tutorial 3 state-of-the-art realization of prognostics in ... · tutorial 3 : state-of-the-art...
TRANSCRIPT
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
Tutorial 3Tutorial 3State-of-the-art Realization of Prognostics in Industry
Presenters: Michel SCHIEBER (Cassidian T&S)Presenters: Michel SCHIEBER (Cassidian T&S)Jean Baptiste LEGER (Predict)Benoit IUNG (Lorraine University, CRAN)
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
The main problematic met
FOREWORD: Why diag 21 ?
• Divergences between requirements and the state of the art;• Discrepancies between policies and methods;• Divergences of interests from actors;• Hardware / Software validation realized at the end of the design processHardware / Software validation realized at the end of the design process.
• No optimized embedded systems maintenance;• Late re design in the de elopment phase hich res lts in project risk ith high
Have for consequences…
• Late re-design in the development phase which results in project risk with high cost;
• No optimized maintenance cycle;• Full life cycle cost factors are not understood.
A worldwide status A worldwide status• By contract only safety aspects are considered;• Platform of exchanges and services on these disciplines in France … in Europe
Europe do not exist;• Increasing interest (European projects) but no federate actions.
diag 21 is a challenge
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
2
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
FOREWORD: towards Health Monitoring Management
ResearchMaintenance
FOREWORD: towards Health Monitoring Management
Largecompanies
/ SME
ResearchCenters
Testability
Spatial Fixed wings
UniversitiesHigh schools
Diagnostics
PrognosticsRevolving wings
gGovernment
agencies& customers
Rail
Availability
Automotives
EurocopterMBDACassidian
ModellingS d di i
Maritime
Diag 21 main objective is to boost synergies between all actors involvedin common problematic in these domains
Cassidian …Standardisation
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
3
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
FOREWORD: diag 21 objectives
diag 21 is an exchange platform which:
• actively participate to the writing of the future testability, diagnostic, prognostic standards,
diag 21 is an exchange platform which:
• share the knowledge between testability, diagnosis and prognosis experts in order move a step further,
• i th d d t d t d d l i• improve methods, processes and means to reduce costs and delays in development, production and support phases,
• establish a network of excellence between customers, system designer, suppliers and research centers,
• create strong partnership between industries and universities and Research Centers
diag 21 is an exchange center
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
4
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
FOREWORD: Structure of the Association
GOVERNMENTAL 11Association
Board
LARGE COMPANIES
ORGANISATIONS 11
55 557
• President• Vice-
President• Secretary• Finance
G i
SIZED FIRM (ECC) SMESMALL AND MEDIUM-
SIZED FIRM (ECC) SME 33
11
COL
Finance• President
scientific committee
Governing Body
High Schools
Laboratories
11
2220
LLEG 5 Working
Scientific Committee
Universities 22
20members
+ the 5 members of
the association
GES
5 Working Groups
General Assembly
Individual Members 11 the association board
International Board(M Schieber)
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
(M. Schieber)
5
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
Standards Improvement
FOREWORD: Expected benefices
p
Methods and processes improvement in order to promote exchanges between different technical fields
Customer relationships / system designer / supplier development & improvement
Quality and diagnostic of failure relevance improvement
New services definition triggers new technical and i l i t teconomical interests
diag 21 a contributor of customer satisfaction improvement
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
6
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
To be in contact
Association diag21Association diag21 TELECOM ParisTech46, rue Barrault 75634 Paris Cedex 1375634 Paris Cedex 13 France
www.diag21.com
Chairman of diag21: J.M. LOUVET (MBDA)[email protected]
Responsible for international exchange: M. SCHIEBER (CASSIDIAN T&S)[email protected]
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
7
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
TUTORIAL OVERVIEWTUTORIAL OVERVIEW
From conventional considerations of Prognostics … tot id ti i h ith d i i kisystem considerations more in phase with decision making
requirements in industrial context …
Wh t t ll th ti i i f i d t i l … What are actually the prognostics vision from industrialviewpoint: dream, reality, something else ?
Towards a reality: Illustration in the case of industry anddefence sector (needs, techniques used, results obtained)
Towards a reality: International standardisation initiatives
Conclusions
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
8
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
TUTORIAL OVERVIEWTUTORIAL OVERVIEW
From conventional considerations of Prognostics … tot id ti i h ith d i i kisystem considerations more in phase with decision making
requirements in industrial context …
Wh t t ll th ti i i f i d t i l … What are actually the prognostics vision from industrialviewpoint: dream, reality, something else ?
Towards a reality: Illustration in the case of industry anddefence sector (needs, techniques used, results obtained)
Towards a reality: International standardisation initiatives
Conclusions
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
9
Prognostics … mainly component oriented
PHM definition
orientedde o
– Estimation of remaining life of a componentcomponent orsubsystemsubsystemyy
Prognostics evaluate the current health of a componentcomponent,and conditional on future load and environmental exposure,estimates at what time the componentcomponent (or subsystemsubsystem) willno longer operate within its stated specificationsP ti i th bilit t f li bl d ffi i tl Prognostic is the ability to perform a reliable and sufficientlyaccurate prediction of the remaining useful life of equipmentequipmentin service The primary function of the prognostic is to projectin service. The primary function of the prognostic is to projectthe current health state of equipmentequipment into the future takinginto account estimates of future usage profiles [Lebold et al.g p [2001]
10A. Saxena, "Prognostics: The science of Prediction", Tutorial, Annual Conference of the PHM Society (PHM2010), Portland, October 10 - 14, 2010
Prognostics … mainly component orientedoriented
Prognostics for Battery, Bearing, Engine …
11N. Scott Clements, "Introduction to Prognostics", Tutorial, Annual Conference of the PHM Society, Montreal, Canada September 25 – 29, 2011
Prognostics … mainly component oriented
Component instrumentation (mainly for ControlControl purposepurpose)
oriented
Components follow different rates and failures modes with behaviorvarying all its life cycle phases (impacting maintenance)
Component supports PHM processes to move to CBM by Component supports PHM processes to move to CBM byconsidering both InformationInformation && itsits supportsupport SoftwareSoftware applicationapplication
12
Prognostics … but the context is mainly system oriented (CBM+)
EngineeredSystem
System semantic deployment• Fleet of ships
mainly system oriented (CBM+)
shipShip
Land Vehicle
p• Ship Health Assessment • Fleet-wide expertise and case
sharing
pPassenger
ShipCargoShip
Ferries Cruise Ship Reefer Ship
Equipment
PrognosticsPrognosticsContextualizationContextualization
engineEngine Pump Propeller
Equipment semantic deployment
DieselEngine GasEngine ControllablePitch Propeller
Azimuth Thruster
• Fleet of engines• Engine Health Monitoring• Engine Predictive
Diagnostics
(Identical Similar(Identical SimilarPropulsionEngine
GeneratorEngine
13B. Iung, JB Léger, M. Monnin, “PHM Applications for fleet management: A summary”, PHM Applications Panel, IEEE PHM conference 2012, Denver, USA
(Identical, Similar,(Identical, Similar,Heterogeneous)Heterogeneous)
Prognostics … but the context is mainly system oriented (CBM+)mainly system oriented (CBM+)
Decision at “System” – “Fleet” levels … using Component PHM processes– Increasing the system control performance vs. services with regards to
ti l diti i i t t i i t (i il bilit )operational conditions, missions, strategic requirements (i.e. availability) …while increasing maintenance considerations (drift observation …)
– … Towards aa proactiveproactive visionvision ofof thethe systemsystem functioningfunctioning to make dynamicdecision and by considering the principle of fleet not only at system level butalso as “system-of-system” level (Population, Aggre./Consolidation)
14M Monnin, J-B Léger, D. Morel, Proactive facility fleet/plant monitoring and management, 24th International Congress on Condition Monitoring and Diagnostics Engineering Management, 2011.
