基于藤蔓树降维剪枝的多部件系统剩余寿命预测

PDF(3697 KB)

振动与冲击 ›› 2024, Vol. 43 ›› Issue (21) : 31-45.

论文

基于藤蔓树降维剪枝的多部件系统剩余寿命预测

董增寿，裴杰，石慧，常春波，刘昕然

作者信息 +

Residual life prediction of multi-component system based on vine tree reduced-dimensional reduction pruning

DONG Zengshou, PEI Jie, SHI Hui, CHANG Chunbo, LIU Xinran

Author information +

文章历史 +

摘要

随着多部件系统日益复杂化和智能化，系统中多个部件间协同工作，其退化的相互影响往往不可忽略。因此，在考虑多部件系统部件间相互影响的基础上，提出了一种构造藤蔓树降维剪枝二元Copula函数来对高维变量间相关结构进行建模的剩余寿命预测方法。首先通过Lévy过程来建模多部件系统的退化模型，利用Akaike信息准则进行Copula函数的最优遴选。其次通过Copula函数来表征高维多变量下两两部件间的双向随机相关性，采用藤蔓树降维剪枝的方法来对高维变量间相关结构进行降维，并进行多部件系统部件的实时剩余寿命预测建模；接着考虑单个部件异质性的影响下，利用贝叶斯参数估计和极大似然估计结合的方法来实现对随机模型以及随机超参数的动态更新。最后通过C-MAPSS涡轮发动机模型来验证所提出模型的合理性和有效性。

Abstract

As multi-component systems become increasingly complex and intelligent, multi-components in the system work in concert with each other, and their degraded interactions are often not negligible. Therefore, based on the consideration of the interactions between the components of a multi-component system, A Remaining Useful Life Prediction method was proposed that constructs a Vine tree reduced-dimensional pruning pair-Copula function to model the relevance structure between high-dimensional variables. The degradation model of the multi-component system is first modelled through a Lévy process, and the optimal selection of the Copula function is carried out using the Akaike information criterion. Secondly, the bi-directional stochastic dependence between two components under high-dimensional multivariate is characterized by Copula function, and the Vine tree reduced-dimensional pruning method is used to downsize the relevance structure between high-dimensional variables and to model the real-time Remaining Useful Life Prediction of the components of the multi-component system; Then the combination of Bayesian parameter estimation and maximum likelihood estimation is employed to achieve dynamic updating of the stochastic model as well as the stochastic hyperparameters under the consideration of the effect of individual component heterogeneity. Finally, the rationality and validity of the proposed model are verified by the C-MAPSS turbine engine model.

导出引用

董增寿, 裴杰, 石慧, 常春波, 刘昕然. 基于藤蔓树降维剪枝的多部件系统剩余寿命预测[J]. 振动与冲击, 2024, 43(21): 31-45

DONG Zengshou, PEI Jie, SHI Hui, CHANG Chunbo, LIU Xinran. Residual life prediction of multi-component system based on vine tree reduced-dimensional reduction pruning[J]. Journal of Vibration and Shock, 2024, 43(21): 31-45

