Distributionally Robust and Transactive Energy Management Scheme for Integrated Wind-Concentrated Solar Virtual Power Plants

Houbo Xiong,Fengji Luo,Mingyu Yan,Lei Yan,Chuangxin Guo,Gianluca Ranzi

DOI: https://doi.org/10.1016/j.apenergy.2024.123148

IF: 11.2

2024-01-01

Applied Energy

Abstract:In the pursuit of a near‑carbon-emission electric sector, concentrated solar power plants (CSP) and wind generators have gained prominence, promising dispatchable electricity for renewable-dominated grids. However, the existing studies focus on the coordinated scheduling of CSP and wind energy, overlooking the critical issue of energy pricing and trading. Moreover, a decentralized model for multiple networks that incorporate both CSP and wind generators, remains under-investigated. Accordingly, this paper proposes a fully decentralized distributionally robust transactive energy management (DRTM) framework for the energy trading, pricing and scheduling across multiple integrated wind-concentrated solar virtual power plants (IWC-VPP), using the alternating direction method of multipliers (ADMM). This model allows each IWC-VPP operator to make independent decisions and share minimal information, ensuring privacy encryption. Based on the distributionally robust optimization (DRO), the DRTM framework can balance robustness and cost-effectiveness in making decisions under uncertainties. For efficient resolution, an adaptive buffer-column and constraint generation (AB-C&CG) algorithm is introduced, which reduces the complexity of the master problem compared to the traditional C&CG. Additionally, a varying penalty factor technique is integrated into ADMM to accelerate computation, and a two-block process is implemented to ensure finite convergence of the entire decentralized framework. Numerical studies on the three-VPP 25-Bus system and four-VPP 156-Bus system validate the effectiveness of the proposed DRTM framework. The simulation results demonstrate the varying penalty factor technique bolsters computational efficiency by up to 46.51% for standard ADMM. Compared with the conventional C&CG, the AB-C&CG significantly reduces the computational consumption by 50.98%, and with the error <0.46%.