陀螺电机轴承小样本非等间隔的寿命预测研究

doi:10.12382/bgxb.2023.0259

摘要/Abstract

摘要：

现有轴承寿命预测的研究成果直接应用于陀螺电机轴承时,存在轴承振动信号难采集、模型使用小样本、非等间隔数据建模精度低等问题。选取陀螺电机电流信号作为可测信号并制定执行标准,截取有效电信号;提取初值化均方根值和Renyi熵作为退化特征以描述轴承寿命。设计的非等间隔、融合经验模态分解(Empirical Mode Decomposition,EMD)和生物地理学算法(Biogeography-based Optimization,BBO)的灰色模型(Grey Models,GM),即非等间隔EMD-BBO-GM(1,1),其包括间隔变化、数据分解、模型构建和参数优化共4个模块,可实现轴承寿命预测功能。选取小微挠性陀螺电机进行轴承寿命预测试验。研究结果表明,新模型使用两种退化特征获得的预测寿命与实际寿命相当,拟合精度不低于98%,预测精度不低于95%;与标准GM(1,1)相比,预测精度提升量为24.975%,间隔变化、数据分解和参数优化3个模块的贡献度分别为90.94%、3.64%、5.42%。

关键词: 陀螺电机轴承, 寿命预测, 电流信号, 灰色模型, 非等间隔数据

Abstract:

The bearing vibration signals are difficult to be collectedwhen the existing research results of bearing life prediction are directly applied to gyro motor bearings, and the modeling accuracy is low when the model uses small samples and unequally spaced data.The current signal of gyro motor is selected as the measurable signal, and an implementation standard is formulated to intercept the effective electric signal. Initializing root mean square (IRMS) and Renyi entropy are extracted as degradation features to describe the bearing life.The designed EMD-BBO-GM (1,1) model consists of interval change module, data decomposition module, model construction module and parameter optimization module, which can realize the function of bearing life prediction.A small and micro flexible gyro motoris selected for the bearing life prediction test. The results show that the predicted life of the model is equivalent to the actual life, the fitting accuracy is not less than 98%, and the prediction accuracy is not less than 95%. Compared with the standard GM(1,1) model, the prediction accuracy of this model is improved by 24.975%, and the contributions of interval change module, data decomposition module and parameter optimization module are 90.94%, 3.64% and 5.42%, respectively.

Key words: gyro motor bearing, life prediction, current signal, grey model, unequally spaced data

中图分类号:

TH133.33⁺1

蔡曜, 王建青, 司玉辉, 王玉琢, 郭伟, 武展. 陀螺电机轴承小样本非等间隔的寿命预测研究[J]. 兵工学报, 2024, 45(7): 2426-2441.

CAI Yao, WANG Jianqing, SI Yuhui, WANG Yuzhuo, GUO Wei, WU Zhan. Prediction of Service Life of Gyro Motor Bearing with Small Sample and Unequally Spaced Data[J]. Acta Armamentarii, 2024, 45(7): 2426-2441.

