Objective The analysis of limited icing capacity of wind turbine in cold wave weather is difficult to predict, resulting in inaccurate wind power prediction and insufficient decision-making basis for wind power dispatching.

Method Through the prediction model of the limited icing capacity of wind turbine, the limited icing capacity of wind turbine in extreme cold wave weather process in Guangxi was analyzed and summarized by using conventional meteorological observation data, wind turbine shutdown actual data and numerical model data.

Result The results show that the reference value and accuracy of icing prediction are effectively improved by integrating the numerical prediction products with the actual data of limited icing capacity and applying regression analysis for real-time correction. In addition, the icing prediction model can effectively respond to the strong cold air system southward affecting the Guangxi wind farm, but the response to the turning weather is insufficient, and the prediction result is larger than the actual data. At the same time, the numerical model prediction results have amplitude deviation and phase deviation, and the predicted value is larger than the actual value in this process. In terms of prediction effect, the model performs better in air temperature prediction than relative humidity and wind speed prediction, and the prediction effect of meteorological elements in high altitude areas is generally better than that in low altitude areas.

Conclusion Based on the above conclusions, some suggestions are put forward, such as strengthening the early warning and prediction ability of cold wave, carrying out the upgrading and transformation of icing capacity prediction system, so as to improve the prediction accuracy of the limited icing capacity of wind turbine in extreme cold wave weather.