Felix Rinker 1,2 Kristof Meixner 1,2 Sebastian Kropatschek 3 Elmar Kiesling 4 Stefan Biffl 1,3 1 ISE TU Wien 2 CDL SQI TU Wien 3 CDP Wien 4 IDPKM WU Wien 5 OvGU Magdeburg

1. Risk and Engineering Knowledge Integration
in Cyber-physical Production Systems Engineering
Felix Rinker1,2 Kristof Meixner1,2 Sebastian Kropatschek3
Elmar Kiesling4 Stefan Biffl1,3
1ISE TU Wien / 2CDL SQI TU Wien
3CDP Wien / 4IDPKM WU Wien
5OvGU Magdeburg

2. Context - Cyber-Physical Production Systems
2
Cyber-Physical Production Systems (CPPS) Engineering
• is the design of CPPSs for particular customers
• according to VDI 2206 involves
engineers of various domains e.g.,
▪ Mechanical Engineering
▪ Electrical Engineering
▪ Automation Engineering

3. Context - Heterogeneous Engineering Data Exchange
▪ Engineering team designs a
complex system model that
can consist of 10.000
instances
▪ Engineers provide their
engineering artifacts coming
from heterogeneous sources
▪ Boundary objects are hard to
manage automatically
▪ Artifact changes are managed
manually
Boundary objects (e.g. Screwdriver) are information used in different ways
in different communities [1]
[1] Star, Susan; Griesemer, James (1989). "Institutional Ecology, 'Translations' and Boundary Objects:
Amateurs and Professionals in Berkeley's Museum of Vertebrate Zoology, 1907-39". Social Studies of
Science. 19 (3): 387–420.
3
Basic Planner
Electrical
Engineer
Automation Engineer
Mechanical
Engineer
Quality
Engineer
Screwdriver
Electric Screwdriver
Corded
Electric Screwdriver
Screwdriver
Controller
Updates Backflow Engineering
Data Artifact
Engineering
Data Exchange Data Delivery

4. Challenges - FMEA and PPR Knowledge Co-evolution
4
Screw on
Dashboard
Robot
Screw Car Body
Dashboard
Car Body with
screwed on
Dashboard
Electric
Screwdriver
Bit
Product, Process, Resource
FMEA Cause & Effect
Failure Mode:
Screw breakaway
torque out of
tolerance
Cause:
Robot not correctly
calibrated
Insufficient
co-evolution
of FMEA and
PPR models
Insufficient
mapping of
system knowledge
C1
C2
M.Pos.Accuarcy
Q.Joining_Quality M.Torque
M.Bit_type
Breakaway_torque
Use Case: Re-Validation after Changes to Engineering Artifacts
• Showcase a screwing process with the change of a torque value
• Focus on the engineering views (quality, mechanics, automation)
and on the effect for the FMEA re-validation
Failure Mode and Effect Analysis (FMEA) Product-Process-Resource (PPR)

5. Research Methodology
We use Design Science, extending our previous work by
(I) conducting a domain analysis to identify requirements
(II) designing a multi-view meta-model to represent relationships between FMEA elements
and PPR assets
(III) providing a method to trace change states and dependencies in the design and
validation lifecycle
Evaluation of the multi-view meta-model and method in a feasibility study on the quality of a
joining process
5
RQ. How and under what conditions do changes to properties of
engineering artifacts necessitate a re-validation of FMEA elements?
S. Biffl, A. Lüder, K. Meixner, F. Rinker, M. Eckhart, and D. Winkler, “Multi-view-Model Risk Assessment in Cyber-Physical Production Systems
Engineering,” in MODELSWARD. SCITEPRESS, 2021
F. Rinker, S. Kropatschek, T. Steuer, E. Kiesling, K. Meixner, P. Sommer, A. Lüder, D. Winkler, and S. Biffl, “Efficient FMEA Re-Validation:
Multi-view Model Integration in Agile Production Systems Engineering,” CDL-SQI, Inst. for Information Systems Eng., TU Wien, Technical
Report CDL-SQI 2021-13, Nov. 2021

