基于核支持向量机的电力系统暂态稳定评估模型

Transient Stability Evaluation Model for Power System Based onKernel Support Vector Machine

随着计算机技术的发展,诸多领域开始向智能化方向转型,空间数据智能技术的应用日益受到关注。电力系统是整个电网运行的重要组成部分,也是电子元器件和电子设备的工作空间,如何对电力系统的稳定性进行评估是保证其稳定运行的关键。针对电力系统的暂态评估问题,融合空间数据智能技术,对样本数据进行数据采集和特征提取模型的构建。利用支持向量机(Support Vector Machine, SVM)算法提高电力系统性能,引入核函数和马氏距离对SVM算法进行优化,建立了基于核SVM(Kernel SVM,KSVM)的电力系统暂态评估模型。在电力系统数据集上进行实验,结果表明所提模型KSVM精确率为95.62%,比卷积神经网络算法高11.36%。

With the development of computer technology,several fields begin to transform to the direction of intelligence,and the application of spatial data intelligence technology is increasingly concerned.The power system is an important part of the whole power grid operation,and it is the working space of electronic components and electronic equipment as well.How to evaluate the stability of power systems is the key point to ensure its stable operation.Aiming to handle the transient evaluation problem of power systems,Data samples are collected and a feature extraction model is constructed,through integrating spatial data intelligence technology into it.Support Vector Machine(SVM)algorithm is used to improves the performance of power systems,and kernel function and Mahalanobis distance are used to optimize the SVM algorithm.A transient evaluation model of power systems based on Kernel Support Vector Machine(KSVM)is established.Experiments are conducted in the power system data sets,and the results show that the accuracy of the proposed model KSVM is 95.62%,which is 11.36%higher than that of the convolutional neural network algorithm.