Research Article
Health Indicator for Predictive Maintenance Based on Fuzzy
CognitiveMaps,GreyWolf,andK-NearestNeighborsAlgorithms
G.Mazzuto ,S.Antomarioni ,F.E.Ciarapica ,andM.Bevilacqua
Dipartimento Ingegneria Industriale e Scienze Matematiche, Universit` a Politecnica Delle Marche, Via Brecce Bianche,
Ancona 60131, Italy
Correspondence should be addressed to G. Mazzuto; g.mazzuto@univpm.it
Received 22 September 2020; Revised 8 October 2020; Accepted 20 January 2021; Published 17 February 2021
Academic Editor: Filippo De Carlo
Copyright©2021G.Mazzutoetal.isisanopenaccessarticledistributedundertheCreativeCommonsAttributionLicense,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
An essential step in the implementation of predictive maintenance involves the health state analysis of productive equipment in
ordertoprovidecompanymanagerswithperformanceanddegradationindicatorswhichhelptopredictcomponentcondition.In
this paper, a supervised approach for health indicator calculation is provided combining the Grey Wolf Optimisation method,
Swarm Intelligence algorithm, and Fuzzy Cognitive Maps. e k-neighbors algorithms is used to predict the Remaining Useful
Lifeofanitem,since,inadditiontoitssimplicity,theyproducegoodresultsinalargenumberofdomains.eapproachaimsto
solvetheproblemthatfrequentlyoccursininterpolationprocedures:theapproximationoffunctionsbelongingtoachosenclass
of functions of which we have no knowledge. e proposed algorithm allows maintenance managers to distinguish different
degradationprofilesindepthwithaconsequentlymorepreciseestimateoftheRemainingUsefulLifeofanitemand,inaddition,
an in-depth understanding of the degradation process. Specifically, in order to show its suitability for predictive maintenance, a
dataset on NASA aircraft engines has been used and results have been compared to those obtained with a neural network
approach. Results highlight how all of the degradation profiles, obtained using the proposed approach, are modelled in a more
detailedmanner,allowingonetosignificantlydistinguishdifferentsituations.Moreover,thephysicalcorespeedandthecorrected
fan speed have been identified as the main critical factors to the engine degradation.
1.Introduction
Although predictive maintenance practices have existed for
many years, only recently, thanks to the emerging Industry
4.0 technologies with increasingly reliable and affordable
smart systems, it has become widely accessible [1]. It has
severaladvantages,includingmachinelifeincreaseby3–5%,
reduced maintenance costs by up to 40%, and returns on
investment up to 10 times [2].
Oneofthemostrelevantstepsinthepredictionprocess
is the choice of the best approach for the item behaviour
assessment, such as data-driven or model-driven approach
[3]. In particular, according to the platform developed by
Pateletal.[4]fortheapplicationofIndustry4.0principlesto
the industrial reality, the data-analytic layer is crucial to
understand a plant functioning. Moreover, if properly
designed,itallowsuserstoidentifythepresenceofinvisible
relationsamongdataprovidedbytheapplicationlayer[5].It
isalsotruethat,accordingtothe“nofreelunch”theorems,a
standard procedure for predictive maintenance does not
exist. Still, it must be chosen among those that best suit the
reality under analysis [6]. In any case, regardless of the
adoptedprocess,foramoreaccurateandoptimalprediction,
it is necessary to gather and analyse appropriately large
amountsofdatawithinatimeframe[7,8]withconsequent
problems deriving from the identification of the most ac-
curate health indicators. e health of a system can be
definedasthedeviationordegradationofanitembehaviour
from its regular operating performance [9].
e calculation of a suitable health indicator (HI) is
fundamental to establish a link between the deviation or
degradationofanitemanditsRemainingUsefulLife(RUL).
us, an accurate HI is a key for a more precise prediction
tool, guaranteeing its reproducibility [10, 11]. is
Hindawi
Mathematical Problems in Engineering
Volume 2021, Article ID 8832011, 21 pages
https://doi.org/10.1155/2021/8832011