基于S变换和神经网络的电能质量多扰动分类识别  

Classification and recognition of power quality multi-disturbance based on S-transform and neural networks

作　　者：谷湘文[1] 高培生[1] 吴为麟[1] 

机构地区：[1]浙江大学电气工程学院,浙江杭州310027

出　　处：《机电工程》2007年第9期46-49,共4页Journal of Mechanical & Electrical Engineering

摘　　要：主要针对谐波扰动与其他扰动共存的情况,提出了一种识别电能质量多扰动共存下的多扰动分类方法。它利用S变换计算基频的幅值信息,从而获得幅值特征向量,再利用傅立叶变换得到频谱的特征向量,然后根据这两组特征向量训练并行的神经网络,得到并行的分类器;它既可以对单种扰动进行分类,也可以对多扰动进行分类。通过仿真试验,验证了此分类器是有效的,可行的。　A classification method which can handle multi-disturbance was presented.Using Fourier transform and S transform,the characteristic vectors of power quality disturbances were obtained.The vectors can be divided to amplitude vectors and spectrum vectors.Based on these vectors,parallel BP neural networks can be trained.The simulation results confirm that this classifier can recognize single disturbance as well as multi-disturbance.

关 键 词：电能质量 多扰动 S变换 傅立叶变换 神经网络 

分 类 号：TM711[电气工程—电力系统及自动化]