all the energy systemrelated issues. Every model has its own strengths and limitations and serves a different purpose. This paper aims to compare the capabilities of two different model formulations to model both the hydro scheduling and the short-term dispatch problems in hydro-dominated power systems. On the one hand, SDDP, a commercial model for hydrothermal generation scheduling with a representation of the transmission network, has been used by the Bolivian system operator for dispatch simulations. Conversely, Dispa-SET, an open-source unit commitment and economic dispatch model with mid-term hydrothermal coordination capability, has been used previously in several Bolivian case studies. In this paper, both models were applied to the same input dataset of the Bolivian electric system considering probabilistic results for 43 weather years from 1984 to 2021. SDDP optimizes the system under all weather years, while Dispa-SET optimises under one full year, for which 43 runs were made. The results show that SDDP generation, reservoir level and spillage fall into the ranges of Dispa-SET results. Some differences that are present mainly lie in the conceptualization of the methods of both models. SDDP prioritizes the dispatch of hydro units, while Dispa-SET, with a higher temporal and technical resolution, maximizes the use of non-dispatchable units such as variable renewables and run-of-river.