图/表 19

参考文献 23

[1]	于军, 刘可, 郭帅, 等. 基于条件深度循环生成对抗网络和动作探索的行星轮轴承剩余寿命预测[J]. 兵工学报, 2020, 41(11):2170-2178.
	YU J, LIU K, GUO S, et al. Remaining useful life prediction of planet bearings based on conditional deep recurrent generative adversarial network and action discovery[J]. Acta Armamentarii, 2020, 41(11):2170-2178. (in Chinese)
[2]	张亚涛, 邱明, 周大威, 等. 基于双应力加速寿命试验的关节轴承寿命预测与可靠性分析[J]. 润滑与密封, 2020, 45(3):51-56. doi: 10.3969/j.issn.0254-0150.2020.03.009
	ZHANG Y T, QIU M, ZHOU D W, et al. Life prediction and reliability analysis of spherical plain bearings based on double stress accelerated life test[J]. Lubrication Engineering, 2020, 45(3):51-56. (in Chinese) doi: 10.3969/j.issn.0254-0150.2020.03.009
[3]	邱明, 牛凯岑, 李军星, 等. 多重应力下滚动轴承剩余寿命预测[J]. 航空动力学报, 2022, 37(5):980-988.
	QIU M, NIU K C, LI J X, et al. Residual life prediction of rolling bearing under multiple stresses[J]. Journal of Aerospace Power, 2022, 37(5):980-988. (in Chinese)
[4]	杨创艳, 马军, 王晓东, 等. 特征融合与灰色回归的滚动轴承性能退化评估[J]. 电子学报, 2022, 50(1):106-115. doi: 10.12263/DZXB.20200826
	YANG C Y, MA J, WANG X D, et al. Feature fusion and grey regression for performance degradation assessment of rolling bearings[J]. Acta Electronica Sinica, 2022, 50(1):106-115. (in Chinese)
[5]	蒋全胜, 许伟洋, 朱俊俊, 等. 基于动态加权卷积长短时记忆网络的滚动轴承剩余寿命预测方法[J]. 振动与冲击, 2022, 41(17):282-291.
	JIANG Q S, XU W Y, ZHU J J, et al. Residual life prediction method of rolling bearing based on DW-CNN-LSTM networks[J]. Journal of Vibration and Shock, 2022, 41(17):282-291. (in Chinese)
[6]	张昌凡, 高见, 刘建华. 基于灰色模型混合算法的滚动轴承剩余使用寿命预测[J]. 机械设计与研究, 2022, 38(4):103-108.
	ZHANG C F, GAO J, LIU J H, et al. Prediction of remaining useful life of rolling bearings based on grey model hybrid algorithm[J]. Machine Design and Research, 2022, 38(4):103-108. (in Chinese)
[7]	苏琪, 王海波, 施晓辰, 等. 基于灰色预测模型的参数寻优方法及能源预测应用[J]. 南昌大学学报(理科版), 2022, 46(3):371-378.
	SU Q, WANG H B, SHI X C, et al. A method and application study of optimizing key parameters of gray prediction model based on emperor butterfly algorithm[J]. Journal of Nanchang University (Natural Science), 2022, 46(3):371-378. (in Chinese)
[8]	姚德臣, 李博阳, 刘恒畅, 等. 基于注意力GRU算法的滚动轴承剩余寿命预测[J]. 振动与冲击, 2021, 40(17):116-123.
	YAO D C, LI B Y, LIU H C, et al. Residual life prediction of rolling bearing based on attention GRU algorithm[J]. Journal of Vibration and Shock, 2021, 40(17):116-123. (in Chinese)
[9]	孙世岩, 张钢, 梁伟阁, 等. 基于时间序列数据扩增和BLSTM的滚动轴承剩余寿命预测方法[J]. 系统工程与电子技术, 2022, 44(3):1060-1068. doi: 10.12305/j.issn.1001-506X.2022.03.40
	SUN S Y, ZHANG G, LIANG W G, et al. Remaining useful life prediction method of rolling bearing based on time series data augmentation and BLSTM[J]. Systems Engineering and Electronics, 2022, 44(3):1060-1068. (in Chinese) doi: 10.12305/j.issn.1001-506X.2022.03.40
[10]	吕明珠. 基于改进LSTM的滚动轴承性能退化趋势预测[J]. 轴承, 2022(4):70-76.
	LÜ M Z. Performance degradation trend prediction of rolling bearing based on improved LSTM[J]. Bearing, 2022(4):70-76. (in Chinese)
[11]	于震梁, 孙志礼, 曹汝男, 等. 基于支持向量机和卡尔曼滤波的机械零件剩余寿命预测模型研究[J]. 兵工学报, 2018, 39(5):991-997. doi: 10.3969/j.issn.1000-1093.2018.05.020
	YU Z L, SUN Z L, CAO R N, et al. Research on remaining useful life predictive model of machine parts based on SVM and kalmanfilter[J]. Acta Armamentarii, 2018, 39(5):991-997. (in Chinese)
[12]	SCHOEN R R, HABETLER T G. Motor bearing damage detection using stator current monitoring[J]. IEEE Transactions on Industry Applications, 1995, 31(6):1274-1279.
[13]	BLODT M, GRANJON P, RAISON B, et al. Models for bearing damage detection in induction motors using stator current monitoring[J]. IEEE Transactions on Industrial Electronics, 2008, 55(4):1813-1822.
[14]	刘小峰, 冯伟, 柏林. 考虑退化轨迹差异性与相似性的轴承RUL预测[J]. 控制与决策, 2021, 36(11):2833-2840.
	LIU X F, FENG W, BO L. Prediction of bearing remaining useful life involving difference and similarity of degradation trajectories[J]. Control and Decision, 2021, 36(11):2833-2840. (in Chinese)
[15]	陈仁祥, 吴昊年, 杨黎霞, 等. 多分类器集成加权均衡分布适配的滚动轴承寿命阶段识别[J]. 仪器仪表学报, 2019, 40(10):66-73.
	CHEN R X, WU H N, YANG L X, et al. Rolling bearing life stage recognition based on multi-classifier integration of the weighted and balanced distribution adaptation[J]. Chinese Journal of Scientific Instrument, 2019, 40(10):66-73. (in Chinese)
[16]	陈佳鲜, 毛文涛, 刘京, 等. 基于深度时序特征迁移的轴承剩余寿命预测方法[J]. 控制与决策, 2021, 36(7):1699-1706.
	CHEN J X, MAO W T, LIU J, et al. Remaining useful life prediction of bearing based on deep temporal feature transfer[J]. Control and Decision, 2021, 36(7):1699-1706. (in Chinese)
[17]	申中杰, 陈雪峰, 何正嘉, 等. 基于相对特征和多变量支持向量机的滚动轴承剩余寿命预测[J]. 机械工程学报, 2013, 49(2):183-189.
	SHEN Z J, CHEN X F, HE Z J, et al. Remaininglife predictions of rolling bearing based on relative features and mmultivariablesupport vector machine[J]. Journal of Mechanical Engineering, 2013, 49(2):183-189. (in Chinese)
[18]	张龙, 宋成洋, 邹友军, 等. 基于Renyi熵和K-medoids聚类的轴承性能退化评估[J]. 振动与冲击, 2020, 39(20):24-31,46.
	ZHANG L, SONG C Y, ZOU Y J, et al. Bearing performance degradation assessment based on renyi entropy and K-medoids clustering[J]. Journal of Vibration and Shock, 2020, 39(20):24-31,46. (in Chinese)
[19]	蔡曜, 司玉辉, 王玉琢, 等. 挠性陀螺EMD-ARIMA漂移模型设计与应用[J/OL]. 北京航空航天大学学报:1-13[2023-03-04].DOI:10.13700/j.bh.1001-5965.2022.0871.
	CAI Y, SI Y H, WANG Y Z, et al. Design and application of EMD-ARIMA drift model for flexible gyro[J/OL]. Journal of Beijing University of Aeronautics and Astronautics:1-13[2023-03-04].DOI:10.13700/j.bh.1001-5965.2022.0871. (in Chinese)
[20]	程立, 马文锁, 夏新涛, 等. 基于改进模糊熵和灰关系的滚动轴承性能退化评估[J]. 机械传动, 2022, 46(1):56-64,86. doi: 10.16578/j.issn.1004.2539.2022.01.007
	CHENG L, MA W S, XIA X T, et al. Performance degradation evaluation of rolling bearing based on improved fuzzy entropy and grey relation[J]. Journal of Mechanical Transmission, 2022, 46(1):56-64,86. (in Chinese) doi: 10.16578/j.issn.1004.2539.2022.01.007
[21]	李文华, 汪立新, 沈强, 等. 基于EMD的MEMS陀螺仪随机漂移分析方法[J]. 北京航空航天大学学报, 2021, 47(9):1927-1932.
	LI W H, WANG L X, SHEN Q, et al. Random drift analysis method of MEMS gyroscope based on EMD[J]. Journal of Beijing University of Aeronautics and Astronautics, 2021, 47(9):1927-1932. (in Chinese)
[22]	舒涛, 张一弛, 丁日显. 基于灰色模型与LSTM网络的旋转机械轴承寿命预测[J]. 系统工程与电子技术, 2021, 43(8):2355-2361. doi: 10.12305/j.issn.1001-506X.2021.08.39
	SHU T, ZHANG Y C, DING R X. Life prediction of bearings in rotating machinery based on grey model and LSTM network[J]. Systems Engineering and Electronics, 2021, 43(8):2355-2361. (in Chinese) doi: 10.12305/j.issn.1001-506X.2021.08.39
[23]	蔡安江, 蔡曜, 郭师虹, 等. 两端式同轨双车运行模式的货位分配策略[J]. 计算机集成制造系统, 2018, 24(12):3165-3177.
	CAI A J, CAI Y, GUO S H, et al. Storage location assignment strategy of double-crane in automated warehouse[J]. Computer Integrated Manufacturing Systems, 2018, 24(12):3165-3177. (in Chinese)

