Coordinated expansion planning of integrated electricity and natural gas distribution systems for enhancing both inherent and dynamic resilience: A distributionally robust approach considering decision-dependent uncertainty

This paper proposes a distributionally robust expansion planning method for Integrated Electricity and Natural Gas Distribution Systems (IEGDS) to enhance system resilience under earthquakes considering decision-dependent uncertainty (DDU). Based on expansion planning framework, the portfolio of coordinate planning includes power lines, tie-lines and gas pipelines, distributed generators (DG) including gas turbines (GT) and dual-fuel units (DF), power-to-gas (P2G) devices, storage devices including energy storage (ES) and gas storage (GS) devices. A modeling method of decision-dependent uncertainty is proposed to characterize the fragility curve of different models of power line and gas pipeline under earthquake damages. To further quantify different resilience enhancing effects among investing devices, inherent resilience (IR) and dynamic resilience (DR) are proposed in this paper. The IR penalty primarily evaluates the long-term load-carrying ability of IEGDS, assessing its ability to withstand disasters. In contrast, the DR metric provides a more refined short-term quantification of load shedding penalty costs by incorporating repair crew dispatch model to accelerate the restoration of critical loads. Then a two-stage distributionally robust model with Kullback-Leibler divergence-based ambiguity set is formulated. In the first stage, investment and operational costs are minimized for the base case. In the second stage, it aims to minimize the penalty of inherent resilience and expected penalty of dynamic resilience based on the planning decisions made in the first stage. To efficiently solve the proposed two-stage large-scale planning model with DDU, a progressive hedging (PH)-based column-and-constraint generation (CCG) method is developed. Numerical results demonstrate the economic benefit and resilience enhancing effect of the proposed model.