Dynamic dimensioning approach for operating reserves: Proof of concept in Belgium. (January 2019)

• Record Type: Journal Article
• Authors: De Vos, K.; Stevens, N.; Devolder, O.; Papavasiliou, A.; Hebb, B.; Matthys-Donnadieu, J.
• Is Part Of: Energy policy. Volume 124(2019)
• Page Start: 272
• Page End: 285
• Publication Date: 2019-01
• Subjects: Balancing -- Dynamic sizing -- Machine learning -- Operating reserves -- System operation; Energy policy -- Periodicals; Politique énergétique -- Périodiques; Electronic journals; 333.79
• DOI: 10.1016/j.enpol.2018.09.031
• Languages: English
• ISSNs: 0301-4215

Abstract

Abstract: This article discusses a new method for the sizing of operating reserves by electric power system operators. Operating reserves are used by system operators to deal with unexpected variations of demand and generation, and maintain a secure operation of the system. This becomes increasingly challenging due to the increasing share of renewable generation based on variable resources. This paper revisits the current sizing method applied in Belgium, which is based on a static approach that determines the required capacity once a year. The presented dynamic sizing method determines the required capacity on a daily basis, using the estimated probability of facing a system imbalance during the next day. This risk is estimated based on historical observations of system conditions by means of machine learning algorithms. A proof of concept is presented for the Belgian system, and demonstrates that the proposed methodology improves reliability management while decreasing the average capacity to be contracted. The method is compliant with European market design, and the corresponding regulatory framework, and is of particular interest for systems with a high share of renewable generation. For these reasons a gradual implementation in Belgium towards 2020 has been decided based on the results of this study. Highlights: The article presents a method for dimensioning operating reserves on daily basis. Advanced statistical tools such as machine learning for estimating the imbalance risks. It improves reliability management while decreasing average reserve capacity. A proof of concept presents a positive business case for the Belgian system in 2020. The Belgian TSO foresees a gradual implementation of the method towards 2020.