基于深度学习的继电器寿命预测方法研究综述

doi:10.16628/j.cnki.2095-8188.2021.12.001

摘要/Abstract

摘要：

近年来,随着科技的不断发展,继电器等电器元器件的剩余使用寿命逐渐成为了研究的热点,准确、高效地分析海量的监测数据是故障检测与健康管理的主要任务。由于传统方法受限于依靠物理模型与先验知识,深度学习方法应运而生。首先介绍了继电器剩余使用寿命的国内外研究现状,其次详细分析了基于深度学习的剩余寿命预测方法,总结了各个方法的优缺点,最后对未来进行了展望。

关键词: 深度学习, 寿命预测, 神经网络, 故障预测与健康管理(PHM)

Abstract:

In recent years,with the continuous development of science and technology,the remaining service life of relay and other electrical components has gradually become a research hotspot.The accurate and efficient analysis of massive monitoring data is the main task of fault detection and health management.Because the traditional methods are limited to rely on the physical models and prior knowledge,the deep learning methods emerge as the times require.First,the current research status of the remaining life of relays at home and abroad is introduced.Second,it analyzes the remaining life prediction method based on deep learning in detail is analyszed,and the advantages and disadvantages of each method are summarized.Finally,the future is looked forward to.

Key words: deep learning, life prediction, neural network, PHM

中图分类号:

TM58

刘百鑫, 王召斌, 乔青云, 朱佳淼, 李朕. 基于深度学习的继电器寿命预测方法研究综述[J]. 电器与能效管理技术, 2021, 0(12): 1-6.

LIU Baixin, WANG Zhaobin, QIAO Qingyun, ZHU Jiamiao, LI Zhen. Review of Relay Life Prediction Methods Based on Deep Learning[J]. LOW VOLTAGE APPARATUS, 2021, 0(12): 1-6.

