Do multi-sector energy system optimization models need hourly temporal resolution? A case study with an investment and dispatch model applied to France

Abstract

Energy system optimization models (ESOMs) increasingly cover the main energy-consuming sectors rather than just electricity, which massively raises calculation time. To reduce the latter, researchers apply various time-series aggregation methods, the pros and cons of which have been analyzed for electricity-only ESOMs but not for ESOMs also covering the main energy-consuming sectors. To address this question we compare the two main time-series aggregation methods: (1) reducing the temporal resolution (from one to two, four or eight hours) and (2) selecting representative periods (one week over one, two or three months, with an hourly resolution). We apply these methods to EOLES_mv, an ESOM covering the main French energy sectors. Both methods cut the calculation time by a similar amount but the former generates much smaller discrepancies for the main output variables (energy mix, system cost and CO2 emissions). These results are at odds with those generally obtained with electricity-only ESOMs, for which reducing the temporal resolution generates significant discrepancies when wind and solar dominate the electricity mix.