面向无人机巡检场景基于 QRGBDT 的架空输电线路弧垂区间预测

Power Line Sag Interval Prediction Basedon QRGBDT for UAV InspectionScenario

随着无人机电力巡检技术的发展,无人机巡检逐渐替代了人工巡检,成为输电线路监测和维护中不可或缺的工具。弧垂作为输电线路安全运行的关键指标,准确的弧垂区间预测对于确保线路安全运行和及时维护具有重大意义。针对新型无人机巡检场景,本研究提出了一种基于分位数极限梯度提升树(QRGBDT)的架空输电线路弧垂区间预测方法。通过将分位数回归模型与极限梯度提升树模型相结合,构建了一个既能有效处理异常值、捕捉数据分布多样性,又具有高度拟合能力和优异泛化性能的弧垂区间预测模型。通过实际应用案例以及与 QRCNN 和 QRRNN 模型的对比,证实了本方法的有效性和优越性。

With the development of drone power inspection technology, drone inspection has gradually replaced manual inspection, becoming an indispensable tool in the monitoring and maintenance of transmission lines. As a key indicator of the safe operation of transmission lines, accurate sag interval prediction is of great significance for ensuring the safe operation of lines and timely maintenance. In view of the new drone inspection scenario, this study proposes an overhead transmission line sag interval prediction method based on Quantile Regression Extreme Gradient Boosting Decision Tree (QRGBDT). By combining the quantile regression model with the extreme gradient boosting decision tree model, a sag interval prediction model is constructed that can effectively handle outliers, capture the diversity of data distribution, and hashigh fitting ability and excellent generalization performance. Through actual application cases and comparison with QRCNN and QRRNN models, the effectiveness and superiority of this method are confirmed.