Robustness of cost-optimal energy system designs: The role of short-term extreme weather events and dispatchable generation

Modeling energy systems typically requires multiple weather years to ensure accurate results. However, running large-scale energy system models with multiple weather years is computationally intensive. To address this, a typical meteorological year is often used, which represents the long-term characteristics of historical weather. However, relying solely on such a representative year is insufficient due to the growing frequency of extreme weather events. It is crucial to consider how energy systems perform under extreme conditions to ensure a secure supply. This study addresses this need by focusing on weather-related extreme conditions of energy systems and employing various methods to generate synthetic weather years for a sector-coupled energy system. By configuring two system scenarios, we aim to identify a robust energy system configuration capable of accommodating all historical years. Additionally, we examine the characteristics of this system configuration and identify critical factors that influence system robustness to weather variability. Our results demonstrate the important role of dispatchable generation technologies in maintaining the security of supply. Furthermore, we find that short-term extreme events, such as 17 or 18 consecutive hours of extremely high residual load, can impose significant stress on the energy system, often exceeding the impact of longer-term extreme events.