计及气象特征的线路覆冰荷载预测OA

To address the frequent occurrence of ice-induced disasters on overhead transmission lines and the limita-tions of conventional icing prediction methods,this paper proposes a novel ice load prediction approach that system-atically incorporates meteorological characteristics.First,based on the load patterns of transmission lines,meteoro-logical conditions are classified into three distinct regime sets:load growth,equilibrium,and decay.The grey rela-tional analysis is employed to integrate the weighted contributions of meteorological parameters into the Mahalanobis distance.Second,to resolve the fuzzy membership relationships between meteorological regimes,two physically in-terpretable indices are empirically defined:the ice growth potential(IGP)factor and ice melting potential(IMP)factor,which quantitatively characterize meteorological characteristics.Finally,membership degrees are combined with monitoring data to construct augmented historical samples embedding meteorological characteristics,and a sup-port vector machine(SVM)is applied for ice load prediction.Case studies compare three training paradigms:me-teorological parameters only,fuzzy classification of meteorological processes,and the proposed method.Results demonstrate that the proposed model achieves higher prediction accuracy.