This document proposes a model-based approach for risk assessment in multi-disciplinary engineering projects. It involves defining metamodels for production system models, link models between artifacts, and metrics. Metrics are defined using the Structured Metrics Metamodel and calculated by executing queries on the system models. Measurement results are stored in the metrics model. The approach aims to support risk assessment across distributed, versioned engineering artifacts represented in AutomationML. Future work includes expanding the metrics, integrating dynamic aspects, and visualizing results.
Model-Based Risk Assessment for Multi-Disciplinary Systems
1. Model-Based Risk Assessment in
Multi-Disciplinary Systems Engineering
Euromicro Conference series on Software Engineering and
Advanced Applications (SEAA) 2015
Arndt Lueder, Nicole SchmidtStefan Biffl, Luca Berardinelli,
Emanuel Maetzler, Manuel Wimmer
2. Introduction
2
Multi-Disciplinary Engineering
Multidisciplinary Domain
Mechanical Engineering
e.g., design, production, and operation of
machinery (powered tools)
Electrical Engineering
e.g., design of complex power system and
electronic circuits
Software Engineering
e.g., design, implementation, testing,
validation of software for machinery
• Heterogeneous document/tool landscape
Mechanical Engineering
Matlab, CAD tools…
Electrical Engineering
EPLAN
Software Engineering
Programming IDEs, Modeling Tools…
= domain = tool = doc
overall system design
mechanical
engineering
electrical
engineering
software
engineering
3. Introduction: Industry 4.0 and its principles
3
Industry 4.0: computerization of
manufacturing. Driving principles
1. Interoperability among mechatronic
systems (a.k.a. cyber physical systems
CPS), humans and factories
2. Virtualization: a virtual copy of the
factory with sensed data
3. Decentralization: the ability of CPSs to
make decisions on their own
4. Real-Time Capability: monitoring,
analysis, planning, execution
5. Service Orientation: OPC Unified
Architecture (SOA)
6. Modularity: flexible adaptation to
changing requirements
= domain = tool = doc
overall system design
mechanical
engineering
electrical
engineering
software
engineering
Industry 4.0
4. Introduction: Engineering of Industrial Production Systems
4
AutomationML (AML) standard for tool
data exchange
AML docs are XML-based artifacts
AML as pivotal language: Tool-
specific docs can be transformed in
AML docs
overall system design
mechanical
engineering
electrical
engineering
software
engineering
Industry 4.0
= domain = tool = doc
XML-based
artifacts
CAEX.xsd
5. Introduction: Engineering of Industrial Production Systems
5
Mechanical/Eletrical/Software
Components Library
«cloned»
Production System Model
«represents»
Lab-sized Production System
“Equipment Center for Distributed Systems,”
http://www.iafbg.ovgu.de/en/technische ausstattung cvs.html,
Institute of Ergonomics, Manufacturing Systems and
Automation at Otto-v.-Guericke University Magdeburg.
6. Introduction: Engineering of Industrial Production Systems
6
Risk: the probability of an
occurring event which can have a
negative impact on system overall
quality
Risk Assessment: collection of
adequate metrics to feed analysis
processes including the
identification of countermeasures
throughout the system engineering
process
Model-Based Risk Assessment:
collection of metrics on (machine
readable) models (e.g., AML ones)
7. Introduction: Engineering of Industrial Production Systems
7
Risk: the probability of an
occurring event which can have a
negative impact on system overall
quality
Risk Assessment: collection of
adequate metrics to feed analysis
processes including the
identification of countermeasures
throughout the system engineering
process
Model-Based Risk Assessment:
collection of metrics on (machine
readable) models (e.g., AML ones)
8. Problem Description
8
Risk management is an error prone
and cumbersome task for industrial
production systems especially when
having distributed models in different
variants and versions
Needs for metrics for AML artifacts
Needs for linking and versioning
support for AML artifacts
Lack of Model-Based foundation for
risk assessment based on AML
= domain = tool
overall system design
mechanical
engineering
electrical
engineering
software
engineering
= doc
9. Contribution: Model-Based Measurement Process for AML
9
Jacquet et al. proposed a generic measurement process model to calculate metrics for
software engineering projects.
10. Contribution: Model-Based Suite for AutomationML
10
Jacquet et al. proposed a generic measurement process model to calculate metrics for
software engineering projects. We contextualized inputs/outputs for each step.
11. Contribution: Model-Based Measurement Process for AML
11
• Our objective
Risk assessment
in multidisciplinary systems engineering projects
based on AML and linked AML artefacts
to reason on a set of distributed engineering artifacts
and their relationships
<<metamodel>>
AttributedGraph
<<metamodel>>
AMLMetamodel
<<model>>
AMLModel
<<model>>
AMLLibrary
conformsTo
Relationship
Legend
<<metamodel>>
LinkMetamodel
<<model>>
LinkModel
connects
Relationship
12. Contribution: Model-Based Measurement Process for AML
12
We need models for...
1. the production system,
2. links among heterogenous, versioned artifacts
3. the metrics definitions
4. metric results
from
Model-Based Co-Evolution of Production Systems
and their Libraries with AutomationML
Berardinelli, Biffl, Maetzler, Mayerhofer, Wimmer
IEEE International Conference on Emerging Technologies
and Factory Automation (ETFA 2015)
AML Metamodel
defines the concepts
and their relationships
for production system
modeling.
Implemented using
Eclipse Ecore
previous work
13. Contribution: Model-Based Measurement Process for AML
13
We need models for...
1. the production system,
2. links among heterogenous, versioned artifacts
3. the metrics definitions
4. metric results
from
Linking and Versioning Support for AutomationML:
A Model-Driven Engineering Perspective
Biffl, Maetzler, Wimmer, Lueder, Schmidt
IEEE International Conference on Industrial Informatics
(INDIN 2015)
Linking Metamodel
defines the concepts and
their relationships for
representing links among
artifacts conforming to
different metamodels.
Implemented using
Eclipse Ecore
previous work
14. Contribution: Model-Based Measurement Process for AML
14
We need models for...
1. the production system,
2. links among heterogenous, versioned artifacts
3. the metrics definitions
4. metric results
from
Structured Metrics Metamodel (SMM) by OMG.
http://www.omg.org/spec/SMM/
“The SMM is a specification for
the definition of measures
and the representation of
their measurement results.
The measure definitions make
up the library of measures and
that serves to establish the
specification upon which all of
the measurements will be
based.” from
http://www.omg.org/spec/SMM/
1.0/
Implemented using Eclipse Ecore
21. Contribution: Model-Based Measurement Process for AML
Future work: exploitation of result, e.g., for impact analysis
May be done with reporting tools such as Eclipse Birt
Connector to EMF models required
http://www.eclipse.org/birt/
22. Conclusion
We discussed foundations for risk assessment in multi-disciplinary
software and systems engineering projects and proposed a model-
based metrics suite for AML models and their inter-model links.
We plan to extend the AutomationML metrics suite for several aspects
Further metrics and queries for AML and link model artifacts
Integrating dynamic aspects of AML model elements specifications through PLCopen
XML and state-machine like notations
Visualization of results using Birt or graphical and textual modeling editors
22