In this work, we consider the problem of predicting the remaining useful life of a piece of equipment, based on data collected from a heterogeneous fleet working under different operating conditions. When the equipment experiences variable operating conditions, individual data-driven prognostic models are not able to accurately predict the remaining useful life during the entire equipment life. The objective of this work is to develop an ensemble approach of different prognostic models for aggregating their remaining useful life predictions in an adaptive way, for good performance throughout the degradation progression. Two data-driven prognostic models are considered, a homogeneous discrete-time finite-state semi-Markov model and a fuzzy similarity–based model. The ensemble approach is based on a locally weighted strategy that aggregates the outcomes of the two prognostic models of the ensemble by assigning to each model a weight and a bias related to its local performance, that is, the accuracy in predicting the remaining useful life of patterns of a validation set similar to the one under study. The proposed approach is applied to a case study regarding a heterogeneous fleet of aluminum electrolytic capacitors used in electric vehicle powertrains. The results have shown that the proposed ensemble approach is able to provide more accurate remaining useful life predictions throughout the entire life of the equipment compared to an alternative ensemble approach and to each individual homogeneous discrete-time finite-state semi-Markov model and fuzzy similarity–based models.
BaraldiPMangiliFZioE. A belief function theory based approach to combining different representation of uncertainty in prognostics. Inf Sci (Ny)2015; 303: 134–149.
2.
McCallJJ. Maintenance policies for stochastically failing equipment: a survey. Manage Sci1965; 11(5): 493–524.
3.
WeiZTaoTZhuoShuD. A dynamic particle filter-support vector regression method for reliability prediction. Reliab Eng Syst Saf2013; 119: 109–116.
4.
ZhaoZWangF-LJiaM-X. Predictive maintenance policy based on process data. Chemom Intell Lab Syst2010; 103(2): 137–143.
5.
ZioE. Prognostics and health management of industrial equipment. In: Diagnostics and prognostics of engineering systems: methods and techniques. IGI-Global, 2012, pp.333–356. DOI: 10.4018/978-1-4666-2095-7.ch017.
6.
Tobon-MejiaDMedjaherKZerhouniN. Hidden Markov models for failure diagnostic and prognostic. In: Prognostics and system health management conference, PHM’11, Shenzhen, China, 24–25 May 2011, pp.1–8. New York: IEEE.
7.
JardineAKSLinDBanjevicD. A review on machinery diagnostics and prognostics implementing condition-based maintenance. Mech Syst Signal Process2006; 20(7): 1483–1510.
8.
HengAZhangSTanACC. Rotating machinery prognostics: state of the art, challenges and opportunities. Mech Syst Signal Process2009; 23(3): 724–739.
9.
BevilacquaMBragliaM. Analytic hierarchy process applied to maintenance strategy selection. Reliab Eng Syst Saf2000; 70(1): 71–83.
10.
Al-DahidiSDi MaioFBaraldiP. Supporting maintenance decision with empirical models based on fleet-wide data. In: The 49th ESReDA seminar on: innovation through human factors in risk assessment & maintenance, Bruxelles, Belgium, 29–30 October 2015, pp.1–12. European Safety, Reliability and Data Association (ESReDA).
11.
SankavaramCPattipatiBKodaliA. Model-based and data-driven prognosis of automotive and electronic systems. In: 2009 IEEE international conference on automation science and engineering (CASE 2009), Bangalore, India, 22–25 August 2009, pp.96–101. IEEE.
12.
ChiangLHRussellEBraatzRD. Fault detection and diagnosis in industrial systems. 2001. Springer Science & Business Media.
13.
SchwabacherMGoebelK. A survey of artificial intelligence for prognostics. In: Proceedings of 2007 AAAI Fall Symposium: AI for Prognostics, Arlington, VA, USA, 9–11 November 2007, pp.107–114. Association for the Advancement of Artificial Intelligence (AAAI).
14.
SiXSWangWHuCH. Remaining useful life estimation—a review on the statistical data driven approaches. Eur J Oper Res2011; 213(1): 1–14.
15.
SikorskaJZHodkiewiczMMaL. Prognostic modelling options for remaining useful life estimation by industry. Mech Syst Signal Process2011; 25(5): 1803–1836.
16.
LeeJWuFZhaoW. Prognostics and health management design for rotary machinery systems—reviews, methodology and applications. Mech Syst Signal Process2014; 42(1–2): 314–334.