Prognostics … as main process for supporting System Pro activity
• Evolution on the Maintenance practices: from conventional to Predictive
supporting System Pro-activity
pstrategies (Condition-Based + Prediction) … and finally Proactive one (E-maintenance) … because anticipation is required (opportunityopportunity spacespace)
15N. Scott Clements, "Introduction to Prognostics", Tutorial, Annual Conference of the PHM Society, Montreal, Canada September 25 – 29, 2011
Prognostics … as main process for supporting System Pro activity
IMS : Intelligent Manufacturing System
supporting System Pro-activity
Enterprise
g g y… Intelligent Maintenance System (E-Maintenance)
Main Change: « Component ispsystems
(ERP, MES, …)Services & Information flows Enterprise
systems
Main Change: « Component isrequesting maintenance action » (/itself;/product
deviation)Plant control system
Good flows
(ERP, MES, …)
Information flows PRODUCTflows
datamaterial data
information
Enterprisesystem
Enterprisesystem
Enterprisesystem
Enterprisesystem
Enterprisesystem
Enterprisesystem
deviation)
Plant controlsystem
Plant controlsystem
Plant controlsystem
Good flows flowsmaterial data
PRODUCT
ReconsideringA t ti
COMPONENTPRODUCT
Plant controlsystem
Plant controlsystem
AutomationParadigm …
Plant controlsystem
Plant controlsystem
Plant controlsystem
COMPONENT
CMMSsystem
Plant controlsystem
system
16
Prognostics … as main process for supporting System Pro activity
MOBILITY concept ... leading to offer new
supporting System Pro-activityp g
capacities for supporting maintenance actions– “on-line” diagnostic based on “mobile” sensorsg
(MEMS with wireless communication) – Plugand Play principleT l l li ti b d RFID T– Tools localisation based on RFID Tags ...
PROACTIVITY concept ... leading to maked i i b i t ti f tdecision by integrating new factors–– “on“on--line”line” prognosticsprognostics ...... toto PHMPHM
(C )– From Predictive Maintenance (CBM) toOpportunistic maintenance ... then to Pro-active Maintenance (LCC approach)active Maintenance (LCC approach)
17
Prognostics … as main process for supporting System Pro activity
Using e-technologies to increase maintenance efficiency velocity PROACTIVITYPROACTIVITY MOBILITYMOBILITY and
supporting System Pro-activity
... To support the moving from “fail andefficiency, velocity, PROACTIVITYPROACTIVITY, MOBILITYMOBILITY .. and
to optimise maintenance related workflow
fix” maintenance practices to “predictand prevent” strategies vs. Proactivityp g yconcept [Lee et al. 2006] [Iung, 2003][Blann, 2003]… while keeping[ , ] p gMaintenance as an Enterprise process(holistic approach) - Integration conceptCarrying out global system-of-interest objectives which ( pp ) g p(i.e. IEC/ISO 62264) for optimisingsystemsystem--ofof--interestinterest objectivesobjectives
y g g y jimposes to the maintenance area: Openness, Integration, Collaboration with the other e-enterprise services
18
yy jjA. Muller, A. Crespo Marquez, B. Iung, On the concept of e-maintenance. Review and current research, ReliabilityEngineering and System Safety (RESS). Volume 93, Issue 8, Pages 1165-1187, 2008
Prognostics … as main process for supporting System Pro activity
The maintenance action is done on the component …
supporting System Pro-activity
The maintenance action is done on the component …but the maintenance decision aims at maintaining orrestoring a Performance, a Service
ROI OEE LCC– ROI, OEE, LCC …– Quality, Productivity, Safety … (Metric of performances)
M i t PP t itt itMaintenance
Action Decision
ProPro--actvityactvity vs. vs. PerformancePerformance
Components degradations
SystemPerformances
HolonicHolonicViewView
degradations Performances
Usage conditionsEnvironment
19Levrat E., Iung B., A. Crespo Marquez, E-maintenance : Overview and Conceptual Framework ; Journal PP&C, specialissue on Maintenance and Facility Management organised by Pr. M. Garetti, Vol. 19, No. 4, June 2008, pp 1–22
Prognostics … within System Engineering consideration
System-of-Interest
Engineering consideration
System-of-Interest
requires a
Design system
requires a
Retirement systemsystem system
requires a
Support system
requires a
Manufacturing system
Maintenancesystem
Commercial system
y
requires aINCOSE Systems Engineering Handbook v. 3.2; ISO/IEC 15288:2008
20
Prognostics … within System Engineering consideration
• Evolution on the engineering practices: towards SystemSystem EngineeringEngineering
Engineering consideration
IINN
(Collaborative Dimension – System Thinking – Life Cycleconsideration)
• From component layer (action level) to system layer (strategy level)
To endow the system_of_interest with the it t b t d
NNTTEE
• From component layer (action level) to system layer (strategy level)
• New way for performance and requirement controlling (conventional ones… but also sustainable ones etc.)
