基于多历史覆冰过程的输电线路覆冰增长预测

Prediction of Icing Growth of Transmission Lines Based on Multi-history Icing Process

应用粒子群优化(PSO)和最小二乘支持向量机(LS-SVM)算法对各历史覆冰过程建立预测模型,利用皮尔逊相关系数法对历史覆冰过程进行相似性筛选,采用径向基神经网络(RBF)建立多历史覆冰过程的覆冰增长率预测模型.实例计算表明,与传统的单历史覆冰过程预测方法相比,基于多历史覆冰过程的输电线路覆冰增长预测具有更好的精度.

Particle swarm optimization (PSO)and least squares support vector machine (LS-SVM)algorithm are applied to predict the icing process,The matching mechanism of similar historical icing process was constructed by Pearson correlation coefficient method.Radial basis function neural network (RBF)is applied to establish a prediction model of icing growth rate in multi period icing process to obtain prediction data of icing thickness.A case study shows that the prediction method of icing growth on transmission line based on multi history icing process has better accuracy compared with the traditional single history icing process prediction method.