6. Background - Stakeholder Views & Artifacts
6
Effect
Cause
Detail Planner
Autom. (AE)
Robot Program
Software Config.
FMEA
model
Stakeholders
Views
Products &
Processes
Mechanical
Resources
Engineering
Artifacts
Process
P
Quality Engineer
(QE)
Abstract
Resrc.
FMEA model
Requirements
Process
Parallel
Engineering
Team Work
space
With
Backflows Team Workspace
Y
a b
Y
a b
Y
a b
Y
a b
Y
a b
XXX
Detail Planner
Mech. (ME)
Y
a b
M-CAD
Bill of Materials
Automation
Resources
ARes1
ARes2
Effect
Cause
Process
P
P
Basic Planner
(BP)
Product Design
Y
a b
Process Design
Abstract
Resrc.
Y
a b
Resource Design
Process
Process
P
P
Abstract
Resrc.
P
P
MRes1
MRes2
ARes1
ARes2
MRes1
MRes2
MRes1
MRes2
F. Rinker, S. Kropatschek, T. Steuer, E. Kiesling, K. Meixner, P. Sommer, A. Lüder, D. Winkler, and S. Biffl,
“Efficient FMEA Re-Validation: Multi-view Model Integration in Agile Production Systems Engineering,”
CDL-SQI, Inst. for Information Systems Eng., TU Wien, Technical Report CDL-SQI 2021-13, Nov. 2021

7. Change Management Process Analysis - UC Laser Welding
7
Identified Requirements for an efficient Multi-view FMEA + PPR (MvFMEA+PPR)
co-evolution and re-validation approach:
R1. FMEA concept representation.
• e.g. failure modes, causes, their relationships and characteristics
R2. PPR concept representation.
• e.g. products, production processes, production resources, relationships and properties
R3. FMEA-to-PPR dependency representation.
• represent links between FMEA concepts and PPR concepts, that are semantically similar
to concepts used in the FMEA
R4. FMEA/PPR change coordination representation.
• represent design and validation states for change coordination, e.g. elements that changed
or have to be re-validated after changes
R5. Efficient FMEA re-validation after PPR changes.
• e.g. efficient identification of FMEA model elements that require re- validation

8. Solution Approach - Multi-view FMEA+PPR meta-model
based on the FMEA Ontology, and the VDI 3682 Ontology-Design-Pattern
8
FMEA
Process FailureMode
causes
isCausedBy
MitigationAction
ControlMethod
examines
has
FailurreMode
isExaminedBy
RPN
hasRpn
hasNewRpn
hasMitigationAction
ProcessOperator
Resource
State
Product
consistsOf
isComposedOf
subClassOf
hasOutput
hasInput
Characteristic
0.. n
Marker
Type
hasMarker
hasType
State
hasState
ValueAttribute
hasValueAttribute
dependency
hasControlMethod
isAssignedTo
View
Stakeholder
hasView
hasView
0.. n
Link
BasicObject
hasType
hasSubProcess
0..n
Z. Rehman and C. Kifor, “An Ontology to Support Semantic Management of FMEA Knowledge,
” International Journal of Computers Communications & Control, vol. 11, no. 4, pp. 507–521, 2016.
C. Hildebrandt, A. Köcher, C. Küstner, C.-M. Lopez-Enriquez, A. W. Müller, B. Caesar, C. S. Gundlach, and A. Fay,
“Ontology Building for Cyber–Physical Systems: Application in the Manufacturing Domain,” IEEE Transactions on Automa

9. Multi-view FMEA+PPR re-validation Method
9
F. Rinker, S. Kropatschek, T. Steuer, E. Kiesling, K. Meixner, P. Sommer, A. Lüder, D. Winkler, and S. Biffl,
“Efficient FMEA Re-Validation: Multi-view Model Integration in Agile Production Systems Engineering,”
CDL-SQI, Inst. for Information Systems Eng., TU Wien, Technical Report CDL-SQI 2021-13, Nov. 2021