System_of_InterestSystem_of_InterestEngineeringEngineering
capacity to be supported EEGGRR
System_ofInterest
EngineeringEngineering
Requirementsof support
Requirementsfrom support
AATTII (Integrated) Logistic(Integrated) Logistic
SupportSupportEngineeringEngineering
Support System
To endow the support system with the
IIOONN To endow the support system with the
capacity to maintain the system_of_interest in operational
conditions
NN
RELATED TO …RELATED TO … 21
Prognostics … within System Engineering considerationEngineering consideration
System_of_InterestSystem_of_Interest
SystemDefinition
SystemDesign &Development
SystemIntegration &Tests
Deployment&Transition
Operation &Support in
Operational ConditionsDismantling
d h h h bSupport Efficiency Safety
SystemSystem--ofof--InterestInterest
To endow the S_o_I with the capacity to be supported Reliability
Maintainability
Support Efficiency, Safety
System_ofInterest
EngineeringEngineering
Requirementsf t
Requirementsf
OperationalAvailability
Life
(Integrated) Logistic(Integrated) LogisticSupportSupport
Support Systemof support from support
To endow the support system with the
LifeCycleCost
Purchase cost SupportSupportEngineeringEngineering
System System System Deployment Operation &
To endow the support system with the capacity to maintain the S_o_I in
operational conditions
Purchase cost
Operation direct cost
Support cost
22
SystemDefinition
SystemDesign &Development
SystemIntegration &Tests
ep oy e t&Transition
Operation &Support in
Operational ConditionsDismantling
Support SystemSupport System
Prognostics … within System Engineering considerationEngineering consideration
"Needs and Requirements" ViewNeeds and Requirements View Purpose (Final Goal) = Why the system does exist ? Mission = What it does (transforms) ? Objectives = How many inputs it transforms? …
"Functional/Logical Architecture" View g A structure of functions that allows the system to perform all the identified operational
scenarios throughout its life-cycle Includes exchanged flows between functions and external world (interface)g ( )
“Organic/Physical Architecture" View A set of concrete components that support both functions and interactions between the A set of concrete components that support both functions and interactions between the
components Includes physical connexions
23
M Monnin, A Voisin, J-B Léger, B Iung, Fleet-wide health management architecture, Annual Conference ofthe Prognostics and Health Management Society, PHM 2011, Montreal, Canada, 2011
Also for the prognostics engineering …
Prognostics … within System Engineering considerationEngineering consideration
To endow the system_of_interest with the Support Efficiency Safety
System_of_InterestSystem_of_InterestE i iE i i
capacity to be supported Reliability
Maintainability
Support Efficiency, Safety
System_ofInterest
EngineeringEngineering
Requirementsof support
Requirementsfrom support
OperationalAvailability
Life
(Integrated) Logistic(Integrated) LogisticSupportSupport
EngineeringEngineering
Support Systempp ppLife
CycleCost
Purchase costTo endow the support system with the
capacity to maintain the system_of_interest in operational
conditions
Purchase cost
Operation direct cost
Support costTo endow the maintenance To endow the support system
MaintenanceMaintenanceE i iE i i
Prognostics
system with the capacity to deploy proactive strategies
pp ywith the capacity to anticipate degradation on the S_O_I
24
EngineeringEngineering Engineering?A. Muller, M-C. Suhner, B. Iung, Proactive Maintenance for industrial system operation based on a formalised prognosisprocess, Reliability Engineering and System Safety (RESS) – Volume 93, Issue 2, February 2008, Pages 234-253
Prognostics … within System Engineering consideration
Conventional Prognostics ClassificationsConventional Prognostics Classifications
Engineering consideration
based on forecastingforecasting techniques techniques …based on forecastingforecasting techniques techniques …
25
Prognostics … within System Engineering considerationEngineering consideration
The criteria expected from the prognosticsf
p p g Taking into account the future scenario Knowledge use for the modelling and operation phases
ScientificPapers on
Prognostics
… towardsP ti… towardsP ti Between 5 and 10
More than 10PERFORMANCE
Prognostics
PrognosticsClassificationbased on systemsystem
PrognosticsClassificationbased on systemsystem
Between 5 and 10
Between 2 and 4
SystemReliabilitybased on system system
considerationsconsiderationsbased on system system considerationsconsiderations
1
NothingComponent
y
Maintenance Scenario
No scenarioUsage Scenario
ComponentReliability
26
I. Geanta, D. Theilliol, B. Iung, M. Schieber, Y. Fusero, « Multi-criteria Cartography Investigation on Diagnostics andPrognostics Techniques Suited for System and Vehicle Health Maintenance « , Conference IEEE AutoTestCon, Anaheim,CAU.S.A, 2012
Prognosticsformalization
Current situation of the
formalization
tt Current situation of the
production system
1
tDiachronicDiachronic
C1C1
C4
DiachronicDiachronicVisionVision
C4
Performance P2
Performance P2
0 480,150,1
Composant 4
0 480,15Degraded 10,1Nominal state
Component 4
Degraded 2
SynchronicSynchronicVisionVision
C2
C3
Performance P
C2C3
Performance
0,270,480,270,48Degraded 2
Failed state
P1 P1
Maintenance
0,18Non Acceptable0,82Acceptable
Performance 1
0,18Non Acceptable0,82Acceptable
Performance 1
3
22 Assessment of the future performances
MaintenanceAlternative
27
Prognosticsformalization
Predict future system
formalizationy
health: Components’
diti ANDconditions ANDperformances
Remaining Usefullives (RULs) of thesystem, sub-system orcomponents
Associated confidencelevel
28
level
Voisin A., Levrat E., Cocheteux P., Iung B., Generic prognosis model for proactive maintenance decision support: application to pre-industrial e-maintenance test-bed; International Journal of Intelligent Manufacturing. Vol. 21, 2, pp 177-193 April 2010
Prognosticsformalizationformalization
G i S b F tiGeneric Sub-Functions
29
PrognosticsFormalization at system level
Maintenance
Formalization … at system level
Maintenance
Action Decision
Components degradations
SystemPerformances
System multiSystem multi--layerslayers/multi /multi Usage conditions
Environment componentscomponents
Sys
tem
Laye
r System
Function
Input Flow (IF) Output
Flow (OF)t0
Component
IFSub-system
FunctionOFIF
ComponentC tIF
Sub-system
FunctionOFS
ubS
yste
mLa
yer
t0t0
30
IF
Component
ComponentFunction OF
Component
ComponentFunction OFIF
Component
ComponentFunction OF
IF
30Com
pone
ntLa
yer
t0t0
t0
PrognosticsFormalization at system level
Degradation-Failure-Flow Analysis
Selection of the i di t d
1 2
Formalization … at system level
indicators and the architecture
