摘要
提出了一种能够提高负荷预测精度的方法。在对历史电网运行数据进行处理时引入大数据挖掘技术,并且使用了支技向量机来进行电力系统短期负荷的预测新方式,对基于粒子群优化的支持向量算法进行了改进,提出基于相似日聚类的支持向量机的方法以对电网的负荷状态进行预测。以湘潭市的电力负荷数据为测试数据,进行两种算法结果的对比。结果表明:本文的算法在对比中具有较大的优势,数据预处理在预测的精度上有着非常重要的关联。
Based on the above algorithm, this paper proposes a method to improve the load forecasting precision. Is crucial to improve the precision of the prediction model of improvement, in the history of power grid operation data for processing large data mining technology is introduced, and using the supported vector machine ( SVM ) for electric power system short-term load forecast new way, in this paper, the support vector based on particle swarm optimization algorithm is improved, based on the similar day clustering support vector machine ( SVM) method to predict the load status of power grid. In order to test the performance of the algorithm in this paper, the results of two algorithms based on the power load data of xiangtan are given at the end of the text. The algorithm in this paper has a great advantage in comparison, and the results show that the data preprocessing of this paper has a very important relation with the accuracy of prediction.
作者
程子华
Cheng Zihua(Guangzhou Electromechanical Technician College, Guangzhou 510435 ,China)
出处
《科技通报》
2019年第5期67-70,共4页
Bulletin of Science and Technology
关键词
支持向量机
电力负载
预测算法
相似日
support vector machine
power load
prediction algorithm
similar day