累积运转时间/h	退化特征编号
	220043		220011		220003
	IRMS	Renyi熵	IRMS	Renyi熵	IRMS	Renyi熵
0	1.000000	12.760610	1.000000	12.154460	1.000000	12.814080
20	1.008089	12.995436	1.009850	12.615788	1.006422	12.772818
55	1.011968	13.121880	1.022386	13.279650	1.014126	12.711260
64	1.011010	12.743217	1.011759	12.638034	1.013960	12.755355
78	1.011944	12.935760	1.017388	12.966638	1.014049	12.732080
123	1.009818	12.543150	1.006930	12.317130	1.013879	12.778550
144	1.008893	12.523078	1.020371	12.452550	1.013687	12.728857
164	1.007781	12.503110	1.020423	12.588730	1.011535	12.676650
187	1.007952	12.988720	1.026127	12.839160	1.001110	12.783950
205	1.007246	12.886852	1.028555	13.131620	1.009151	12.791182
226	1.008892	12.832840	1.018645	13.257020	1.011227	12.794500
243	1.008743	12.984810	1.017111	12.748890	1.015668	13.270220
267	1.008647	13.022864	1.025812	12.938729	1.017248	13.285421
280	1.014697	13.086070	1.021895	13.311210	1.017211	13.316950
309	1.010683	12.652590	1.019687	12.844010	1.011412	12.657030
333	1.012474	12.868410	1.019994	12.949610	1.019678	12.477780
358	1.012305	12.870947	1.026154	12.968855	1.018409	12.734486
374	1.012103	12.873710	1.018154	12.990050	1.014866	13.053870
396	1.008989	13.402536	1.022165	13.114836	1.035628	13.588148
419	1.010530	13.151701	1.023789	13.055913	1.027330	13.330807
442	1.007410	13.649570	1.013922	13.180680	1.046018	13.913390
451	1.101613	14.792932	1.105410	13.773929	1.085554	14.365252
467	1.063090	13.910431	1.155980	13.599005	1.197124	14.607567
479	1.086249	14.063920	1.196534	16.275224	1.129003	14.951632
492	1.099945	14.180621	1.213297	17.048331	1.170872	15.468065
510	1.171855	15.227048	1.259082	17.635756	1.212425	15.998750
521	1.123821	14.871357	1.298111	18.252993	1.199307	16.130718
536	1.177895	15.920309	1.356910	18.876250	1.222185	16.437612
556	1.207621	15.569185			1.247730	16.821767
580	1.212569	15.908466			1.290866	17.289143
603	1.275265	16.501803			1.346950	18.039797
621	1.278898	16.834760			1.338238	17.917427
641	1.273854	17.033510			1.375423	18.512040
661	1.331158	17.865882
681	1.366120	18.297010