参考文献

[1] PAN, D., J.-B. LIU, and J. CAO, Remaining useful life estimation using an inverse Gaussian degradation model[J]. Neurocomputing, 2016. 185: p. 64-72.
[2] ZIO, E., Prognostics and Health Management (PHM): Where are we and where do we (need to) go in theory and practice[J]. Reliability Engineering & System Safety, 2022. 218: p. 108119.
[3] 曹现刚, 等, 基于KPCA-LSTM的旋转机械剩余使用寿命预测[J]. 振动与冲击, 2023. 42(24): p. 81-91.
CAO Xiangang, et al., Remaining Useful life prediction of rotating machinery based on KPCA-LSTM[J]. Journal of Vibration and Shock, 2023. 42(24): p. 81-91.
[4] LEI, Y., et al., Machinery health prognostics: A systematic review from data acquisition to RUL prediction[J]. Mechanical Systems and Signal Processing, 2018. 104: p. 799-834.
[5] 王焱, 等, 基于改进ECANet-TCN和迁移学习的轴承剩余寿命预测[J]. 振动与冲击, 2023. 42(21): p. 149-159.
WANG Yan, et al., Bearing residual life prediction based on improved ECANet-TCN and transfer learning[J]. Journal of Vibration and Shock, 2023. 42(21): p. 149-159.
[6] MANJURUL ISLAM, M.M., A.E. PROSVIRIN, and J.-M. KIM, Data-driven prognostic scheme for rolling-element bearings using a new health index and variants of least-square support vector machines[J]. Mechanical Systems and Signal Processing, 2021. 160: p. 107853.
[7] ZHANG, H., X. Xi, and R. PAN, A two-stage data-driven approach to remaining useful life prediction via long short-term memory networks[J]. Reliability Engineering & System Safety, 2023. 237: p. 109332.
[8] XING, J., H. ZHANG, and J. ZHANG, Remaining useful life prediction of – Lithium batteries based on principal component analysis and improved Gaussian process regression[J]. International Journal of Electrochemical Science, 2023. 18(4): p. 100048.
[9] KOU, L., et al., Prediction system of rolling contact fatigue on crossing nose based on support vector regression[J]. Measurement, 2023. 210: p. 112579.
[10] ZHANG, S., Q. ZHAI, and Y. Li, Degradation modeling and RUL prediction with Wiener process considering measurable and unobservable external impacts[J]. Reliability Engineering & System Safety, 2023. 231: p. 109021.
[11] WANG, H., et al., Remaining Useful Life Prediction and Optimal Maintenance Time Determination for a Single Unit Using Isotonic Regression and Gamma Process Model[J]. Reliability Engineering & System Safety, 2021. 210: p. 107504.
[12] Li, H., E. DELOUX, and L. DIEULLE, A condition-based maintenance policy for multi-component systems with Lévy Copulas dependence[J]. Reliability Engineering & System Safety, 2016. 149: p. 44-55.
[13] SUN, B., et al., An improved inverse Gaussian process with random effects and measurement errors for RUL prediction of hydraulic piston pump[J]. Measurement, 2021. 173: p. 108604.
[14] ZHAO, Y. and Y. WANG, Remaining useful life prediction for multi-sensor systems using a novel end-to-end deep-learning method[J]. Measurement, 2021. 182: p. 109685.
[15] CHEN, X. and M. ZENG, Convolution-Graph Attention Network With Sensor Embeddings for Remaining Useful Life Prediction of Turbofan Engines[J]. IEEE Sensors Journal, 2023. 23(14): p. 15786-15794.
[16] LI, N., et al., Remaining useful life prediction based on a multi-sensor data fusion model[J]. Reliability Engineering & System Safety, 2021. 208: p. 107249.
[17] LU, B., Z. CHEN, and X. ZHAO, Data-driven dynamic predictive maintenance for a manufacturing system with quality deterioration and online sensors[J]. Reliability Engineering & System Safety, 2021. 212: p. 107628.
[18] Dekker, R., V. FRANK, and R.E. WILDEMAN. A Review of Multi-Component Maintenance Models with Economic Dependence[J]. in Erasmus University Rotterdam, Erasmus School of Economics (ESE), Econometric Institute. 1996.
[19] DINH, D.-H., P. DO, and B. IUNG, Degradation modeling and reliability assessment for a multi-component system with structural dependence[J]. Computers & Industrial Engineering, 2020. 144: p. 106443.
[20] DINH, D.-H., P. DO, and B. IUNG, Multi-level opportunistic predictive maintenance for multi-component systems with economic dependence and assembly/disassembly impacts[J]. Reliability Engineering & System Safety, 2022. 217: p. 108055.
[21] OAKLEY, J.L., K.J. WILSON, and P. PHILIPSON, A condition-based maintenance policy for continuously monitored multi-component systems with economic and stochastic dependence[J]. Reliability Engineering & System Safety, 2022. 222: p. 108321.
[22] KONG, X., J. YANG, and L. LI, Reliability analysis for multi-component systems considering stochastic dependency based on factor analysis[J]. Mechanical Systems and Signal Processing, 2022. 169: p. 108754.
[23] LI, H., et al., Condition-based maintenance strategies for stochastically dependent systems using Nested Lévy Copulas[J]. Reliability Engineering & System Safety, 2022. 217: p. 108038.
[24] LIU, S. and H. JIANG, Engine remaining useful life prediction model based on R-Vine Copula with multi-sensor data[J]. Heliyon, 2023. 9(6): p. e17118.
[25] JOE, H., Families of m-Variate Distributions with Given Margins and m(m-1)/2 Bivariate Dependence Parameters[J]. Lecture Notes-Monograph Series, 1996. 28: p. 120-141.
[26] BEDFORD, T. and R.M. COOKE, Vines: A New Graphical Model for Dependent Random Variables[J]. The Annals of Statistics, 2002. 30(4): p. 1031-1068.
[27] JIAXIN, Z. and S. MICHAEL, On the quantification and efficient propagation of imprecise probabilities with Copula dependence[J]. International Journal of Approximate Reasoning, 2020. 122(prepublish).
[28] Wang, Z., et al., Forecasted Scenarios of Regional Wind Farms Based on Regular Vine Copulas[J]. Journal of Modern Power Systems and Clean Energy, 2020. 8(01): p. 77-85.
[29] 胡启国, 周松. 基于Vine Copula模型的失效相关机械零件可靠性分析[J].机械强度, 2019,41(06):1365-1371.
HU Qiguo, ZHOUSong. Reliability Analysis of Failure Related Mechanical Parts Based On Vine Copula Model[J]. Journal of Mechanical Strength, 2019,41(06):1365-1371.
[30] ZHAO, Y., et al., Modeling multivariate dependence by nonparametric pair-Copula construction in composite system reliability evaluation[J]. International Journal of Electrical Power & Energy Systems, 2021. 124: p. 106373.
[31] SCHOUTENS, W., Exotic options under Lévy models: An overview[J]. Journal of Computational and Applied Mathematics, 2006. 189(1): p. 526-538.
[32] LIU, W., et al., Reliability Analysis of Operational Metro Tunnel Based on a Dynamic Bayesian Copula Model[J]. Journal of Computing in Civil Engineering, 2020. 34(3).
[33] Y. Zhao, Q. Liu, J. Kuang, et al. Modeling multivariate dependence by nonparametric pair-copula construction in composite system reliability evaluation[J]. International Journal of Electrical Power & Energy Systems, 2021, 124: 106373.
[34] 彭宅铭, 程龙生, 姚启峰. 多传感器数据融合的复杂系统退化模式挖掘[J]. 振动与冲击, 2022. 41(13): p. 239-245+251.
PENG Zhaiming, CHENG Longsheng, YAO Qifeng. Degradation mode mining for complex system based on multi-sensor data fusion[J]. Journal of Vibration and Shock, 2022. 41(13): p. 239-245+251.
[35] ZHANG, Y., et al., Health status assessment and remaining useful life prediction of aero-engine based on BiGRU and MMoE[J]. Reliability Engineering & System Safety, 2022. 220: p. 108263.
[36] 宋海龙, 等, 基于LightGBM的航空发动机剩余使用寿命预测[J]. 现代计算机, 2021. 27(35): p. 47-52.
Song Hailong, et al., Remaining Useful Life Prediction for Aero-Engine Based on LightGBM[J]. Modern Computer, 2021, 27(35): 47-52.