Distribution system planning considering peak shaving of energy station

• Published in: Applied energy Vol. 312; p. 118692
• Main Authors: He, Shuaijia, Gao, Hongjun, Liu, Junyong, Zhang, Xi, Chen, Zhe
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 15-04-2022
• Subjects: Decentralized; Distribution systems; Distributionally robust optimization; Energy station; Peak shaving; Distribution systems; Decentralized; Peak shaving; Energy station; Distributionally robust optimization
• ISSN: 0306-2619; 1872-9118
• DOI: 10.1016/j.apenergy.2022.118692

•It is the first time to adopt energy station to shave the peak electric load in the DS planning.•The input energy of energy station is scheduled to participate in the peak shaving of DS.•This paper formulates the DS and ES planning in a decentralized framework.•The TOU price is introduced by an explicit coupling formula to regulate the exchanged power.•An easily reformulated solution method is proposed to cope with the PD uncertainty of loads.
Energy stations (ESs) connected in a distribution system (DS) may lead to great impacts on the planning scheme of DS. In this context, this paper carries out a long-term DS planning model considering the peak shaving of ES, which is achieved by scheduling the input energy of ES. By regarding DS and ES as different stakeholders, a decentralized framework is devised to shave the electric peak loads in the DS planning, where the coupling relationship between the TOU price and exchanged power (e.g., the input power of ES) is clearly expressed. Specifically, an explicit adjustment formula is developed to represent the coupling relationship based on the concept of elasticity. In addition, an easily reformulated solution method is developed to address the probability distribution (PD) uncertainty of electric and cooling loads in the uncertainty-moment-based distributionally robust optimization (DRO) planning model. The chance-constrained power balance is expressed in a second order conic (SOC) format based on the conditional value at risk (CVaR) method and duality theory. Then, the SOC constraints are linearized according to the polyhedral linearization method. Furthermore, the bilinear terms of the planning model are respectively linearized by the McCormick and big-M methods. Finally, the proposed planning model is tested on a modified IEEE 33-node DS with an ES and a practical 99-node DS with an ES. Numerical results show that the proposed planning model is effective in managing PD uncertainties of loads as well as reducing costs of DS.