Semantic Web for Advanced Engineering

Semantic Web for Advanced Engineering

Marta SabouVienna University of Technology,

Institute of Software Technology and Interactive Systems, Christian Doppler Laboratory for „Software Engineering Integration for Flexible Automation Systems“ (CDL-Flex)

2Dimensions not to scale. Adapted from Stefan Biffl.

Semantic Web

Software Engineering

AutomationEngineering

MechanicalEngineering

ElectricalEngineering

MechatronicEngineering

Business Informatics

CDL-Flex

My Research Universe

Human Computation

Content

What is the Fourth Industrial Revolution?– What are scenarios where Semantic Web technologies could be used?

To what extent can Semantic Web (SW) technologies be used to support the scenario of multi-disciplinary engineering?

What are challenges of applying SW technologies?

The Fourth Industrial Revolution

Source: Forschungsunion Wirtschaft und Wissenschaft, Acatech,”Securing the future of German manufacturing industry. Recommendations for implementing the strategic initiative INDUSTRIE 4.0 .Final report of the Industrie 4.0. Working Group.”, 2013

Cyber-Physical Systems (CPS)

Flexible, adaptive manufacturing (CPPS)Smart, distributed transportation systems

Estimated Economic Impact

Potential boost to the European Union’s gross domestic product by €110 billion annually over the next five years.

Manufacturing

70% of global trade In EU:

– 2 million businesses– 34 million jobs– 60% of EU economic growth

(Some) Challenges:– Shorter time to market– Increased product diversification and customization – Highly flexibilized (mass-) production– Higher product quality– Improved efficiency

Towards Flexible Production Systems and Processes

Source: Forschungsunion Wirtschaft und Wissenschaft, Acatech,”Securing the future of German manufacturing industry. Recommendations for implementing the strategic initiative INDUSTRIE 4.0. Final report of the Industrie 4.0. Working Group.”, 2013

Production System: Product-Process-Resource

Production step: Slicing

Material: Body with slices

Production resource:Slicing robot

Material: Breadbody

Production step: Baking

Product: Bread

Production resource: Oven

Characteristics of modern, flexible production systems

plug-and-participate capabilities of production resources – the integration and use of new or changed production resources

during production system use without any changes within the rest of the production system

self-* capabilities of production resources – self-programming of production process control, self-maintenance in

case of technical failures, or self-monitoring for quality late freeze of product-related production system behaviour

– fixing the characteristics of an ordered product at the latest possible point before production step execution, e.g., enabling to change the ordered colour of a car until the start of painting

Engineering Phase

Business Requirements

Cyber Physical Production System (CPPS)

Integrate Business Requirements in Engineering

Deploy created artifacts

Production TransportSales

Process Eng.

Electrical Eng.

CAD, Pipe & Instrumentation

Electrical Plan

Tool Data

Tool Data

Customer Representative

Software Eng.

Customer Reqs. & Review

Tool Data

Software Dev. EnvironmentTool Data

Control Eng.

PLC programTool Data

Project Manager

Engineering Cockpit

PLC

Test/Operation Phase

Operator

SCADATool Data

Multi-Model DashboardTool Data

Diagnosis Analysis

Tool Data

OPC UA ServerConfig

ERP SystemTool Data

Production Planning

Tool Data

Business Manager

Production Manager

Control Eng.



Access runtime information

Access engineering information


Engineering Cockpit

OPC UA Server (augmented)

Business Manager

Enrich runtime information

1

2

3

4

Scenario 1:Engineering Tool Network

Scenario 2:Multi-disciplinary

Reuse

Scenario 3:Flexible

Production

Scenario 4:Maintenance

Support

Production System Life-Cycle

Sc1: Discipline-crossing Engineering Tool Networks - fault free information propagation and reuse in engineering networks covering different engineering disciplines, engineers, and engineering tools during the creation of a production system.

Sc2: Use of existing Artifacts for Plant Engineering - identification and selection of reusable production system components.

Engineering Phase






Process Eng.

Electrical Eng.


