基于多源信息融合与HMM的剩余寿命预测

doi:10.12305/j.issn.1001-506X.2022.05.38

摘要/Abstract

摘要：

针对设备剩余使用寿命预测问题, 提出一种基于多源信息融合与隐马尔可夫模型的预测方法。首先, 针对发动机结构复杂、监控数据参数多等问题, 提出一种基于传感器信噪比和主成分分析(principal component analysis, PCA)降维的多源传感器数据融合方法。在此基础上, 利用样本数据训练高斯混合隐马尔可夫模型, 同时为降低模型偏差并避免过拟合风险, 提出一种“定制”策略训练方法, 训练后的模型可用于系统健康状态识别和剩余使用寿命预测。最后, 通过美国国家航空航天局公开的航空发动机仿真数据集对所提方法进行了验证, 并与几种具有代表性且预测精度较高的文献方法进行了比较分析, 验证了方法的有效性。

关键词: 多源信息融合, 隐马尔可夫模型, 剩余使用寿命, 模型训练

Abstract:

Aiming at the problem of equipment remaining useful life prediction, a prediction method based on multi-source information fusion and hidden Markov model is proposed. Firstly, a multi-source sensor data fusion method based on signal-to-noise ratio of sensor and principal component analysis (PCA) dimensionality reduction is proposed to solve the problems of complex engine structure and multiple monitoring data parameters.On this basis, the Gaussian mixture hidden Markov model is trained using the sample data. At the same time, in order to reduce the deviation of model and avoid the risk of over fitting, a "customized" strategy training method is proposed. The trained model can be used for system health status recognition and remaining useful life prediction. Finally, the effectiveness of the proposed method is verified by the aeroengine simulation data set published by National Aeronautics and Space Administration, and compared with several representative literature methods with high prediction accuracy.

Key words: multi-source information fusion, hidden Markov model, remaining useful life, model training

中图分类号:

V239

黄林, 龚立, 姜伟, 王康勃. 基于多源信息融合与HMM的剩余寿命预测[J]. 系统工程与电子技术, 2022, 44(5): 1747-1756.

Lin HUANG, Li GONG, Wei JIANG, Kangbo WANG. Remaining useful life prediction based on multi-source information fusion and HMM[J]. Systems Engineering and Electronics, 2022, 44(5): 1747-1756.

