Scarcity events analysis in adequacy studies using CN2 rule mining

F. Heymann, R. Bessa, M. Liebensteiner, K. Parginos, J.C. Martín Hinójar, P. Dueñas

Energy and AI Vol. 8, pp. 100154-1 - 100154-8

Summary:

Dealing with scarcity events is nowadays gaining relevance in electricity market studies, as traditionally predictable generation and consumption patterns are fading. Policymakers and system planners use therefore adequacy studies to a) understand if the current market design will attract sufficient generation capacity to meet electricity demand in the future and b) to comprehend what drives system inadequacy or resource scarcity when future scenarios lack adequate capacity. This work addressed the latter and showcases a first in-its-kind rule-based methodology that filters scarcity events from a large set of electricity market simulations. In this proof-of-concept, a rule-mining algorithm is applied to outputs from ENTSO-E's Pan-European electricity market model, which is run for 700 model scenarios, each covering 8760 time steps. The developed methodology shows how to unveil potential reasons behind scarcity events in an automated, interpretable, and scalable manner.

Keywords: Electricity markets; Data mining; Security of supply; Renewable energy; Interpretability

DOI reference: https://doi.org/10.1016/j.egyai.2022.100154

Published on paper: May 2022.

Published on-line: March 2022.

Citation:
F. Heymann, R. Bessa, M. Liebensteiner, K. Parginos, J.C. Martín Hinójar, P. Dueñas, Scarcity events analysis in adequacy studies using CN2 rule mining. Energy and AI. Vol. 8, pp. 100154-1 - 100154-8, May 2022. [Online: March 2022]