« Layered » A hit t
3Architecture
Cumulated Architecture 4
5
Knowledge
31
PrognosticsFormalization at system levelFormalization … at system level
Functional analysis Functional analysis– Process approach
Flow based performance assessment– Flow based performance assessment– Causality (physical laws)
Dysfunctional analysis Dysfunctional analysis– Component degradation
Flow deviation– Flow deviation Event-Driven Indicators I f h i (d d ti ti ) Inference mechanisms (degradation propagation)
– Causal (in time throughout components)T l (th h t ti )
32
– Temporal (throughout time)
32
PrognosticsFormalization at system level
System performance has to modeled in a generic wayf f f
Formalization … at system level
– Performances-Requirements definition: performances represent thefinality fulfillment (objectives) which are modeled through outputoutput flowsflows. Performance indicator: output flow’s properties
Motor
– Functional point of view: multi-level hierarchical decomposition
To accumulatestrip
electric power supply
Accumulated strip System
To carryi
rotation stripStrip on bobbin
Accumulated strip
Accumulation unit
System performance
rotationBelt
rotationPulley-belt
To produce rotation
electric power supply Motor
rotation
System performance
Sub-system /Component performancePulley belt
system • Torque• Rotation speed
To carry rotation
Belt rotation
AC Motor
Pulley-belt
performance
To pull stripStrip on bobbin
Accumulated strip
system
Roller
PrognosticsFormalization at system level
• Dysfunctional point of view : performance loss arises from
Formalization … at system level
the deviation of output flow properties• Dysfunctional causality modeled by generic causal relations. They model
the causes of output flow deviation:the causes of output flow deviation:• The deviation of an input flow property• The support degradation
DDD PP• Knowledge from particularizing relationships extracted from FMECA and
HAZOP analyses
flowOutputcomponentflowInput DevDegopDev PropPr
Belt rotationBelt rotationHAZOP analyses
NO Angular Velocity (AC motor rotation)NO
CauseDeviationAttribute
Belt rotation
NO Angular Velocity (AC motor rotation)NO
CauseDeviationAttribute
Belt rotation
lengthening (belt)
LESS Angular Velocity (AC motor rotation)LESS
Slip (belt)NO
Angular Velocity(Wob) lengthening (belt)
LESS Angular Velocity (AC motor rotation)LESS
Slip (belt)NO
Angular Velocity(Wob)
MORE Angular Velocity (AC motor rotation)MORE
lengthening (belt)
MORE Angular Velocity (AC motor rotation)MORE
lengthening (belt)
34Léger J-B., Iung B.; Methodological approach to modelling of degradation detection and failure diagnosis in complexproduction system. In: Proceedings of the 9th workshop on Principles of Diagnosis DX98,, pp. 209-216, May 24-27, 1998,Cap Cod, Massachusetts, USA
Prognostics Execution… at system level: test case for deploymentlevel: test case for deployment
35Iung B, Levrat E, Crespo Marquez A, Erbe H. Conceptual framework for e-maintenance: Illustration by e-maintenance technologies and platform. Annual Review in Control, vol.33, issue 2, pp220-229, 2009
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
TUTORIAL OVERVIEWTUTORIAL OVERVIEW
From conventional considerations of Prognostics to From conventional considerations of Prognostics … tosystem considerations more in phase with decision makingrequirements in industrial context …
… What are actually the prognostics vision from industrialviewpoint: dream, reality, something else ?
Towards a reality: Illustration in the case of industry anddefence sector (needs, techniques used, results obtained)
Towards a reality: International standardisation initiatives
Conclusions Conclusions
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
36
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
DIAG 21 Industrial Survey on Prognostics: Context and objectivey g j DIAG21 Prognostics Working Group objective
To identify the industrial considerations (i.e. needs, issues,challenges, interests) on Prognostics
To compare the academic Prognostics vision with the industrial one inaccordance to the previous considerationsaccordance to the previous considerations
To contribute by a joint industrial – academic work to the formalisationand engineering of the Prognostics to face with considerations
DIAG21 Prognostics WG main statement Prognostics is understood differently by industry and academia P ti i “ t fi ld” ( t ll k ) f t f Prognostics is a “emergent field” (not well known) for most of
industrial people
DIAG21 Prognostics WG first challenge DIAG21 Prognostics WG first challenge To collect the industrial vision on Prognostics (considerations-based) To develop a “white paper”
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
To promote the Prognostics in industry
37
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
DIAG 21 Industrial Survey on Prognostics: Context and objectivey g j Development of Interviews based on a specific questionnaire
What are the objectives of the questionnaire but also of the interview? What are the professional profile of the interviewee industrial? What are the main issues – challenges related to the application
domain in which the interviewee is working on (but also the maindomain in which the interviewee is working on (but also the maincriterion characterising these issues - challenges) ?
What are the levers for change but also the areas of progress inrelation these issues – challenges ?
How the interviewee industrial is considering Prognostics with regardsto these issues – challenges ?to these issues challenges ? Is it already implemented ?
Yes : Prognostics definition, Prognostics added-value ...g g No: Why ? What are the obstacles? ...
What the interviewee is expecting from Prognostics academicit ?
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
community ?
38
Tutorial 3 : State-of-the-Art Realization of Prognostics in IndustryDIAG 21 Industrial Survey on Prognostics: Who are provided fulfilledy g pquestionnaires in a first step of procedure ? About 20 answers directly from individual interviews
About 15 answers from exchanges issued of the industrial panelon Prognostics developed during DIAG21 Forum (November 2011in Paris)in Paris)
24%4%5% 5%
Transport
Prestataire
Transport
Service
19%Energie
Aéronautique/Spatial
Nano/Micro Electronique
Energy
Aeronautics/Spatial
Electronics
24%
19%
Naval
Fabricant de Machines
Naval
Machine Tools
Industrial representative of Design, Manufacturing, Integration,Operation, Maintenance areas
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
39
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
DIAG 21 Industrial Survey on Prognostics: What are the main reports?DIAG 21 Industrial Survey on Prognostics: What are the main reports?
