Under the global climate deterioration scenario

frequent cold wave events and their associated large-scale snow/rain weather conditions increasingly trigger icing failures in critical power grid components

heightening power system outage risks. To enhance power system resilience against cold waves

this paper proposes a two-stage robust preventive scheduling method incorporating source-grid failure probability information. By integrating icing characteristics of generation and grid components under cold wave conditions

critical component failure models are established as follows: a hidden Markov model (HMM) quantifies wind turbine outage probabilities on the generation side

while a dynamic ice accretion-melting process model evaluates transmission line failures on the grid side

considering thermal ice-melting properties. By utilizing information entropy theory

uncertainty sets are constructed to encapsulate wind turbine outages and transmission line faults. A cost-minimizing two-stage robust preventive scheduling model is developed and optimized through the column-and-constraint generation (C&CG) algorithm to coordinate generation scheduling and contin-gency reserves. Case studies on the modified IEEE-RTS 79 system demonstrate that the proposed probability modeling method can be adopted to effectively quantify renewable energy and transmission line failure risks

reduce the conservatism in uncertainty sets

and lower the preventive scheduling costs compared to conventional approaches.