Coordinated Planning of Grid-Connected Distributed PVs and ESSs in Industrial Parks Considering Long- and Short-Term Uncertainties
Yingzi Wu,Zhiyi Li,Xuanyi Xiao,Sajjad Golshannavaz
DOI: https://doi.org/10.1016/j.est.2024.114501
IF: 9.4
2024-01-01
Journal of Energy Storage
Abstract:Highly flexible energy storage systems (ESSs) can effectively enhance the accessible capacity of distributed photovoltaics (PVs) into distribution networks. However, the uncertainties in source-load dynamics across multiple timescales, particularly within industrial parks, impose constraints on the accessible capacity of PVs and undermine system economic efficiency. To address these challenges, this study proposes a two-stage coordinated planning model for distributed PVs and ESSs, incorporating long- and short-term uncertainties to trade-off between economic performance and operation robustness. Firstly, an incomplete probability distribution set for long-term load uncertainty parameters is established to reduce complexity of high-dimensional modeling, caused by economic and market factors. An adaptive clustering algorithm based on the class homogeneity (CH) index is proposed to characterize long- and short-term uncertainties. Subsequently, a two-stage coordinated planning model incorporating information gap decision theory (IGDT) is developed to obtain planning schemes under different uncertainty levels and facilitate a multidimensional evaluation of these planning schemes. In the first stage, the optimal siting and sizing of PVs and ESSs are determined using a second-order cone programming (SOCP) model. In the second stage, indicators such as extreme output probability, voltage deviation, and environmental impact are employed to assess different planning schemes. Finally, the case study on a 44-bus distribution system in southern China demonstrates that the proposed planning approach enhances the accessible capacity of PVs and improves economic efficiency under long- and short-term uncertainties. Compared to existing methods that consider only long-term or short-term uncertainties, the proposed approach improves economic efficiency by 3.4 % and 8.5 %, respectively.