图/表 11

图1

图2

图3

图4

表1

图5

图6

图7

图8

图9

表2

参考文献 30

1	LEI Y G , LI N P , GUO L , et al. Machinery health prognostics: a systematic review from data acquisition to RUL prediction[J]. Mechanical Systems and Signal Processing, 2018, 104, 799- 834. doi: 10.1016/j.ymssp.2017.11.016
2	ORDÓNEZ C , LASHERAS F S , ROCA-PARDINAS J , et al. A hybrid ARIMA-SVM model for the study of the remaining useful life of aircraft engines[J]. Journal of Computational and Applied Mathematics, 2019, 346, 184- 191. doi: 10.1016/j.cam.2018.07.008
3	KUMAR A , CHINNAM R B , TSENG F . An HMM and polynomial regression based approach for remaining useful life and health state estimation of cutting tools[J]. Computers & Industrial Engineering, 2019, 128, 1008- 1014.
4	SON J B , ZHOU S Y , SANKAVARAM C , et al. Remaining useful life prediction based on noisy condition monitoring signals using constrained Kalman filter[J]. Reliability Engineering & System Safety, 2016, 152, 38- 50.
5	COSTA P R D O , AKCAY A , ZHANG Y , et al. Attention and long short-term memory network for remaining useful lifetime predictions of turbofan engine degradation[J]. International Journal of Prognostics and Health Management, 2019, 10 (4): 211477960.
6	ELLEFSEN A L , BJORLYKHAUG E , AESOY V , et al. Remaining useful life predictions for turbofan engine degradation using semi-supervised deep architecture[J]. Reliability Engineering & System Safety, 2019, 183, 240- 251.
7	RABINER L R . A tutorial on hidden Markov models and selec-ted applications in speech recognition[J]. Proceedings of the IEEE, 1989, 77 (2): 257- 286. doi: 10.1109/5.18626
8	张袁鹏, 郑岱堃, 李昕哲, 等. 基于隐马尔可夫模型的动态规划检测前跟踪算法[J]. 系统工程与电子技术, 2019, 41 (11): 2479- 2487.
	ZHANG Y P , ZHENG D K , LI X Z , et al. Dynamic programming track-before-detect algorithm based on hidden Markov model[J]. Systems Engineering and Electronics, 2019, 41 (11): 2479- 2487.
9	JIANG J , CHEN R Y , CHEN M , et al. Dynamic fault prediction of power transformers based on hidden Markov model of dissolved gases analysis[J]. IEEE Trans.on Power Deliver, 2019, 34 (4): 1393- 1400. doi: 10.1109/TPWRD.2019.2900543
10	ZHU J L , GE Z Q , SONG Z H . HMM-driven robust probabilistic principal component analyzer for dynamic process fault classification[J]. IEEE Trans.on Industrial Electronics, 2015, 62 (6): 3814- 3821.
11	YIAKOPOULOS C , GRYLLIAS K , CHIOUA M , et al. An on-line SAX and HMM-based anomaly detection and visualization tool for early disturbance discovery in a dynamic industrial process[J]. Journal of Process Control, 2016, 44, 134- 159. doi: 10.1016/j.jprocont.2016.05.007
12	GALAGEDARAGE D M , KHAN F . Process fault prognosis using hidden Markov model-Bayesian networks hybrid model[J]. Industrial & Engineering Chemistry Research, 2019, 58 (27): 12041- 12053.
13	WANG F , TAN S , YANG Y W , et al. Hidden Markov model-based fault detection approach for a multimode process[J]. Industrial & Engineering Chemistry Research, 2016, 55 (16): 4613- 4621.
14	DU Y , WU T H , MAKIS V . Parameter estimation and remaining useful life prediction of lubricating oil with HMM[J]. Wear, 2017, 376, 1227- 1233.
15	SOUALHI A , CLERC G , RAZIK H , et al. Hidden Markov models for the prediction of impending faults[J]. IEEE Trans.on Industrial Electronics, 2016, 63 (5): 3271- 3281. doi: 10.1109/TIE.2016.2535111
16	LE T T , CHATELAIN F , BÉRENGUER C . Multi-branch hidden Markov models for remaining useful life estimation of systems under multiple deterioration modes[J]. Proceedings of the Institution of Mechanical Engineers, Part O: Journal of Risk and Reliability, 2016, 230 (5): 473- 484. doi: 10.1177/1748006X15624584
17	RAMASSO E , ROMBAUT M , ZERHOUNI N . Joint prediction of continuous and discrete states in time-series based on belief functions[J]. IEEE Ttrans.on Cybernetics, 2012, 43 (1): 37- 50.
18	周俊. 数据驱动的航空发动机剩余使用寿命预测方法研究[D]. 南京: 南京航空航天大学, 2017.
	ZHOU J. Research on data driven residual service life prediction method of aero-engine[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2017.
19	AN S J , LIU W Q , VENKATESH S . Fast cross-validation algorithms for least squares support vector machine and kernel ridge regression[J]. Pattern Recognition, 2007, 40 (8): 2154- 2162. doi: 10.1016/j.patcog.2006.12.015
20	SMOLA A J , SCHÖLKOPF B . A tutorial on support vector regression[J]. Statistics and computing, 2004, 14 (3): 199- 222. doi: 10.1023/B:STCO.0000035301.49549.88
21	周志华. 机器学习[M]. 北京: 清华大学出版社, 2016.
	ZHOU Z H . Machine learning[M]. Beijing: Tsinghua University Press, 2016.
22	李航. 统计学习方法[M]. 北京: 清华大学出版社, 2012.
	LI H . Statistical learning method[M]. Beijing: Tsinghua University Press, 2012.
23	CAMCI F , CHINNAM R B . Health-state estimation and prognostics in machining processes[J]. IEEE Trans.on Automation Science and Engineering, 2010, 7 (3): 581- 597. doi: 10.1109/TASE.2009.2038170
24	SAXENA A, GOEBEL K, SIMON D, et al. Damage propagation modeling for aircraft engine run-to-failure simulation[C]//Proc. of the IEEE International Conference on Prognostics and Health Management, 2008.
25	WANG T Y, YU J B, SIEGEL D, et al. A similarity-based prognostics approach for remaining useful life estimation of engineered systems[C]//Proc. of the IEEE International Conference on Prognostics and Health Management, 2008.
26	MALHOTRA P, TV V, RAMAKRISHNAN A, et al. Multi-sensor prognostics using an unsupervised health index based on LSTM encoder-decoder[EB/OL]. [2021-02-01]. https://arxiv.org/abs/1608.06154v1.
27	WEAKLIEM D L . A critique of the Bayesian information criterion for model selection[J]. Sociological Methods & Research, 1999, 27 (3): 359- 397.
28	梁泽明, 姜洪权, 周秉直, 等. 多参数相似性信息融合的剩余寿命预测[J]. 计算机集成制造系统, 2018, 24 (4): 5- 11.
	LIANG Z M , JIANG H Q , ZHOU B Z , et al. Multi-variable similarity-based information fusion method for remaining useful life prediction[J]. Computer Integrated Manufacturing Systems, 2018, 24 (4): 5- 11.
29	BABU G S, ZHAO P, LI X. Deep convolutional neural network based regression approach for estimation of remaining useful life[C]//Proc. of the International Conference on Database Systems for Advanced Applications, 2016: 214-228.
30	彭开香, 皮彦婷, 焦瑞华, 等. 航空发动机的健康指标构建与剩余寿命预测[J]. 控制理论与应用, 2020, 37 (4): 713- 720.
	PENG K X , PI Y T , JIAO R H , et al. Health indicator construction and remaining useful life prediction for aircraft engine[J]. Control Theory & Applications, 2020, 37 (4): 713- 720.

