摘要(Abstract)：
针对传统电弧故障检测方法依赖人为设定阈值、存在保护误动作以及现有人工智能方法需要预先提取特征、计算量大的问题,提出一种基于改进AlexNet模型的串联型电弧故障识别方法;该方法直接采用原始电流波形作为模型输入,避免数据特征预处理;利用Inception结构对AlexNet模型结构进行改进,减少网络参数,并采用随机梯度下降算法与学习率自适应调整方法对模型训练策略进行优化,分别利用已知负载与未知负载对所提方法进行试验验证。结果表明,该方法电弧故障识别准确率达到97.5%以上。
关键词(KeyWords)： 电弧故障;改进AlexNet模型;随机梯度下降;故障识别
基金项目(Foundation): 国家自然科学基金项目(51607055);; 河北省自然科学基金项目(E2020202204);; 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)开放基金项目(EERI_PI2020002);; 特种电机与高压电器教育部重点实验室(沈阳工业大学)开放基金项目(KFKT202003);; 浙江省重点研发计划项目(2020C03103);; 浙江省基础公益研究计划项目(LGG20E070002)
作者(Author): 朱晨,王尧,谢振华,班云升,傅炳,田明
DOI: 10.13349/j.cnki.jdxbn.20210621.001
参考文献(References)：
[1] 酆昆.建筑群内电气线路引起火灾常见情况分析[J].科技创新与应用，2019(20):42-43.
[2] 李杰，曾叙砜，孙领.上海市火灾的统计规律及原因分析[J].安全与环境学报，2019,19 (5):1853-1859.
[3] MEDORA N K,KUSKO A.Arcing faults in low and medium voltage electrical systems:why do they persist[C]//2011 IEEE Symposium on Product Compliance Engineering Proceedings,October 10-12,2011,San Diego,USA.New York:IEEE,2011:1-6.
[4] MENDLER R,HASAN R,TROLLE J B,et al.Hazardous-rated electrical equipment and the arc-flash hazard[J].IEEE Transactions on Industry Applications,2015,51(5):4335-4341.
[5] CRNKO T,DYRNES S.Arcing fault hazards and safety suggestions for design and maintenance[J].IEEE Industry Applications Magazine,2001,7(3):23-32.
[6] TISSERAND E,LEZAMA J,SCHWEITZER P,et al.Series arcing detection by algebraic derivative of the current[J].Electric Power Systems Research,2015,119:91-99.
[7] ARTALE G,CATALIOTTI A,COSENTINO V,et al.Arc fault detection method based on CZT low-frequency harmonic current analysis[J].IEEE Transactions on Instrumentation and Measurement,2017,66(5):888-896.
[8] 刘晓明，徐叶飞，刘婷，等.基于电流信号短时过零率的电弧故障检测[J].电工技术学报，2015,30(13) :125-133.
[9] 王尧，韦强强，葛磊蛟，等.基于电弧电流高频分量的串联交流电弧故障检测方法[J].电力自动化设备，2017,37(7):191-197.
[10] 马少华，鲍洁秋，蔡志远，等.基于信息维数和零休时间的电弧故障识别方法[J].中国电机工程学报，2016,36(9):2572-2579.
[11] YANG K,CHU R B,ZHANG R C,et al.A novel methodology for series arc fault detection by temporal domain visualization and convolutional neural network[J].Sensors,2019,20(1):162.
[12] 苏晶晶，许志红.基于EMD和PNN的故障电弧多变量判据诊断方法[J].电力自动化设备，2019,39(4):106-113.
[13] 郭凤仪，李坤，陈昌垦，等.基于小波近似熵的串联电弧故障识别方法[J].电工技术学报，2016,31(24):164-172.
[14] LIU Y W,WU C J,WANG Y C.Detection of serial arc fault on low-voltage indoor power lines by using radial basis function neural network[J].International Journal of Electrical Power and Energy Systems,2016,83:149-157.
[15] SIEGEL J E,PRATT S,SUN Y B,et al.Real-time deep neural networks for internet-enabled arc-fault detection[J].Engineering Applications of Artificial Intelligence,2018,74:35-42.
[16] WANG Y K,ZHANG F,ZHANG X H,et al.Series AC arc fault detection method based on hybrid time and frequency analysis and fully connected neural network[J].IEEE Transactions on Industrial Informatics,2018,15(12):6210-6219.
[17] 周志炜，许志红.一种新型碳化路径试验装置[J].电器与能效管理技术，2015(21):60-65,70.
[18] LIU W J,ZHANG X B,JI R P,et al.Arc fault detection for AC SSPC in MEA with HHT and ANN[C]//2016 IEEE International Conference on Aircraft Utility Systems (AUS),Octorber 10-12,2016,Beijing,China.New York:IEEE,2016:7-12.
[19] WANG Y K,ZHANG F,ZHANG S W.A new methodology for identifying arc fault by sparse representation and neural network[J].IEEE Transactions on Instrumentation and Measurement,2018,67(11):2526-2537.
[20] SIEGEL J E,PRATT S,SUN Y B,et al.Real-time deep neural networks for internet-enabled arc-fault detection[J].Engineering Applications of Artificial Intelligence,2018,74:35-42.
[21] LEE S,KANG S,KIM T,et al.Improving the performance of serial arc detection using VMD and deep neural network[C]//2019 IEEE 62nd International Midwest Symposium on Circuits and Systems (MWSCAS),August 4-7,2019,Dallas,USA.New York:IEEE,2019:614-617.