图/表 4

参考文献 45

[1]	孙强, 岳继光. 基于不确定性的故障预测方法综述[J]. 控制与决策, 2014, 29(5):769-778.
[2]	ZHANG S J, KANG R, LIN Y H. Remaining useful life prediction for degradation with recovery phenomenon based on uncertain process[J/OL]. https://ideas.repec.org/a/eee/reensy/v208y2021ics0951832021000119.html .
[3]	年夫顺. 关于故障预测与健康管理技术的几点认识[J]. 仪器仪表学报, 2018, 39(8):1-14.
[4]	刘惠, 刘振宇, 郏维强, 等. 基于深度学习的剩余使用寿命预测技术研究现状与挑战[J/OL]. 计算机集成制造系统:1-31[2021-12-15]. http://kns.cnki.net/kcms/detail/11.5946.TP.20201202.1751.015.html .
[5]	JARDINE A K S, LIN D, BANJEVIC D. A review on machinery diagnostics and prognostics implementing condition-based maintenance[J]. Mechanical Systems and Signal Processing, 2005, 20(7):1483-1510. doi: 10.1016/j.ymssp.2005.09.012
[6]	王召斌, 尚尚, 翟国富. 航天继电器贮存可靠性退化试验与评价方法研究[J]. 电器与能效管理技术, 2015(8):1-7.
[7]	刘兰强. 电磁继电器在汽车电器保护电路中的应用[J]. 科技资讯, 2018, 16(28):73-74.
[8]	王佳炜, 王召斌, 黄周霖. 继电器寿命预测方法综述[J]. 电器与能效管理技术, 2018(4):1-5,12.
[9]	VICHARE N M, PECHT M G. Prognostics and health management of electronics[J]. IEEE Transactions on Components and Packaging Technologies, 2006, 29(1):222-229. doi: 10.1109/TCAPT.2006.870387
[10]	孙训俊. 电器产品贮存寿命分析方法的研究[D]. 天津:河北工业大学, 2014.
[11]	王召斌, 蔡昭文, 翟国富, 等. 航天继电器加速贮存退化可靠性建模和统计分析[J]. 低压电器, 2012(17):1-7.
[12]	季文强. 基于深度学习和不确定性量化的数据驱动剩余寿命预测方法研究[D]. 合肥:中国科学技术大学, 2020.
[13]	HHUANG Z, XU Z, WANG W, et al. Remaining useful life prediction for a nonlinear heterogeneous Wiener process model with an adaptive drift[J]. IEEE Transactions on Reliability, 2015, 64(2):687-700. doi: 10.1109/TR.2015.2403433
[14]	PANDEY M D, YUAN X X, NOORTWIJK J M. The influence of temporal uncertainty of deterioration on life-cycle management of structures[J]. Structure and Infrastructure Engineering, 2009, 5(2):145-156. doi: 10.1080/15732470601012154
[15]	KHAROUFEH J P. Explicit results for wear processes in a Markovian environment[J]. Operations Research Letters, 2003, 31(3):237-244. doi: 10.1016/S0167-6377(02)00229-8
[16]	GEBRAEEL N Z, LAWLEY M A, LI RONG, et al. Residual-life distributions from component degradation signals:A Bayesian approach[J]. IIE Transactions, 2005, 37(6):543-557. doi: 10.1080/07408170590929018
[17]	张菲菲, 李志刚. 基于BP神经网络的继电器剩余寿命预测[J]. 低压电器, 2012(1):11-14.
[18]	李华, 孙东旺, 贺鹏举, 等. 基于超程时间回归模型的继电器寿命预测方法[J]. 电工技术学报, 2013, 28(S2):414-417,423.
[19]	王婷云, 李志刚, 左帅, 等. 基于Wiener过程的电子器件退化建模及剩余寿命预测[C]// 第五届电工产品可靠性与电接触国际会议论文集; 2014:391-395.
[20]	刘小勇. 基于深度学习的机械设备退化状态建模及剩余寿命预测研究[D]. 哈尔滨:哈尔滨工业大学, 2018.
[21]	LIU Y, ZHANG G, ZHAO C, et al. Mechanical condition identification and prediction of spring operating mechanism of high voltage circuit breaker[J]. IEEE Access, 2020, 8:210328-210338. doi: 10.1109/Access.6287639
[22]	张蕾, 章毅. 大数据分析的无限深度神经网络方法[J]. 计算机研究与发展, 2016, 53(1):68-79.
[23]	XIA M, LI T, LIU L, et al. Remaining useful life prediction of rotating machinery using hierarchical deep neural network[C]// 2017 IEEE International Conference on Systems,Man,and Cybernetics (SMC), 2017:2778-2783.
[24]	王海龙, 夏筱筠, 孙维堂. 基于EMD与卷积神经网络的滚动轴承故障诊断[J]. 组合机床与自动化加工技术, 2019(10):46-48,52.
[25]	BABU G S, ZHAO P, LI X L. Deep convolutional neural network based regression approach for estimation of remaining useful life[J]. Database Systems for Advanced Applications, 2016, 9642:214-228.
[26]	谢勇, 韦震, 邓力明, 等. 电磁继电器超行程自动测试系统的研究[C]// 中国航天电子技术研究院科学技术委员会2020年学术年会论文集, 2020: 7.
[27]	崔和臣. 交流接触器电寿命预测的深度学习模型及实验[D]. 温州:温州大学, 2019.
[28]	CUI H, WU Z, WU G, et al. Convolutional neural networks for electrical endurance prediction of alternating current contactors[J]. IEEE Transactions on Components,Packaging and Manufacturing Technology, 2019, 9(9):1785-1793. doi: 10.1109/TCPMT.5503870
[29]	李俊峰, 汪月乐, 胡升, 等. 基于DBN,SVM和BP神经网络的光谱分类比较[J]. 光谱学与光谱分析, 2016, 36(10):3261-3264.
[30]	赵光权, 葛强强, 刘小勇, 等. 基于DBN的故障特征提取及诊断方法研究[J]. 仪器仪表学报, 2016, 37(9):1946-1953.
[31]	张国辉. 基于深度置信网络的时间序列预测方法及其应用研究[D]. 哈尔滨:哈尔滨工业大学, 2017.
[32]	葛强强. 基于深度置信网络的数据驱动故障诊断方法研究[D]. 哈尔滨:哈尔滨工业大学, 2016.
[33]	ZHANG C, LIM P, QIN A K, et al. Multiobjective deep belief networks ensemble for remaining useful life estimation in prognostics[J]. IEEE Transactions on Neural Networks and Learning Systems, 2017, 28(10):2306-2318. doi: 10.1109/TNNLS.2016.2582798
[34]	JIAO R, PENG K, DONG J, et al. A health indicator construction method based on deep belief network for remaining useful life prediction[C]// 2019 Prognostics and System Health Management Conference (PHM-Qingdao), 2019:1-6.
[35]	张国强. 基于DBN的汽车电磁继电器寿命预测方法研究[D]. 哈尔滨:东北林业大学, 2019.
[36]	李忠志. 基于改进DBN和Softmax回归的电磁继电器寿命预测方法研究[D]. 哈尔滨:东北林业大学, 2020.
[37]	韩春阳. 银基触点材料电性能模拟实验与寿命预测方法研究[D]. 镇江:江苏科技大学, 2020.
[38]	Naseer S, YASIR S, SHEHZAD K, et al. Enhanced network anomaly detection based on deep neural networks[J]. IEEE Access, 2018, 6:48231-48246. doi: 10.1109/ACCESS.2018.2863036
[39]	邴绍强, 王振, 段鸿杰, 等. 基于循环神经网络的抽油杆柱寿命预测新方法[J]. 电脑知识与技术, 2019, 15(35):178-182,187.
[40]	张子贤, 李敏, 苗红霞, 等. 基于RNN的故障预测算法及在GIS上的应用[J]. 计算机测量与控制, 2020, 28(12):27-31,36.
[41]	HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8):1735-1780. doi: 10.1162/neco.1997.9.8.1735
[42]	欧明阳, 杨代军, 张存满. 基于循环神经网络的PEMFC寿命预测[J]. 电池, 2020, 50(2):123-126.
[43]	WEI F, WAN Z, HE H, et al. Ultrafast active response strategy against malfunction attack on fault current limiter[J]. IEEE Transactions on Smart Grid, 2020, 11(3):2722-2733. doi: 10.1109/TSG.5165411
[44]	GUAN X, WEN Y, DONG Z, et al. Deterioration behavior analysis and LSTM-based failure prediction of GIB electrical contact inside various insulation gases[J]. IEEE Access, 2020, 8:152367-152376. doi: 10.1109/Access.6287639
[45]	MA M, MAO Z. Deep-convolution-based LSTM network for remaining useful life prediction[J]. IEEE Transactions on Industrial Informatics, 2021, 17(3):1658-1667. doi: 10.1109/TII.9424

方法	优点	缺点
DNN	模型简单	易产生数据过拟合,计算量大
CNN	可以处理高维数据	数据降维时易造成信息丢失
DBN	反向微调	难以预测长时间序列
RNN	可以利用历史信息	易产生梯度消失