As distributed solar generation becomes increasingly prevalent, the temporal granularity of energy system models directly influences design quality and operational feasibility. Traditional hourly modeling has dominated the industry due to computational constraints, yet it misses rapid fluctuations in solar irradiance, load demand, and inverter cycling that occur at sub-hourly timescales.
This research quantifies the performance gap between minute-resolution and hourly modeling across single buildings and aggregated microgrids. The key finding is that minute-level temporal detail substantially improves capacity expansion results and component sizing decisions. Inverters emerge as a critical system bottleneck, particularly sensitive to sub-hourly dynamics; their oversizing or undersizing using hourly data can lead to suboptimal configurations and underestimated costs.
The study demonstrates that unserved energy—load periods when renewable supply cannot meet demand—is consistently underestimated in hourly models. This underestimation directly affects system reliability assessments and sizing decisions. For single-building photovoltaic systems, the effect of temporal resolution is pronounced; for larger aggregated systems, diversity effects partially dampen the impact, though inverter sizing remains critical.
To address computational complexity, the researchers tested three reduction techniques: time-series averaging, regular sampling, and clustering-based aggregation. Results show that carefully combined approaches—such as averaging with selective sampling or clustering methods—preserve 95% of minute-resolution accuracy while achieving computational runtimes comparable to traditional hourly models.
These findings have immediate practical implications. Grid planners and solar developers can adopt hybrid modeling strategies: deploy minute-resolution analysis where solar penetration is high or inverter constraints are tight, then apply validated aggregation methods for broader system studies. The research suggests future improvements through enhanced hourly datasets with subhourly detail, enabling better accuracy without prohibitive computational costs.
The balance between temporal fidelity and practical computational burden is essential as energy systems continue rapid renewable integration.