Industrial Prognostics vision is not totally in phase with thescientific one (i.e. Prognostics is an item within a decision loopsuch as PHM, CBM+ not a finality by itself)
Only some of the interviewees use Prognostics capacities. Forthe others, it is a long-term project, a dream or a danger (withregards to the investment the benefits the ROI)regards to the investment, the benefits, the ROI)
Some industrial refused to communicate on this subject
Prognostics development thinking is similar to the historical wayof deciding (i.e. the difficulty) to evolve from corrective topreventive maintenancepreventive maintenance
Prognostics is often associated to the issues – challenges relatedto SERVICES
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
to SERVICES
40
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
DIAG 21 Industrial Survey on Prognostics: What are the main reports?DIAG 21 Industrial Survey on Prognostics: What are the main reports? Services with regards to: Product/System Life Cycle consideration Prognostics capacities could be integrated from Design
phase (PHM Engineering) Prognostics can take advantage of the Return of Experience Prognostics can take advantage of the Return of Experience
data Prognostics can serve to prove delivery of expected
f i l f ti iperformances in normal functioning Prognostics can impact the product quality by mastering the
component degradation and its evolution ...p g Circular Economy paradigm Economical considerations Prognostics capacities to support maintenance cost control Prognostics capacities to support maintenance cost control Prognostics metrics to assess its impact on CoO, OEE, ROI
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
41
Tutorial 3 : State-of-the-Art Realization of Prognostics in IndustryDIAG 21 Industrial Survey on Prognostics: What are the prognostics key y g p g yvision with regards to industrial requirements ? Component availability vision: Prognostics can be a support to
anticipate a component failure by controlling its actual andp p y gfuture degradation according to its “operational context”
System productivity vision: Prognostics can be a support toy p y g ppadapt the maintenance planning (dynamic change) with regardsnot only of the components availability but also of the globalrequirements/performances expectedrequirements/performances expected
Mission vision
Safety-Risk vision Safety Risk vision
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
42
Tutorial 3 : State-of-the-Art Realization of Prognostics in IndustryDIAG 21 Industrial Survey on Prognostics: What are the prognostics key y g p g yvision with regards to industrial requirements ?
Operation - Business vision Operation Business vision
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
43
Tutorial 3 : State-of-the-Art Realization of Prognostics in IndustryDIAG 21 Industrial Survey on Prognostics: What are the facts and obstacles toy gface with for Prognostics deployment? Prognostics in general is based on the RUL principle ... but it is
very different in relation to the technologies considered due tovery different in relation to the technologies considered due tothe degradation to be observed (i.e. Prognostics restricted towell known components such as Engine, Bearings)
Prognostics starts from the current situation (i.e. Indicators)leading to consider Condition Monitoring and HealthManagement as two main up stream processesManagement as two main up-stream processes Relevance of the CM, HM, HUMS systems ? Consideration
of these systems in the whole engineering.of these systems in the whole engineering. Sources of knowledge in general
Indicators definition? Information system standardisationyand interoperability (i.e. OSA/CBM)
How the indicators aggregation is solved to supply “currenti i ” diff b i l l ?
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
situation” at different abstraction levels?
44
Tutorial 3 : State-of-the-Art Realization of Prognostics in IndustryDIAG 21 Industrial Survey on Prognostics: What are the facts and obstacles toy gface with for Prognostics deployment? Prognostics and uncertainties on the input data, the results
(degree of confidence) What is the forecasting horizon well(degree of confidence). What is the forecasting horizon welladapted to minimise the uncertainties (i.e. 1 hour, 1 day)? Is itrealistic to consider Prognostics at system level taking intoaccount the sum of all the uncertainties at each level?
The importance of the forecasting model (method) to be usedand depending on the system data, the system knowledge, thecontext data, ... Is the model selected the most appropriatedtaking into account the available data the technologies thetaking into account the available data, the technologies, thehorizon ...? Verification/Validation phases of the model
Prognostics engineering in general ? Without integration of Prognostics engineering in general ? Without integration ofPrognostics at design phase, the Prognostics added value isvery limited (towards a System thinking based PHM
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
Engineering)
45
Tutorial 3 : State-of-the-Art Realization of Prognostics in IndustryDIAG 21 Industrial Survey on Prognostics: What are the facts and obstacles toy gface with for Prognostics deployment?
Prognostics and metrics [A. Saxena, 2008] required to be ableh i f P i l “i l”to assess the impact of Prognostics not only on “internal”
considerations but also on “external” ones (i.e. ROI, OEE,maintenance cost decreasing) in relation to temporal scalemaintenance cost decreasing) in relation to temporal scale
Prognostics deployment requires to use sometimes emergentICT ... which are not yet necessary reliableICT ... which are not yet necessary reliable
Not to be intrusive (i.e. certification issue for IVHM)
Effort quantification (on all the Prognostics life cycle)
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
46
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
TUTORIAL OVERVIEWTUTORIAL OVERVIEW
From conventional considerations of Prognostics to From conventional considerations of Prognostics … tosystem considerations more in phase with decision makingrequirements in industrial context …
… What are actually the prognostics vision from industrialviewpoint: dream, reality, something else ?
