CVaR-based planning of park-level integrated energy system considering extreme scenarios of energy prices

•A novel scenario generation method capturing extreme scenarios is proposed.•The correlation between energy prices is modelled by the Frank-Copula function.•A conditional-value-at-risk-based planning method is proposed.•The effectiveness and advantages of the proposed method are validated. The uncertainty and correlation of energy prices pose challenges to the planning and operation of park-level integrated energy systems. Especially, the extreme scenarios in both electricity and natural gas prices would result in significant economic losses, which have not been fully studied in the planning of park-level integrated energy systems. In this regard, this paper proposes a conditional-value-at-risk-based planning method for the park-level integrated energy system that considers both the extreme scenarios of energy prices and the correlation between the electricity price and natural gas price. First, the Gaussian mixture model is used to model the probability distribution function of the electricity price and natural gas price with high accuracy. Then, the Frank-Copula function is adopted to model the correlation between energy prices in different energy markets. Furthermore, a novel scenario generation method preserving extreme scenarios of energy prices is proposed, where extreme price scenarios are identified in the scenario generation process and subsequently preserved in scenario reduction. Finally, a conditional-value-at-risk-based planning model is proposed to optimize the overall investment and operation costs of the park-level integrated energy system, where the operational risks introduced by renewable generation and energy markets are effectively handled by the conditional value at risk. Case studies demonstrate that compared with the model without the consideration of extreme scenarios of energy prices, the potential risk loss of the proposed model is reduced by 3% in the same generated scenarios. The proposed model is more beneficial in enhancing the system’s resilience to the risk.