17.
LuoSPattipatiJQiaoKR. Model-based prognostic techniques applied to a suspension system. IEEE Trans Syst Man Cybern: Part A2008; 38(5): 1156–1168.
18.
GebraeelJElwanyNPanA. Residual life predictions in the absence of prior degradation knowledge. IEEE Trans Reliab2009; 58(1): 106–117.
19.
LiuJWangWMaF. A data-model-fusion prognostic framework for dynamic system state forecasting. Eng Appl Artif Intell2012; 25(4): 814–823.
GebraeelJPanN. Prognostic degradation models for computing and updating residual life distributions in a time-varying environment. IEEE Trans Reliab2008; 57(4): 539–550.
22.
LiYKurfessTRLiangSY. Stochastic prognostics for rolling element bearings. Mech Syst Signal Process2000; 14(5): 747–762.
23.
LiYBillingtonSZhangC. Adaptive prognostics for rolling element bearing condition. Mech Syst Signal Process1999; 13(1): 103–113.
24.
Di MaioFZioE. Failure prognostics by a data-driven similarity-based approach. Int J Reliab Qual Saf Eng2014; 20(1): 1350001
25.
TranVTYangB-S. Machine fault diagnosis and prognosis: the state of the art. Int J Fluid Mach Syst2009; 2(1): 61–71.
26.
LiWPhamH. An inspection-maintenance model for systems with multiple competing processes. IEEE Trans Reliab2005; 54(2): 318–327.
27.
SchwabacherM. A survey of data-driven prognostics. In: Proceedings of the AIAA infotech@aerospace conference, Arlington, VA, 26–29 September 2005, pp.1–5. The American Institute of Aeronautics and Astronautics (AIAA).
28.
CobleJBHinesJW. Prognostic algorithm categorization with PHM challenge application. In: 2008 international conference on prognostics and health management, PHM 2008, Denver, CO, USA, 6–9 October 2008. New York: IEEE.
29.
HeimesFO. Recurrent neural networks for remaining useful life estimation. In: International conference on prognostics and health management. PHM 2008, Denver, CO, USA, 6–9 October 2008, pp.1–6. New York: IEEE.
30.
Di MaioFTsuiKLZioE. Combining relevance vector machines and exponential regression for bearing residual life estimation. Mech Syst Signal Process2012; 31: 405–427.
31.
Medina-OlivaGVoisinAMonninM. Predictive diagnosis based on a fleet-wide ontology approach. Knowledge-Based Syst2014; 68: 40–57.
32.
MonninMAbichouBVoisinA. Fleet historical cases for predictive maintenance. In: International conference on acoustical and vibratory methods in surveillance and diagnostics, Surveillance 6, Compiègne, France, 25–26 October 2011. Sciencesconf.org.
33.
AgarwalVLybeckNJBickfordR. Development of asset fault signatures for prognostic and health management in the nuclear industry. In: IEEE conference on prognostics and health management (PHM), Cheney, WA, USA, 22–25 June 2014. IEEE.
34.
Al-DahidiSDi MaioFBaraldiP. Remaining useful life estimation in heterogeneous fleets working under variable operating conditions. Reliab Eng Syst Saf2016; 156: 109–124.
35.
HuCYounBDWangP. Ensemble of data-driven prognostic algorithms with weight optimization and k-fold cross validation. In: Annual conference of the prognostics and health management (PHM) society, Portland, Oregon, 10–16 October 2010, pp. 1–12.
36.
PolikarR. Ensemble based systems in decision making. IEEE Circuits Syst Mag2006; 6(3): 21–45.
37.
Di MaioFHuJTseP. Ensemble-approaches for clustering health status of oil sand pumps. Expert Syst Appl2012; 39(5): 4847–4859.
38.
BaraldiPMangiliFZioE. Investigation of uncertainty treatment capability of model-based and data-driven prognostic methods using simulated data. Reliab Eng Syst Saf2013; 112: 94–108.
39.
CadiniFZioEAvramD. Monte Carlo-based filtering for fatigue crack growth estimation. Probabilistic Eng Mech2009; 24(3): 367–373.
40.
WangTYuJSiegelD. A similarity-based prognostics approach for Remaining Useful Life estimation of engineered systems. In: International conference on prognostics and health management, Denver, CO, USA, 6–9 October 2008, pp.1–6. IEEE.
