基于滑动离散傅里叶变换的串联直流电弧故障识别

发布时间：2018-02-07 15:03

本文关键词： 串联直流电弧电弧故障故障识别滑动离散傅里叶变换　出处：《电工技术学报》2017年19期 　论文类型：期刊论文

【摘要】：在直流供电系统中,串联直流电弧故障因难于被检测而成为威胁系统安全运行的最主要因素。针对现有故障电弧识别方法运算较为复杂、实时性相对较差的不足,提出一种基于滑动离散傅里叶变换(DFT)的串联直流电弧故障识别方法。时域与频域分析表明,燃弧起始阶段是直流电弧故障识别的最佳时期,根据该阶段电弧电流的频谱特征确定滑动DFT分析频率点为40 k Hz、80 k Hz和100 k Hz。综合考虑算法分辨率与实时性要求,采用200μs时间窗口对电弧电流进行滑动DFT分析,并用200μs时间窗口进行滑动平均降噪处理。结果表明,发生电弧故障前后滑动DFT频谱在三个特征频率点均有明显变化,验证了滑动DFT算法用于串联直流电弧故障识别的可行性。
[Abstract]:In the DC power supply system, the series DC arc fault is the main factor threatening the safe operation of the system because it is difficult to detect. A series DC arc fault identification method based on sliding discrete Fourier transform (DFT) is proposed. The time domain and frequency domain analysis show that the starting stage of arc ignition is the best time for DC arc fault identification. According to the spectrum characteristics of arc current in this stage, the frequency points of sliding DFT analysis are determined to be 40 kHz and 100k Hz. Considering the resolution and real-time requirement of the algorithm, the sliding DFT analysis of arc current is carried out by using 200 渭 s time window. The sliding average noise reduction is carried out with 200 渭 s time window. The results show that the sliding DFT spectrum changes obviously at the three characteristic frequency points before and after the arc fault occurs. The feasibility of the sliding DFT algorithm in series DC arc fault identification is verified.
【作者单位】：省部共建电工装备可靠性与智能化国家重点实验室(河北工业大学);天津工业大学电气与自动化学院;天津大学电气自动化与信息工程学院;浙江大学电气工程学院;
【基金】：国家自然科学基金(51607055) 河北省自然科学基金(E2015202143) 河北省教育厅青年科学基金(QN2014148)资助项目
【分类号】：TM501.2

【相似文献】