累积运转时间/h	退化特征编号
	220043		220011		220003
	IRMS	Renyi熵	IRMS	Renyi熵	IRMS	Renyi熵
0	1.000000	12.760610	1.000000	12.154460	1.000000	12.814080
20	1.008089	12.995436	1.009850	12.615788	1.006422	12.772818
55	1.011968	13.121880	1.022386	13.279650	1.014126	12.711260
64	1.011010	12.743217	1.011759	12.638034	1.013960	12.755355
78	1.011944	12.935760	1.017388	12.966638	1.014049	12.732080
123	1.009818	12.543150	1.006930	12.317130	1.013879	12.778550
144	1.008893	12.523078	1.020371	12.452550	1.013687	12.728857
164	1.007781	12.503110	1.020423	12.588730	1.011535	12.676650
187	1.007952	12.988720	1.026127	12.839160	1.001110	12.783950
205	1.007246	12.886852	1.028555	13.131620	1.009151	12.791182
226	1.008892	12.832840	1.018645	13.257020	1.011227	12.794500
243	1.008743	12.984810	1.017111	12.748890	1.015668	13.270220
267	1.008647	13.022864	1.025812	12.938729	1.017248	13.285421
280	1.014697	13.086070	1.021895	13.311210	1.017211	13.316950
309	1.010683	12.652590	1.019687	12.844010	1.011412	12.657030
333	1.012474	12.868410	1.019994	12.949610	1.019678	12.477780
358	1.012305	12.870947	1.026154	12.968855	1.018409	12.734486
374	1.012103	12.873710	1.018154	12.990050	1.014866	13.053870
396	1.008989	13.402536	1.022165	13.114836	1.035628	13.588148
419	1.010530	13.151701	1.023789	13.055913	1.027330	13.330807
442	1.007410	13.649570	1.013922	13.180680	1.046018	13.913390
451	1.101613	14.792932	1.105410	13.773929	1.085554	14.365252
467	1.063090	13.910431	1.155980	13.599005	1.197124	14.607567
479	1.086249	14.063920	1.196534	16.275224	1.129003	14.951632
492	1.099945	14.180621	1.213297	17.048331	1.170872	15.468065
510	1.171855	15.227048	1.259082	17.635756	1.212425	15.998750
521	1.123821	14.871357	1.298111	18.252993	1.199307	16.130718
536	1.177895	15.920309	1.356910	18.876250	1.222185	16.437612
556	1.207621	15.569185			1.247730	16.821767
580	1.212569	15.908466			1.290866	17.289143
603	1.275265	16.501803			1.346950	18.039797
621	1.278898	16.834760			1.338238	17.917427
641	1.273854	17.033510			1.375423	18.512040
661	1.331158	17.865882
681	1.366120	18.297010