41.
ZioEDi MaioF. A data-driven fuzzy approach for predicting the remaining useful life in dynamic failure scenarios of a nuclear system. Reliab Eng Syst Saf2010; 95(1): 49–57.
42.
AngstenbergerL. Dynamic fuzzy pattern recognition with applications to finance and engineering. In: International series in intelligent technologies, 2001, vol. 17. The Netherlands: Springer.
43.
BaraldiPCammiAMangiliF. Local fusion of an ensemble of models for the reconstruction of faulty signals. IEEE Trans Nucl Sci2010; 57(2): 793–806.
44.
SaxenaACelayaJSahaB. Metrics for offline evaluation of prognostic performance. Int J Progn Heal Manag2010; 1: 1–20.
45.
Al-DahidiSDi MaioFBaraldiP. A switching ensemble approach for remaining useful life estimation of electrolytic capacitors. In: Safety and reliability: Methodology and applications - Proceedings of the 26th European safety and reliability conference, ESREL 2016, Glasgow, UK, 25–29 September 2016, pp. 2000–2005.
ChenCT. Dynamic preventive maintenance strategy for an aging and deteriorating production system. Expert Syst Appl2011; 38(5): 6287–6293.
48.
MoghaddassRZuoMJ. A parameter estimation method for a condition-monitored device under multi-state deterioration. Reliab Eng Syst Saf2012; 106: 94–103.
49.
ShuMHHsuBMKapurKC. Dynamic performance measures for tools with multi-state wear processes and their applications for tool design and selection. Int J Prod Res2010; 48(16): 4725–4744.
50.
MoghaddassRZuoMJ. An integrated framework for online diagnostic and prognostic health monitoring using a multistate deterioration process. Reliab Eng Syst Saf2014; 124: 92–104.
51.
Al-DahidiSDi MaioFBaraldiP. A novel ensemble clustering for operational transients classification with application to a nuclear power plant turbine. Int J Progn Heal Manag2015; 6(SP3): 1–21.
52.
BaraldiPDi MaioFRigamontiM. Unsupervised clustering of vibration signals for identifying anomalous conditions in a nuclear turbine. J Intell Fuzzy Syst2013; 28(4): 1723–1731.
53.
BaraldiPDi MaioFZioE. Unsupervised clustering for fault diagnosis in nuclear power plant components. Int J Comput Intell Syst2013; 6(4): 764–777.
54.
KendallMStuartA. The advanced theory of statistics, vol. 2. London: Charles Griffin and Company Limited, 1968.
55.
CompareMBaraldiPCannarileF. Homogeneous finite-time, finite-state, semi-Markov modelling for enhancing Empirical Classification System diagnostics of industrial components (under review). Probabilistic Eng Mech2015.
56.
RigamontiMBaraldiPZioE. Particle filter-based prognostics for an electrolytic capacitor working in variable operating conditions. IEEE Trans Power Electron2016; 31(2): 1567–1575.
57.
KulkarniCBiswasGKoutsoukosX. Integrated diagnostic/prognostic experimental setup for capacitor degradation and health monitoring. in AUTOTESTCON (proceedings), Orlando, FL, 13–16 September 2010, pp.351–357. New York: IEEE.
58.
VenetPDarnandHGrelletG. Detection of faults of filter capacitors in a converter. Application to predictive maintenance. 15th Int Telecommun Energy Conf INTELEC ’931993; 2: 229–234.
59.
AbdennadherKVenetPRojatG. A real-time predictive-maintenance system of aluminum electrolytic capacitors used in uninterrupted power supplies. IEEE T Ind Appl2010; 46(4): 1644–1652.
60.
LahyaniAVenetPGrelletG. Failure prediction of electrolytic capacitors during operation of a switchmode power supply. IEEE T Power Electr1998; 13(6): 1199–1207.
61.
GasperiML. Life prediction model for aluminum electrolytic capacitors. IEEE Ind Appl Conf 1996 Thirty-First IAS Annu Meet IAS ’96, Conf Rec1996; 3(1): 1347–1351.
62.
WolfgangE. Examples for failures in power electronics systems. In: ECPE tutorial on reliability of power electronic systems, Nuremberg, Germany, 19–20 April 2007, ECPE European Center for Power Electronics e.V.
