Evaluating Representative Days Selection for Capacity Planning with Variable Renewable Options

Capacity planning models are necessary to study pathways to meeting decarbonization goals and meeting load growth. As these models are computationally costly, representative days are frequently used to reduce the problem size. The choice of representative days becomes even more important as variable renewable energy sources like wind and solar are increasingly becoming a part of the electricity mix. Ultimately, the set of selected days will vary based on the method used and number of days chosen. In turn, the outputs of the planning model can vary based on the set of days used - this work focuses on quantifying this variation. First, this paper introduces some existing methods for set selection, weighing additional considerations required when addressing variable renewables and regional variations. A series of metrics is then proposed for comparing the outputs of the planning models from different representative day sets. Finally, they are applied, in PyPSA, to Canada’s Maritime provinces as a $\mathbf{2 5}$-year grid planning problem and compared to a 365-day approach. Results indicate that there are a wide range of near-optimal solutions with contrasting generation mixes, and that capacity planning would be better informed by generating results over multiple sets of representative days, rather than a single set.