面向不间断供电(UPS)系统的电能质量分析技术  

Power quality analysis technology for Uninterruptible Power Supply(UPS)system

作　　者：邓卜侨 谢岫峰 纪明阳 艾青 王康 DENG Buqiao;XIE Xiufeng;JI Mingyang;AI Qing;WANG Kang(Beijing FibrLink Communications Co.,Ltd.,Beijing 100070,China)

机构地区：[1]北京中电飞华通信有限公司,北京100070

出　　处：《电子设计工程》2025年第1期12-16,共5页Electronic Design Engineering

基　　金：北京中电飞华通信有限公司科技项目(52680021N01H)。

摘　　要：针对现行UPS电能质量检测过程存在的准确率低、实时性差且成本高的缺点,文中基于VMD-SAE-1DCNN模型提出了一种UPS电能质量检测与识别算法。对于电能信号非线性与非平稳的特点,使用变分模态算法对原信号进行分解,从而得到本征模态信号。同时,采用稀疏自编码器对本征模态信号进行特征提取,通过建立多层一维卷积神经网络模型对特征进行训练,提升了运算效率。实验测试结果表明,所提算法的迭代次数与运行时间在对比算法中均为最优,分类准确率可达97%以上,充分证明了改进算法的有效性。In response to the shortcomings of low accuracy,poor real⁃time performance,and high cost in the current UPS power quality detection process,a UPS power quality detection and recognition algorithm is proposed based on the VMD⁃SAE⁃1DCNN model.For the nonlinear and non⁃stationary characteristics of electric energy signals,the variational modal algorithm is used to decompose the original signal to obtain the intrinsic modal signal.At the same time,the sparse Auto Encoder is used to extract the characteristics of the intrinsic mode signal,and the multi⁃layer one⁃dimensional Convolutional Neural Network model is established to train the characteristics,which improves the operation efficiency.The experimental test results show that the proposed algorithm has the best iteration times and running time in the comparison algorithm,with a classification accuracy of over 97%,fully proving the effectiveness of the proposed algorithm improvement.

关 键 词：UPS 变分模态算法 稀疏自编码器 卷积神经网络 电能质量分析 

分 类 号：TN99[电子电信—信号与信息处理] TP391[电子电信—信息与通信工程]