序号	变量	描述
1	T2/℃	风扇入口总温
2	T24/℃	低压压气机出口总温
3	T30/℃	高压压气机出口总温
4	T50/℃	低压涡轮出口总温
5	P2/kPa	风扇入口压力
6	P15/kPa	外涵总压
7	P30/kPa	高压压气机出口总压
8	Nf/rpm	风扇物理转速
9	Nc/rpm	核心机物理转速
10	epr	发动机压比(P50/P2)
11	Ps30/kPa	高压压气机出口静压
12	Phi	燃油流量与P30比值
13	NRf/rpm	风扇换算转速
14	NRc/rpm	核心机换算转速
15	BPR	涵道比
16	farB	燃烧室燃气比
17	htBleed	引气焓值
18	Nf_dmd/rpm	设定风扇转速
19	PCNfR_dmd/rpm	设定核心机换算转速
20	W31/(kg/s)	高压涡轮冷却引气流量
21	W32/(kg/s)	低压涡轮冷却引气流量

预测方法	分数	RMSE	MAPE	FPR/%	FNR/%	A/%
多源信息融合	-	-	5.02	-	-	-
多层感知机	17 972	37.562 9	-	-	-	-
深度卷积网络	1 286	18.448 0	-	-	-	-
长短时记忆网络	1 263	-	18	34	30	36
DBN-HMM	-	33.818 1	24.544 8	51	31	18
本文	588	19.4	14.8	28	21	51

[1]	张瑜, 吴凯, 郭杰, 葛志闪, 张宝超. 基于数据质量评估的自适应序贯航迹关联算法[J]. 系统工程与电子技术, 2022, 44(11): 3477-3485.
[2]	苗成林, 李彤, 吕军, 常成. 基于隐马尔可夫模型的动态跳频信道接入算法[J]. 系统工程与电子技术, 2019, 41(8): 1873-1880.
[3]	杜镇宇, 刘方正, 李翼宏. 基于HMM的APT攻击路径预测[J]. 系统工程与电子技术, 2019, 41(4): 826-834.
[4]	张袁鹏, 郑岱堃, 李昕哲, 孙永健. 基于隐马尔可夫模型的动态规划检测前跟踪算法[J]. 系统工程与电子技术, 2019, 41(11): 2479-2487.
[5]	尚永爽，许爱强，李文海，王怡苹，盛沛. 部分可观测信息条件下系统最佳检修策略分析[J]. Journal of Systems Engineering and Electronics, 2012, 34(4): 749-753.
[6]	陈海洋, 高晓光, 梅军峰. 广义隐马尔可夫模型的快速前向后向算法[J]. Journal of Systems Engineering and Electronics, 2012, 34(10): 2175-2179.
[7]	杨广, 吴晓平, 宋业新, 田树新. 基于粗糙集理论的多源信息融合故障诊断方法 [J]. Journal of Systems Engineering and Electronics, 2009, 31(8): 2013-2019.
[8]	周欣然, 滕召胜, 易钊. 基于混合协同粒子群优化的广义T-S模糊模型训练方法[J]. Journal of Systems Engineering and Electronics, 2009, 31(5): 1189-1193.