Temporal and spatial tradeoffs in power system modeling with assumptions about storage: An application of the POWER model

Abstract

As the number and complexity of power system planning models grows, understanding the impact of modeling choices on accuracy and computational requirements becomes increasingly important. This study examines empirically various temporal and spatial tradeoffs using the POWER planning model for scenarios of a highly renewable US system. First, the common temporal simplification of using a representative subset of hours from a full year of available hours is justified using a reduced form model. Accuracy losses are generally ≤6%, but storage is sensitive to the associated model modifications, highlighting the need for proper storage balancing constraints. Cost tradeoffs of various temporal and spatial adjustments are then quantified: four temporal resolutions (1- to 8-h-average time blocks); various representative day subset sizes (1 week–6 months); two spatial resolutions of site-by-site versus uniform fractional buildout across all solar and wind sites; and multiple spatial extents, ranging from California to the contiguous US. Most tradeoffs yield <15% cost differences, with the effect of geographic aggregation across increasing spatial extents producing the largest cost reduction of 14% and 42% for the western and contiguous US, respectively. These results can help power system modelers determine the most appropriate temporal and spatial treatment for their application.