Towards a reality: Illustration in the case of industry anddefence sector (needs, techniques used, results obtained)
Towards a reality: International standardisation initiatives
Conclusions Conclusions
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
47
Prognostics in Industry
out
its
agre
emen
tou
t it
s ag
reem
ent
Process Industry Oil refinery S l i d
Manufacturing Milling machine i
ed o
r tr
ansm
itte
d w
itho
ed o
r tr
ansm
itte
d w
itho Steel industry
Chemical Pulp & paper
Automotive manufacturing
Nuclear fuel manufacturing
ty a
nd m
ust
not
be u
sed
ty a
nd m
ust
not
be u
sed
Nuclear fuel recycling …
Power Energy
manufacturing …
Defense Frigates aircraft
nt is
PR
EDIC
T’s
prop
erty
nt is
PR
EDIC
T’s
prop
erty
gy Hydropower plant Gas turbine/steam
turbine
Frigates, aircraft carriers…
Ground vehicle UAV
s co
nfid
enti
al d
ocum
ens
conf
iden
tial
doc
umen Nuclear power plant
Wind farm …
UAV …
…
© T
his
© T
his
48PHM Europe 2012 - Dresden - 03/07/2012
Why Prognostic?
out
its
agre
emen
tou
t it
s ag
reem
ent
For all industry Anticipate failure and breakdown
Increase uptime Dynamic planning for maintenance
ed o
r tr
ansm
itte
d w
itho
ed o
r tr
ansm
itte
d w
itho
y p g Decrease maintenance costs
Decrease energy consumptionIn Process industry
Anticipate performances and product quality Increase competitiveness
ty a
nd m
ust
not
be u
sed
ty a
nd m
ust
not
be u
sedp
In Power energy Anticipate 2 years main shutdown/outage
Maintenance optimisation Anticipate safety impact
Decrease nuclear risk
nt is
PR
EDIC
T’s
prop
erty
nt is
PR
EDIC
T’s
prop
erty
In Manufacturing Anticipate decrease of product quality
Decrease waste or product defect Optimal change of tools
Decrease product costs
s co
nfid
enti
al d
ocum
ens
conf
iden
tial
doc
umenIn Defence
Anticipate mission impact Decrease mission abort Increase survivability
© T
his
© T
his
49PHM Europe 2012 - Dresden - 03/07/2012
Prognostics based on effect impacts
out
its
agre
emen
tou
t it
s ag
reem
ent
ed o
r tr
ansm
itte
d w
itho
ed o
r tr
ansm
itte
d w
itho
Identification of
ty a
nd m
ust
not
be u
sed
ty a
nd m
ust
not
be u
sedIdentification of
consequences from an event
nt is
PR
EDIC
T’s
prop
erty
nt is
PR
EDIC
T’s
prop
ertyfrom an event
s co
nfid
enti
al d
ocum
ens
conf
iden
tial
doc
umen
Applicable to all Industry
© T
his
© T
his
50PHM Europe 2012 - Dresden - 03/07/2012
pp y
Prognostics based on Trend
out
its
agre
emen
tou
t it
s ag
reem
entHealth
indicator
ed o
r tr
ansm
itte
d w
itho
ed o
r tr
ansm
itte
d w
itho
Trend analysis
ty a
nd m
ust
not
be u
sed
ty a
nd m
ust
not
be u
sed
Date before next f il
nt is
PR
EDIC
T’s
prop
erty
nt is
PR
EDIC
T’s
prop
ertyfailure
s co
nfid
enti
al d
ocum
ens
conf
iden
tial
doc
umen
Applicable to Process parts of any Industry
© T
his
© T
his
51PHM Europe 2012 - Dresden - 03/07/2012
pp p y y
Prognostic based on multi-trends
out
its
agre
emen
tou
t it
s ag
reem
ent
ed o
r tr
ansm
itte
d w
itho
ed o
r tr
ansm
itte
d w
itho
ty a
nd m
ust
not
be u
sed
ty a
nd m
ust
not
be u
sed
nt is
PR
EDIC
T’s
prop
erty
nt is
PR
EDIC
T’s
prop
erty
s co
nfid
enti
al d
ocum
ens
conf
iden
tial
doc
umen
Applicable to Process parts of any Industry
© T
his
© T
his
52PHM Europe 2012 - Dresden - 03/07/2012
pp p y y
Empirical prognostic
out
its
agre
emen
tou
t it
s ag
reem
ent
Based on empirical law and table
ed o
r tr
ansm
itte
d w
itho
ed o
r tr
ansm
itte
d w
itho
Empirical law
ty a
nd m
ust
not
be u
sed
ty a
nd m
ust
not
be u
sedp
and curve
nt is
PR
EDIC
T’s
prop
erty
nt is
PR
EDIC
T’s
prop
erty
s co
nfid
enti
al d
ocum
ens
conf
iden
tial
doc
umen
Applicable to Industry having long time record t l ti i t l t t
© T
his
© T
his
53PHM Europe 2012 - Dresden - 03/07/2012
system or long time experimental tests
Prognostic based on conditional probability Failures and states
simulated from i d l d
out
its
agre
emen
tou
t it
s ag
reem
ent
experience model and real data
ed o
r tr
ansm
itte
d w
itho
ed o
r tr
ansm
itte
d w
itho
ty a
nd m
ust
not
be u
sed
ty a
nd m
ust
not
be u
sed
nt is
PR
EDIC
T’s
prop
erty
nt is
PR
EDIC
T’s
prop
erty
Experience Model
Real Data
s co
nfid
enti
al d
ocum
ens
conf
iden
tial
doc
umen
Real Data
Applicable to Industry having long experience or fl t id t
© T
his
© T
his
54PHM Europe 2012 - Dresden - 03/07/2012
fleet-wide systems
Systemic PrognosticTaking into
out
its
agre
emen
tou
t it
s ag
reem
ent
Taking into account missions and environment
ed o
r tr
ansm
itte
d w
itho
ed o
r tr
ansm
itte
d w
itho
System approach of
ty a
nd m
ust
not
be u
sed
ty a
nd m
ust
not
be u
sedpp
prognostics
nt is
PR
EDIC
T’s
prop
erty
nt is
PR
EDIC
T’s
prop
erty
Multi-scenariiPrognostic
Prognostic adjustment contin o ls based on eal
s co
nfid
enti
al d
ocum
ens
conf
iden
tial
doc
umencontinuoulsy based on real
data
Applicable to Defense, Maritime and Wind Offshore
© T
his
© T
his
55PHM Europe 2012 - Dresden - 03/07/2012
pp ,
Prognostic for ageing mngt. Big Picture with CASSIDIAN T&S
out
its
agre
emen
tou
t it
s ag
reem
ent
Big Picture with CASSIDIAN T&S
ed o
r tr
ansm
itte
d w
itho
ed o
r tr
ansm
itte
d w
itho
ty a
nd m
ust
not
be u
sed
ty a
nd m
ust
not
be u
sed
Prognostic
nt is
PR
EDIC
T’s
prop
erty
nt is
PR
EDIC
T’s
prop
erty
s co
nfid
enti
al d
ocum
ens
conf
iden
tial
doc
umen
Applicable to Industry having long life-cycle
© T
his
© T
his
56PHM Europe 2012 - Dresden - 03/07/2012
pp y g g y
Some examples
out
its
agre
emen
tou
t it
s ag
reem
ent
Nuclear Fuel Recycling Prognostic of bearing failure on a rotational furnace based on speed analysisPrognostic: 6 months
ed o
r tr
ansm
itte
d w
itho
ed o
r tr
ansm
itte
d w
itho
Nuclear Fuel Manufacturing Prognostic of motor failure on a milling machine based on current analysis
Prognostic: 3 months
ty a
nd m
ust
not
be u
