输电线路覆冰预测研究

发布时间：2019-03-04 14:41

【摘要】：架空输电线路是电网中传输电能的重要组成部分,必须确保输电线路安全稳定的运行,然而近年来国内外电力行业中架空输电线路多次遭到严重的冰灾事故,给电网的安全运行带来了危害,严重的会造成大面积的停电事故并且使电网恢复困难。因此对输电线路的覆冰情况进行预测,能及时了解导线覆冰的趋势并给出预警,能有效防止严重的覆冰事故,给输电导线防冰、除冰提供依据。本文首先阐述了输电线路产生覆冰的条件、覆冰的机理和过程,对输电线路覆冰进行了分类,综合分析了影响导线覆冰的各种因素,同时阐述了导线防冰、除冰的常用措施。利用输电线路的影响因素对输电线路进行覆冰预测,将输电线路影响因素的历史数据作为预测模型的输入量,首先用RBF神经网络和支持向量机(SVM)模型分别对输电线路覆冰进行了预测,结果显示支持向量机的预测精度高于RBF神经网络。支持向量机中的惩罚参数和核参数对算法精度影响很大,采用遗传算法(GA)和粒子群算法(PSO)对惩罚参数和核参数进行优化,同时遗传算法中的交叉概率和变异概率影响寻优能力,粒子群算法中的惯性因子和加速因子影响寻优能力,对两算法中的参数进行改进,利用改进后的算法对SVM中惩罚参数和核参数进行优化,利用优化后的SVM对导线覆冰进行预测,结果表明优化后的预测时间都增加了,预测精度也提高了。其中基于改进遗传算法的优化时间相对最少,而基于遗传粒子群结合算法的预测时间最长,基于改进遗传算法和改进粒子群算法优化后的预测精度大体相同,基于遗传粒子群结合算法优化后的预测精度最高。从线路覆冰预测的结果可以看出,支持向量机在历史数据样本较小时预测能力比神经网络要好,更适合处理小样本问题,同时利用优化措施对支持向量机进行优化后,增加了预测算法复杂度,提高了覆冰厚度预测的精度。
[Abstract]:Overhead transmission line is an important part of electric energy transmission in power grid. It is necessary to ensure the safe and stable operation of transmission line. However, in recent years, the overhead transmission line in the domestic and foreign power industry has suffered many serious ice disaster accidents. It has brought harm to the safe operation of the power grid, which will cause large-scale blackout and make the power grid difficult to recover. In order to predict the ice-coating situation of transmission lines, the trend of wire ice-coating can be understood in time and early-warning can be given, which can effectively prevent serious ice-coating accidents and provide the basis for anti-icing and de-icing of transmission lines. In this paper, the conditions, mechanism and process of ice-coating on transmission lines are described firstly, the ice-coating of transmission lines is classified, and the factors that affect the ice-coating of transmission lines are comprehensively analyzed. At the same time, the commonly used measures of anti-icing and de-icing of conductors are also expounded. The influence factors of the transmission line are used to predict the icing of the transmission line, and the historical data of the influencing factors of the transmission line are taken as the input of the prediction model. Firstly, RBF neural network and support vector machine (SVM) model are used to predict icing of transmission lines. The results show that the prediction accuracy of support vector machine is higher than that of RBF neural network. The penalty parameters and kernel parameters in support vector machine have great influence on the accuracy of the algorithm. Genetic algorithm (GA) and particle swarm optimization (PSO) are used to optimize the penalty parameters and kernel parameters. At the same time, the crossover probability and mutation probability of genetic algorithm affect the optimization ability, the inertia factor and acceleration factor of particle swarm optimization algorithm affect the optimization ability, and the parameters of the two algorithms are improved. The improved algorithm is used to optimize the penalty parameters and kernel parameters in the SVM, and the optimized SVM is used to predict the wire icing. The results show that the prediction time and the prediction precision are increased after the optimization. The optimization time based on the improved genetic algorithm is the least, and the prediction time based on the genetic particle swarm optimization algorithm is the longest, and the prediction precision of the improved genetic algorithm and the improved particle swarm optimization algorithm is about the same as that of the improved genetic algorithm and the improved particle swarm optimization algorithm. The prediction accuracy based on genetic particle swarm optimization algorithm is the highest. From the result of line icing prediction, it can be seen that support vector machine has better prediction ability than neural network in small historical data sample, and is more suitable to deal with small sample problem. At the same time, support vector machine is optimized by using optimization measures. The complexity of prediction algorithm is increased and the accuracy of ice thickness prediction is improved.
【学位授予单位】：山西大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM752

【相似文献】