卢明,赵书杰,刘振声,杨晓辉,李哲,宋礼斌.基于灰色投影优化随机森林算法的输电线路舞动预警方法[J].电测与仪表,2020,57(9):45-51,57. LU Ming,ZHAO Shujie,LIU Zhensheng,YANG Xiaohui,LI Zhe,SONG Libin.Early warning method of transmission line galloping based on random forest optimized by grey relation projection[J].Electrical Measurement & Instrumentation,2020,57(9):45-51,57.
基于灰色投影优化随机森林算法的输电线路舞动预警方法
Early warning method of transmission line galloping based on random forest optimized by grey relation projection
The transmission line galloping poses a great challenge to the safe and stable operation of power grids. In order to achieve accurate early warning of transmission line galloping, an early warning method based on random forest optimized by grey relation projection was proposed in this paper. The input values of early warning method include internal factors (conductor splitting number, diameter, spacing, etc.) and external factors (wind speed, wind direction angle, humidity, etc.). In addition, in view of the difficulties such as fewer gallop-ing samples, great differences in galloping terrain and difficult to evaluate objectively, this paper proposed a weighted grey relational projection method in order to optimize the selection of historical data which is similar to the predicted terrain. The historical data of transmission line galloping was used to verify the effectiveness of the proposed model. The results show that the early warning results of this model have obvious advantages in accuracy and false alarm rate when comparing with traditional random forest algorithm and BP neural network. This method can provide a new solution for the early warning of transmission lines galloping.