摘要
针对智能变电站中数据体量大、种类多、速度快的特点,对智能变电站中的数据分类方法进行了研究,提出了基于安全指标和遗传模拟退火支持向量机的两级分类方法。首先,构建了智能变电站安全指标分类规则库,使用其对变电站数据进行初次粗糙分类,缩小数据规模;其次,依据智能变电站故障隐患数据样本,使用支持向量机训练出二类分类器,并采用遗传算法和模拟退火算法对其性能进行优化,完成智能变电站数据的二次分类,得到的正类数据为正常数据,负类数据即为需要重点进行下一步分析的异常数据。实验表明,该方法在智能变电站数据分类上取得了良好的效果,并且能够有效地控制数据的规模。
Due to grid data had the natures of high volume, multiple types and high speed, data classification method of intelligent substation had been studied. Based on this, the two-level analysis method based on safety indicators and genetic simulated annealing support vector machine had been proposed. Firstly, constructed a rule base of safety indicators for intelligent substation to reduce the data size roughly;Secondly, used support vector machine model to train the fault hidden data samples which produced in intelligent substation, and optimized the model with genetic algorithms and simulated annealing algorithm. The positive class data as normal data, while the negative class data was abnormal data need to focus on for further analysis. For this reason, it is appropriate for application to reduce the data size of intelligent substation.
出处
《东北电力大学学报》
2014年第2期61-65,共5页
Journal of Northeast Electric Power University
基金
吉林省科技发展计划(20120338)
关键词
智能变电站
数据分类
安全指标
支持向量机
Intelligent substation
Data classification
Safety indicators
Support vector machines
