A day-ahead market model for power systems: benchmarking and security implications

TL;DR

The paper critiques OPF-based security assessments by introducing a social-welfare-based day-ahead market (DAM) within the EuroEM framework and evaluating its security implications with the Cascades cascading-failure model. The DAM captures profit-driven bidding while ensuring grid-consistent redispatch, enabling a market-realistic security analysis. Benchmarking against economic dispatch (ED) and unit commitment (UC) on a three-zone IEEE-118 system shows that DAM elevates price levels and storage/gas activity, and significantly increases the cumulative demand-not-served (DNS) relative to OPF, highlighting potential security risks overlooked by traditional analyses. The findings suggest that market-cleared dispatch can push the system toward more critical conditions, underscoring the need for enhanced reserves, expansion planning, and market-informed security assessments. The work provides a transparent, open framework for evaluating market impacts on security and offers a path toward more realistic, risk-aware forecasting in power-system operations.

Abstract

Power system security assessments, e.g. via cascading outage models, often use operational set-points based on optimal power flow (OPF) dispatch. However, driven by cost minimization, OPF provides an ideal, albeit unrealistic, clearing of the generating units, disregarding the complex interactions among market participants. The security of the system, therefore, may be overestimated. To address this gap, we introduce a market model with a social-welfare-based day-ahead market clearing mechanism. The security implications are analyzed via Cascades, a cascading outage analysis framework. We apply this framework to the IEEE-118 bus system with three independent control zones. The results show that market dispatch leads to an increase in demand not served of up to 80% higher than OPF, highlighting a security overestimation. Operators can use this information to properly allocate reserves and perform efficient expansion planning strategies.

Figures (4)

• Figure 1: Overview of the EuroEM algorithm. The long-term positions optimization sets the bidding structure for the day-ahead market model (DAM). The results from the DAM clearing are passed to the redispatch and security models.
• Figure 2: Comparison of the clearing prices in the ED and DAM models. Very high prices ($\ge$100 €/MWh) are removed from the histograms to avoid disrupting the scale. a) histogram of the clearing prices of the ED model; b) histogram of the clearing prices of the DAM model; c) number of hours with very high prices in the ED and DAM models.
• Figure 3: Comparison of the system performance between the DAM, ED, and UC models, and optimal power flow (OPF) based dispatch. The comparison includes a) the cumulative number of failures of total branches, interconnectors, adjacent branches, and intra-zonal branches; b) the demand not served (DNS).
• Figure 4: Risk curves of the cascading simulation model using OPF (black), economic dispatch (ED, yellow), unit commitment (UC, blue), and day-ahead market dispatch (DAM, red). The curves that are more pronounced on the right side exhibit worse cumulative demand not served DNS. The models use the same set of cascade-initiating contingencies in the four assessments.