基于聚类算法的电网覆冰灾害分级预警模型研究

Study on Clustering Algorithm-based Icing Disasters Classification Early Warning Model for Power Grid

对电网覆冰灾害进行了分级预警模型的研究,采用图像检测法测得输电线路覆冰厚度的实时数据,以此为约束条件,采用聚类算法计算出输电线路的综合覆冰厚度,针对不同的覆冰厚度,对覆冰灾害进行了等级划分,并发出相应等级的预警信号。通过对某220 k V架空输电线路覆冰厚度数据的实时监测,验证了该预警模型的有效性。电网覆冰灾害分级预警模型的建立,使覆冰灾害预警更为准确、有效,且节约线路维护检修成本,为高压输电线路状态评估、安全性评价等工作提供参考依据。

In this paper, a graded early warning model on icing in power grid has been studied. Firstly, image detection method has been used to collect real-time image data of icing thickness on transmission lines. Secondly, based on those image data and via clustering algorithm, icing thickness could be calculated. Finally, according to different icing thickness, icing disasters could be classified into different levels with corresponding warning signals. The validation of this model has been tested via the real-time image data of icing on a 220 kV transmission line. The establishment of the model could help to predict icing disasters in a more appropriate, effective and cost-saving way, and to premisely evaluate the condition and safety of high-voltage transmission lines′ operation.