With the intensification of global warming

the extreme generation scenarios of renewable energy occur frequently under the influence of medium-range weather processes

which brings great security risks to the power system. Aiming at the shortcomings of controllable generation of renewable energy scenarios under medium-range weather processes

this paper proposes a multi-day wind power extreme scenario generation framework based on conditional diffusion model. Two conditional diffusion models are used to describe the correlation between meteorological and wind power sequential output. A progressive generation model is then constructed

which translates meteorological conditions into daily output characteristic sequences

and further generates multi-day wind power scenarios. Based on the first conditional diffusion model

the framework proposes a daily output sequence feature threshold identification method for multi-day wind power extreme generation events

and then realizes the generation of multi-day wind power extreme scenarios through the second conditional diffusion model. The proposed model has a more stable training process. The effectiveness of the proposed generation framework is demonstrated using climate model data from Xinjiang

China in the global climate model (GCM). This framework achieves the controllable generation of extreme wind power scenarios under the influence of medium-range weather processes.