sed
ty a
nd m
ust
not
be u
sed
Nuclear Power Energy Prognostic of pump failure on a coolant system based on pressure and flow-
Prognostic: 3 months
Prognostic: 3 months
nt is
PR
EDIC
T’s
prop
erty
nt is
PR
EDIC
T’s
prop
erty
rate analysisNavy
Prognostic of diesel engine failures on embedded power plant and on propulsion system based on stress real-time evaluation
Prognostic: 3 months
Prognostic : 2 months
s co
nfid
enti
al d
ocum
ens
conf
iden
tial
doc
umen
p p yPower Energy
Prognostic of main condensers fouling on a sea water cooling system based on pressure/flowrate balance driftPrognostic : 1 month
© T
his
© T
his
57PHM Europe 2012 - Dresden - 03/07/2012
Main pre-requirements
out
its
agre
emen
tou
t it
s ag
reem
ent
Prognostics based on effect impacts System approach Dysfunctional analysis
FMEA, HAZOP, FTA, RCA…
ed o
r tr
ansm
itte
d w
itho
ed o
r tr
ansm
itte
d w
itho
, O , , CPrognostics based on trend analysis
Identification of relevant indicators Indicator isn’t direct measure from sensors
Data cleaning and conditioning Reconciliation of data when sensor faults
ty a
nd m
ust
not
be u
sed
ty a
nd m
ust
not
be u
sed
Empirical prognostics Long experience or experimentation
Prognostics based on conditional probabilities System approach Dysfunctional analysis
nt is
PR
EDIC
T’s
prop
erty
nt is
PR
EDIC
T’s
prop
erty Dysfunctional analysis
Long experience with statistics coupled with real dataSystemic prognostics
Mission modelling System approach Experience on impact factors and corroboration
s co
nfid
enti
al d
ocum
ens
conf
iden
tial
doc
umen Experience on impact factors and corroboration
Prognostics for ageing management Ageing modelling Long experience or experimentation Continuous improvement
© T
his
© T
his
58PHM Europe 2012 - Dresden - 03/07/2012
Prognostics ROI
out
its
agre
emen
tou
t it
s ag
reem
ent
Return on investment: 10 timesBenefits starting between 3 to 6 months after solutions commissioning
ed o
r tr
ansm
itte
d w
itho
ed o
r tr
ansm
itte
d w
ithoReduction in maintenance costs: 25% to 30%
Proactive maintenance program can provide a cost saving of 8% to 12% over a program utilizing preventive maintenance strategiesCost saving opportunities of 30% to 40% could easily be realized
ty a
nd m
ust
not
be u
sed
ty a
nd m
ust
not
be u
sedCost saving opportunities of 30% to 40% could easily be realized
regarding a reactive maintenance approach and material condition.Elimination of breakdowns: 70% to 75%Reduction in downtime: 35% to 45%
nt is
PR
EDIC
T’s
prop
erty
nt is
PR
EDIC
T’s
prop
erty
Increase in production: 20% to 25%Maintenance planning postponed from 30% to 40% -> only 3
maintenance cycles to save one
s co
nfid
enti
al d
ocum
ens
conf
iden
tial
doc
umen
•93% OEE•97% plant throughput
•86% OEE•91% plant throughput
•67% OEE•74% plant throughput
© T
his
© T
his
59PHM Europe 2012 - Dresden - 03/07/2012
p g p•3% downtime
p g p•13% downtime
p g p•34% asset downtime
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
TUTORIAL OVERVIEWTUTORIAL OVERVIEW
From conventional considerations of Prognostics to From conventional considerations of Prognostics … tosystem considerations more in phase with decision makingrequirements in industrial context …
… What are actually the prognostics vision from industrialviewpoint: dream, reality, something else ?
Towards a reality: Illustration in the case of industry anddefence sector (needs, techniques used, results obtained)
Towards a reality: International standardisation initiatives
Conclusions Conclusions
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
60
Tutorial 3 : State-of-the-Art Realization of Prognostics in IndustryTowards a reality: International standardisation initiativesyActual state
The fault detection calculus methods are multiples and each new standards bring some new ( 1 22)ways to calculate (ex: IEEE 1522),
Their used is still base on a fuzzy definition of the faults catalog which is the base of the proposed calculus.
No or not precise fault catalog based on function or Failure modes
The fault catalog is defined between the supplier and the client (No standards)
The notion of fault catalog for software is not existing or not provenThe notion of fault catalog for software is not existing or not proven
Many (or too many >20) norms standard industrial and military which concern testability. Most of them are not applicable or obsolete
MIL HDBK 472 (1984)
IEEE 1522 (2004): remove 2010
Some results show that the calculated results of isolation and localization of faults areSome results show that the calculated results of isolation and localization of faults are different from the real found in field during return of experience. Even more the results attached to the embedded fault isolation algorithms bring uncertainty of isolation even during support and maintenance.
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
61
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
Availability Test DefinitionDiagnostic Development Safety
Reliability Requirements
To reduce the
Maintainability RequirementsTo Reduce
Safety Requirements
To limit isk
The missing link
ReliabilityFault Detection TestTo reduce
R i Ti /C t
number of failures Down Time
FunctionalR d d
Tests, BIT, BITE, Monitoring
To limit risk
ab yFault Isolation TestRepair Time/Cost
Testability
Redundancyg
effectiveness
DiagnosabilityThe Test and Safety AnalysisProcess
Safety AnalysisMaintainability
Maintenance
Fault Coverage SeverityC iti lit
SAFETYPreventive Corrective Predictive
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
Detection Rate Isolation RateCriticality
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
Actual engineering processActual engineering processNo validation milestone.