Electrical Plan

Tool Data

Tool Data


Software Eng.


Tool Data


Control Eng.


Project Manager

Engineering Cockpit

PLC


Operator

SCADATool Data


Diagnosis Analysis

Tool Data

OPC UA ServerConfig

ERP SystemTool Data

Production Planning

Tool Data

Business Manager

Production Manager

Control Eng.






Engineering Cockpit


Business Manager


1

2

3

4

Scenario 1:Engineering Tool Network


Reuse

Scenario 3:Flexible

Production


Support

Production System Life-Cycle

Sc3: Flexible Production System Organization – aims at run-time flexibility of production systems. Enables the integration of advanced knowledge about the production system and the product within the production system control at production system run-time.

Sc4: Maintenance and Replacement Engineering – combines engineering and run-time information of a production system towards improved maintenance capabilities of production system components.

Semantic Needs for Industrie4.0 Scenarios

Production System Engineering Needs & Scenarios SC1 SC2 SC3 SC4

N1 Explicit engineering knowledge representation ✔ ✔ ✔ ✔

N2 Engineering data integration ✔ ✔ ✔ ✔

N3 Engineering knowledge access and analytics ✔ ✔ ✔ ✔

N4 Efficient access to semi-structured data in the organization and on the Web

✔ ✔

✔

N5 Support for multi-disciplinary engineering process knowledge

✔ ✔ ✔ ✔

N6 Provisioning of integrated engineering knowledge at production-system run-time

✔ ✔

Content

What is the Fourth Industrial Revolution?– Industry 4.0 = flexibility through cyber-physical systems– Manufacturing an important area– Need for flexible production systems and processes (CPPS)– Several scenarios and needs for semantic technologies in the

complex life-cycle of production systems

To what extent can Semantic Web technologies be used to support the scenario of multi-disciplinary engineering?

16

CPPS Engineering - Complex scenario

Data complexity Engineering data From different disciplines Complex dependencies Changing Large (40+K signals)

Distributed and concurrent engineering

Different disciplines Different terminology Different tools Heterogeneous data

models and formats

Software Eng.Mechanical Eng. Electrical Eng.

17

CPPS Engineering – Example Tasks & QueriesData analysis across data from engineering disciplines

– Comprehensive statistics (across disciplines)– Constraint Checking – Defect detection

Change propagation and notification across disciplines

Software Eng.Mechanical Eng. Electrical Eng.

Which machine functions are needed to produce Product X with Production Process Y?

Which sensors are not linked to a software variable?

Which component contains more than one signal?

18 Software Eng.Mechanical Eng. Electrical Eng.

Solution Idea: Common Concepts

19

Common Concepts provide a common vocabulary to speak about the data in common They link distributed and heterogeneous (local) data models.

Common Concepts

Data Integration Solution

Ontology Based Information Integration (OBII) Approach Integration of data from

heterogeneous sources using [Cal01, Wac01].

Three components of the OBII approach: – (1) Local ontologies - to

represent data specific to a data source.

– (2) A common ontology - to represent the aggregation of relevant concepts

– (3) The mapping between local ontologies and the common ontology.

Technology Stack

Common Concepts in Production Systems Engineering

Ontology Classification Schemes

Product Production Process

Production Resource

Physical Objects

Ont. of product types, OntoCAPE, eClassOWL

OntoCAPE, ISO 15926

Ont. of resource types, OntoCAPE, CCO, AMLO, ManufOnto, eClassOWL, ISO 15926, AutomOnto

