Data-driven alternating current optimal power flow: A Lagrange multiplier based approach. (November 2022)

• Record Type: Journal Article
• Authors: Lei, Xingyu; Yu, Juan; Aini, Habaer; Wu, Wencui
• Is Part Of: Energy reports. Volume 8(2022)Supplement 8
• Page Start: 748
• Page End: 755
• Publication Date: 2022-11
• Subjects: Alternating current optimal power flow -- Data-driven -- Lagrange multipliers -- Stacked extreme learning machine; Power resources -- Periodicals; Energy industries -- Periodicals; Power resources; Periodicals; Electronic journals; 621.04205
• DOI: 10.1016/j.egyr.2022.10.024
• Languages: English
• ISSNs: 2352-4847

Abstract

Abstract: This paper proposes a data-driven Alternating Current Optimal Power Flow (AC-OPF) method assisted by Lagrange multipliers. Stacked Extreme Learning Machine (SELM) is introduced for AC-OPF learning to avoid the time-consuming training and hyperparameter adjustment process of deep neural networks. Instead of incorporating the prior physical information into the neural networks algorithm, we developed a new neural network structure for the SELM learning based on the Lagrange multipliers of the AC-OPF problem. Case studies of several IEEE benchmark systems demonstrate that the AC-OPF learning performance is improved by introducing additional Lagrange multipliers information, while the proposed method outperforms other alternatives with almost 99% learning accuracy and fast computation.