Abstract
In many regions, energy models are pointing towards variable renewable energy (VRE), mainly solar photovoltaics (PV) and wind power, as potential backbone of future power systems. However, such models have not typically been able to explicitly include, from the outset, geospatial aspects around future VRE power plant siting in the cost-optimisation. For instance, should one build far from the grid in search of excellent resources, or rather stay close to existing infrastructure to keep costs low? Here, we present an OSeMOSYS-FlexTool workflow for capacity expansion and dispatch optimisation that explicitly includes the geospatial dimension of VRE buildout. Applying the model to Kenya, we identify clear but nontrivial links between site characteristics and siting preferences in the optimisation results, which moreover differ markedly for solar PV and wind power. This highly replicable approach has important implications for power system planning, especially in countries whose grids will be VRE-heavy in the future.
