•  
  •  
 

Keywords

transmission line; micrometeorology; icing prediction; support vector regression machine; improvedwhale optimization algorithm; small sample

Abstract

The transmission lines in the micrometeorological area are more prone to icing, so it is extremely destructive to the safe operation of the power grid system. In view of the characteristics that icing monitoring data in micrometeorological areas is scarce, and interference is strong, RF-GSWOA-SVRM, a prediction method for transmission line icing in micrometeorological areas based on random forest (RF), global search whale optimization algorithm (GSWOA), and support vector regression machine (SVRM), is proposed to improve the accuracy of icing prediction. Firstly, the RF algorithm is used to extract the correlation between transmission line icing and micrometeorological data, thereby reducing the overfitting phenomenon caused by a single meteorological factor and the superposition effect of multiple meteorological factors. Then, to address the issue that the SVRM algorithm is highly sensitive to the selection of the kernel function and the setting of the penalty factor, the traditional whale algorithm is optimized to obtain GSWOA, thereby avoiding the kernel function and penalty factor from falling into local optimal solutions. Furthermore, the two parameters of the SVRM algorithm are optimized via GSWOA, and a short-term icing prediction model based on RF-GSWOA-SVRM is established. Finally, by taking the online monitoring data of transmission lines in a single micrometeorological area of Henan power grid as an example, a comparative analysis is conducted to verify the effectiveness of the proposed method. This model is applied to the transmission line icing prediction in similar micrometeorological areas of a certain region, and high prediction accuracy is achieved, demonstrating that the model has certain general applicability.

DOI

10.19781/j.issn.1673-9140.2026.01.004

First Page

36

Last Page

45

Download

Share

COinS