Abstract:

Addressing the conflict between the spatiotemporal uncertainty of extreme ice disasters and the economic constraints of transmission network hardening, this paper proposes a collaborative planning and operation method for enhancing grid resilience. First, regarding disaster modeling, a multi-stage dynamic icing model fusing Generalized Extreme Value Distribution (GEVD) and spatiotemporal random fields is constructed to capture the non-stationary evolution of disasters. Simultaneously, a Continuous-Time Markov Chain (CTMC) is employed to characterize the cascading failure mechanism of multi-split conductors, describing the dynamic decay of transmission capacity under ice accumulation. Second, regarding decision optimization, a two-stage stochastic optimization framework incorporating Conditional Value-at-Risk (CVaR) is proposed. The first stage (planning) co-optimizes differentiated line hardening levels and active de-icing device layout to minimize investment costs. The second stage (operation) minimizes operational costs and tail risks by coordinating generation redispatch, hierarchical load shedding, and active de-icing/repair resources. Case studies on a modified IEEE RTS-96 system demonstrate that the proposed targeted investment strategy saves approximately 33% of capital costs compared to a uniform maximum-hardening scheme. Furthermore, the multi-resource collaborative operation strategy reduces total operational costs by 74.9% under extreme scenarios. The research findings validate that this approach can effectively enhance the economic efficiency and resilience of transmission grids.

Key words: extreme ice disaster, grid resilience, economic assessment, continuous-time Markov chain, two-stage collaborative optimization