Hakan Ergun

Long-term energy system models are widely used to support decarbonisation planning, but their policy relevance depends on how well they represent the interaction between technology choices and grid constraints. Energy planning models often simplify the transmission grid through aggregated inter-regional exchanges, potentially overlooking nodal congestion that affects renewable integration, flexibility needs, and siting.This paper develops and tests an iterative bidirectional soft-linking framework between FENICE (Future Energy transition multI-seCtor modEl), a regional multi-sector long-term energy system model, and an optimal power flow (OPF) model. FENICE provides regional investment, demand, storage, and transmission-expansion decisions, while the OPF model evaluates the power flows through a nodal representation of the transmission grid. The OPF assessment is performed with hourly resolution for each representative year, and the resulting nodal flows are aggregated into regional flow constraints for subsequent FENICE iterations.The framework is applied to the Italian energy system. Compared with the uncoupled FENICE results, the iterative framework yields different technology portfolio and spatial allocation of investments. In 2050, wind capacity decreases by 25%, battery capacity more than doubles, thermoelectric generation capacity rises ten-fold, and curtailment increases by 60%. The differences concentrate in specific regions and transfer corridors affected by recurrent congestion, especially along island-to-mainland and south-to-north routes.These results provide evidence that simplified grid representations can distort policy-relevant trade-off results and may affect decisions on renewable siting, storage needs, dispatchable capacity, and transmission constraints. Embedding a detailed transmission grid representation in long-term planning can therefore allow network constraints to influence investment pathways, rather than leaving transmission feasibility as a separate assessment.

The shift towards a sustainable power system presents formidable challenges in both planning and operation due to the intermittent nature of renewable energy sources.This paper focuses on Transmission Expansion Planning (TEP) formulations without intertemporal constraints and evaluates five system state aggregation methods — one ex-ante and four ex-post — across three test systems of varying size and spatial correlation. Through comparative analysis of accuracy and computational cost, we show that ex-post methods carry substantially higher pre-processing overhead than the ex-ante baseline. For the simplest system, no method outperforms the ex-ante approach; for larger systems, the continuous hourly investment (CHI) method achieves the best accuracy among ex-post methods but at a computational cost 5–17× higher than the ex-ante baseline. In contrast to earlier studies favouring ex-post clustering, this was not found to be consistently superior regarding any attribute, challenging assumptions about the universal effectiveness of some methods. We propose a practical method-selection framework to guide practitioners in choosing appropriate temporal aggregation strategies under different system conditions.