10. Feasibility Study: MvFMEA+PPR - Coordination Links and Marker
10
Cause-Effect
M.Torque
Screw on
Dashboard
Robot
Screw Car Body
Dashboard
Electric
Screwdriver
Bit
Resources
FMEA Cause & Effect
Failure Mode:
Screw breakaway
torque out of
tolerance
Cause:
Robot not correctly
calibrated
M.Pos.Accuarcy
M.Bit_type
Breakaway_torque
Drive
Robot
Controller
Screwdriver
Controller
M.Torque A.Screw.Curve
Products & Process
A.Motion.Accel
M.Pos.Accuarcy
Q.Joining_Quality
Cause Fault
Car Body with
screwed on
Dashboard
FMEA to PPR Dependency PPR to PPR Dependency
Process Resource
Product
Element changed Element to validate
Process-Resource
Product-Process
Characteristic
FMEA re-validation states
Multi-view FMEA+PPR meta-model

11. Feasibility Study - Neo4J Graph & Cypher Query
11
https://github.com/tuw-qse/fmea-revalidation-resources/blob/main/neo4j/multi_view-
fmea-ppr.neo4j

12. Feasibility Study - Evaluation of Re-validation Capabilities
12
5- point Likert scale (++, +, o, -, --), where ++/-- indicate very high/low capabilities,
to evaluate the fulfillment of the requirements in comparison with alternative approaches
Traditional approaches
a. FMEA+EA: FMEA re-validation based on Engineering Artifacts (EAs)
• Requires manual mapping and co-evolution of FMEA models and PSE artifacts
b. FMEA+TS: FMEA re-validation in Tool Suites (TSs)
• Manages engineering objects in a data base as a basis for co-evolution with FMEA m

13. Limitations
▪ Feasibility study focused on a single use case derived from projects at large PSE
companies in the automotive industry.
▪ This may introduce bias due to the specific selection of FMEA re-validation
challenges and approaches.
—> conduct case studies in wider variety of application contexts
▪ The expressiveness of re-validation concepts and dependencies used in the
evaluation can be considered a limitation
▪ Evaluation environment involved a limited number of stakeholders
▪ Ability to manage FMEA with many asset types and links to large PPR models
remains an open issue
—> investigate the effectiveness of the approach in larger settings
13

14. Conclusion and Future Work
▪ Advanced engineering use cases require multi-view risk and engineering knowledge
integration capabilities to handle the co-evolution of discipline-specific models and
knowledge
▪ Our work enables agile re-validation of artifact changes in a Multi-view FMEA and PPR
Model using methods and technologies such as Coordination Markers and Neo4J
▪ Future Work
▪ Provide a graphical model change reviewing interface for domain experts to
check the completeness and correctness of model changes (Rinker et al., 2020)
▪ Investigate approaches that are semantically more expressive than Neo4J, such
as Semantic Web technologies
▪ Investigate the usability and usefulness of making implicit domain expert
knowledge sufficiently explicit, to automate the multi-view change management
and analyses
14
Felix Rinker, Laura Waltersdorfer, Manuel Schüller, Stefan Biffl, Dietmar Winkler:
A Multi-Model Reviewing Approach for Production Systems Engineering Models.
MODELSWARD (Revised Selected Papers) 2020: 121-146

15. Contacts
Felix Rinker
felix.rinker@tuwien.ac.at
https://qse.ifs.tuwien.ac.at/frinker
15
Elmar Kiesling
elmar.kiesling@wu.ac.at
Stefan Biffl
stefan.biffl@tuwien.ac.at
Kristof Meixner
kristof.meixner@tuwien.ac.at
Sebastian Kropatschek
sebastian.kropatschek@acdp.at