模型参数	设定值
新陈代谢子列维度Q	10
BBO算法迭代次数In	400
栖息地数量L	200
最大迁入率、迁出率P_λ_,_μ	1
AS迁移算子概率P_AS	0.4
最大突变率Pm_m	0.05

模型参数	设定值
新陈代谢子列维度Q	10
BBO算法迭代次数In	400
栖息地数量L	200
最大迁入率、迁出率P_λ_,_μ	1
AS迁移算子概率P_AS	0.4
最大突变率Pm_m	0.05

间隔变化后累积工作时间/h 及平均值	实测值间隔变换后的IRMS	滚动建模次数 q	拟合精度/ %	IRMS 预测值	预测精度/%
227.8	1.018482	1	98.38	0.927759	91.09
248.9	1.019250	2	96.50	1.001069	98.22
263.9	1.024688	3	99.24	1.004260	98.01
289.6	1.021164	4	99.20	0.948433	92.88
302.5	1.020182	5	99.02	1.022177	99.80
329.7	1.019952	6	99.64	1.012181	99.24
354.0	1.025169	7	99.60	1.017137	99.22
379.6	1.019175	8	99.62	1.023771	99.55
394.8	1.021946	9	99.77	1.013074	99.13
417.2	1.023662	10	99.03	1.013478	99.01
440.5	1.014566	11	99.76	1.027540	98.72
464.1	1.146814	12	91.91	1.093822	95.38
471.5	1.171187	13	98.34	1.157042	98.79
487.8	1.207881	14	98.17	1.226958	98.42
497.9	1.228304	15	96.23	1.257111	97.65
509.7	1.258319	16	97.21	1.243437	98.82
526.9	1.321238	17	98.44	1.290234	97.65
536.0	1.356910	18	97.36	1.322070	97.43
平均值			98.190		97.723