R i t
System Definition
Design Development
Requirement DefinitionConcept
Only one validation task
Validation
Integration
Demonstration
Test Definition Diagnostic Development
Maintenance Procedures
Integrated Diagnostic
System FunctionalTest
FD = 95%FI= 90% T1 Faults Injection
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
PreliminaryFMECA
CriticalFMECA
Requirement AllocationFD + FI prediction
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
Model driven Engineering process
R i t 5 * T t bilit
Model driven Engineering process
System Definition
Design Development
Requirement DefinitionConcept
ConceptA hit t
5 * Testability validation task
Validation
g p
Integration
ArchitectureValidationTestabilityDiagnosability
Demonstration
Analysis BITE Softwareversus Hard.
Test Definitionsbased on Failure Modes
Test Definitionsbased on Function
Demonstration
Diagnostic Development Test Strategy (Iteration)
Maintenance ProceduresMaintenanceRequirements
Modeling Tests Modeling with Test. IntegratedDiagnosticModeling with Test System Functional
Tests.
R i t
FunctionalModeling Fault Injection
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
CriticalFMECA
PreliminaryFMECA
Requirement AllocationsFD + FI prediction
FD/FIDemonstration
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
Standards Support InteroperabilityStandards Support Interoperability
• ISO Standards• ISO 13374-1 Condition Monitoring and diagnostics of machines –
Data processing, communication and presentation – Part 1: General guidelines (OSACBM-based)
• ISO 13374-2 Condition monitoring and diagnostics of machines –Data processing, communication and presentation – Part 2: Data processing
• PLCS (Product Life Cycle Support) – An ISO standard being developed within the STEP community focusing on information exchange for the life cycle support
• OASI PLCS DEX on Aviation Maintenance
• Historical Maintenance activities (similar to MAI)• Non maintenance activities that may affect future maintenancey• Estimation of system state• Activities affecting product inventories
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
Reference Diagnostic Modeling and Application Development (John Sheppard Montana State University, TIM Wilmering, The Boeing Company)
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
Standards Support InteroperabilityStandards Support Interoperability
IEEE Standards SCC20 DMCIEEE P1232 Artificial Intelligence Exchange and Tie to All Test
Environments (AI-ESTATE)
IEEE P1636 Software Interface for Maintenance Information Collection and Analysis (SIMICA)
IEEE P1636.1 Exchanging Test Results and Session Information via the Extensible Markup Languuage (XML)p g g ( )
IEEE 1636.2 Exchanging Maintenance Action Information via the Extensible Markup Language (XML)
Machinery Information Management Open System Alliance (MIMOSA)
OSA-EAI V3.2.1..OSA EAI V3.2.1..
OSA-CBM V3.2.1 ..
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
Reference Diagnostic Modeling and Application Development (John Sheppard Montana State University, TIM Wilmering, The Boeing Company)
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
StandardsStandards
Ministry of Defence - Defence Standard 25-24 Health and Usage Monitoring Capability for Land Platforms (HUMS) Issue 1 PublicationMonitoring Capability for Land Platforms (HUMS) Issue 1 Publication Date 24 December 2004
MoD STAN 23-09 Generic Vehicle Architecture
Working group SAE HM-1 g g p• Design (Andy Hess, Tim Wilmering)
• Architecture (Emmanuel Nwadiogbu, Sonia Vohnout)
• V&V, Certification (Ravi Rajamani)
• Data & Information Management (Bob Hess)
HM-1 Design Processes (Wilmering Timothy J timothy j wilmering@boeing com)HM-1 Design Processes (Wilmering, Timothy J [email protected])
VHM design
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
Reference Diagnostic Modeling and Application Development (John Sheppard Montana State University, TIM Wilmering, The Boeing Company)
Tutorial 3 : State-of-the-Art Realization of Prognostics in IndustryASD governance structure for e-Business Standardsg
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
EADS PLM standardization November 4st 2010
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
ASD NetworkASD Network
S5000FOperational and maintenance data feedback
LogisticSupport
s an
d en
t Orderadmini-Provisioning
dataLogs matand dataPro-
i i iOperational
dataLSAdata
S3000LS2000M
Analysisactivities
syst
ems
equi
pme
In ser-vice
stration
IP DataSubsets
visioning
Equipmentidentificationd i d t S3000L
Des
ign
of
supp
ort vice
use
TechnicalPubli-
IETM, other media
design data
S4000M
S1000D
D Publi-cationsDesign
data
S4000M
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
Source: EADS - Saab
© 2010 Saab AB
69
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
TUTORIAL OVERVIEWTUTORIAL OVERVIEW
From conventional considerations of Prognostics to From conventional considerations of Prognostics … tosystem considerations more in phase with decision makingrequirements in industrial context …
… What are actually the prognostics vision from industrialviewpoint: dream, reality, something else ?
Towards a reality: Illustration in the case of process industryand defence sector (needs, techniques used, resultsobtained, …)
Towards a reality: International standardisation initiativesy
Conclusions
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
70
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
CONCLUSIONCONCLUSION• Prognostics – A emergent process within PHM context … but with
potential great added value
• Prognostics – Towards its required integration within decision-making loop … thus considering not only the component level butalso the eq i ements/se i es/pe fo man es one (CBM+ ision)also the requirements/services/performances one (CBM+ vision)
• Prognostics – System Engineering thinking … thus PHMEngineering has to be considered within System EngineeringEngineering has to be considered within System Engineering
• A lot of issues and challenges
• Uncertainty Management and Degree of Confidence• Uncertainty Management and Degree of Confidence
• Performance assessment with regards to Prognostics metrics
• D d bl ICT t t P ti d l t• Dependable ICT to support Prognostics deployment
• Synergy between industry and academia to develop methods,models tools
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.
models, tools …
71
Tutorial 3 : State-of-the-Art Realization of Prognostics in Industry
Tutorial 3Tutorial 3State-of-the-art Realization of Prognostics in Industry
Presenters: Michel SCHIEBER (Cassidian T&S)Presenters: Michel SCHIEBER (Cassidian T&S)Jean Baptiste LEGER (Predict)Benoit IUNG (Lorraine University, CRAN)
Ce document est la propriété de l’association diag 21. Il ne peut être communiqué à des tiers et /ou reproduit sans une autorisation préalable écrite et son contenu ne peut être divulgué. diag 21 2009.This document and the information contained herein is proprietary information of “diag 21 partnership” and shall not be disclosed or reproduced without the prior an authorization. diag 21 2009.