A multi-scenario distributionally robust model for resilience-oriented offshore wind farms and transmission network integrated planning considering typhoon disasters

Abstract

Existing resilience-oriented offshore wind farms and transmission network integrated planning (ROWF&TNIP) models lack detailed characterization of the uncertainties associated with wind power and grid faults during typhoon disasters, and tend to be relatively conservative in enhancing resilience. To address these limitations, this paper proposes a multi-scenario distributionally robust model for ROWF&TNIP considering typhoon disasters. This model accounts for multiple uncertainties in wind power and grid faults under both normal operation scenario (NOS) and typhoon disaster scenario (TDS), and enhances resilience in a less conservative manner. Firstly, the multi-scenario distributionally robust uncertainty sets for offshore wind farms (OWF) output and grid fault are established: a conditional value-at-risk (CVaR) based multi-scenario budget uncertainty set to capture the uncertainties of wind turbine outputs and turbine failures under NOS and TDS, and a 1-norm grid fault uncertainty set to represent the uncertain probability distribution of four types of fault: high-probability faults, high-loss faults, cascading faults under TDS and fault-free state under NOS. Subsequently, a multi-scenario distributionally robust ROWF&TNIP model is formulated, utilizing the worst-case expected load-shedding cost under TDS as resilience index, the planning and expected generation cost under TDS and NOS as economic index. This model coordinates resilience and economic efficiency under the most adverse realization of uncertain OWF outputs and grid faults. To further mitigate the conservatism of the ROWF&TNIP model, short-term source-grid-load measures, including preventive unit commitment, differential load-shedding and an innovative differential hardening model, are integrated to the planning model. A column and constraint generation (C&CG) based decomposition algorithm is developed to solve the model. In case study section, a series of comparative and sensitivity analyses are conducted on the IEEE-30 bus system and a Chinese 81-bus system to demonstrate the effectiveness of the proposed model and reveal how key parameters of the model influence the resilience and economy of the planning results.