Structure Ont. of product structure, OntoCAPE NF

Ont. of resource structures, OntoCAPE, ManufOnto, EquipOnt, CCO, AMLO,

ISO 15926, AutomOnto

Functionality NFOnt. of production

process types, OntoCAPE, ISO 15926, ManufOnto

Ont. of production resource capabilities (skills), AMLO, ManufOnto, EquipOnt,

ISO 15926

Process NFOnt. of production process structures,

OntoCAPE, ISO 15926, ManufOntoManufOnto

Materials Ont. of bills of materials, eClassOWL NF NF

Observations, Measurements NF

Ont. of process states and its observability, SSN,

OntoCAPE, ISO 15926Ont. of resource states, SSN, AutomOnto

Quantities, dimensions, units

Ont. of product characteristics,

eClassOWL

Ont. of production processes characteristics,

OntoCAPE, ISO 15926

Ont. of production resource characteristics, ManufOnto, CCO, SSN,

AutomOnto

Typical Ontology Modelling Needs

Modelling Part-Whole relations – containment hierarchies are a well-accepted and frequently

occurring organizational paradigm from modelling part-whole relations in mechatronic engineering settings

– No built-in support in OWL but several ODPs Modelling connections between components

– interface-based composition describes the capabilities expected from an interface and can enable reasoning tasks about the correctness of a system’s structure.

Modelling component roles – component roles refer to their functions and behaviour that they play

in the system

Technology Stack

26

Semantic Mappings between Engineering Concepts

Evaluation of Mapping Alternatives

Source: Kovalenko O, Euzenat J (2016) Semantic Matching of Engineering Data Structures. In Biffl S, Sabou M (Eds.) Semantic Web for Intelligent Engineering Applications. Springer

Technology Stack

Automating Cross-Disciplinary Defect Detection and Data Analysis

29

30

Constraint Checking across Engineering Disciplines

“All safe software variables should be linked to exactly two sensors” “Check that all sensors have PLC variables defined”

SELECT ?sensor ?sensor_id WHERE {

?sensor a hw:Sensor . ?sensor hw:hasKeyValue ?sensor_id .

?hw_var a hw:Variable . ?hw_var hw:isDefinedOnDevice ?sensor . ?hw_var hw:hasItemName ?hw_var_name .

OPTIONAL { ?var a cs:GlobalVariable . ?var cs:hasName ?cs_var_name .

FILTER (?hc_var_name = ?cs_var_name) . }

FILTER (!bound(?cs_var)) }SELECT ?kks ?signal WHERE {

{SELECT ?kks WHERE { ?kks :hasSignal ?signal }

GROUP BY ?kks HAVING (COUNT (?signal) >= 2)} ?kks :hasSignal ?signal}}

31

Signal List -Version 1 Version 2

Knowledge Change Management – “Change Compression”

25 deletes, 30 updates, 15 insertions.From syntactic level

Pump XA_20 was moved to sector AH1 To semantic level

In real-life scenarios scale is a major issue, e.g:• 40,000 signals• 2,500 deletes, 3,000 updates, 1,500 insertions.

Knowledge Change Management : Change Propagation

Browsing and Querying of Cross-disciplinary Engineering Data

33

AutomationML Analyzer: relies on Linked Data technologies to enable efficient integration, browsing, querying, and analysis of diverse engineering models represented in AutomationML.

Browser based Visualisation

34

Browsable internal links

Different Views on Data

Querying Integrated AutomationML Data

35

Predefined SPARQL queries enable monitoring, analysis, validation and defect detection tasks

Semantic Web Capabilities for Engineering Settings

Semantic Web Capabilities & Needs N1 N2 N3 N4 N5 N6

C1 Formal semantic modeling ++ + ++ + + +

C2 Intelligent, web-scale knowledge integration

+ ++ ++ ++ ++

C3 Browsing and exploration of distributed data set

+ ++ + +

C4 Quality assurance of knowledge with reasoning

++ ++

C5 Knowledge reuse + + ++

++ +

Summary

What is the Fourth Industrial Revolution?– Industry 4.0 = flexibility through cyber-physical systems– Manufacturing an important area– Need for flexible production systems and processes (CPPS)– Several scenarios and needs for semantic technologies in the complex life-

cycle of production systems To what extent can Semantic Web technologies be used to support the

scenario of multi-disciplinary engineering?– CPPS engineering is a complex scenario– Data integration is crucial– There is a good match between Semantic Web technology capabilities and

the needs of Industrie 4.0 scenarios What are challenges of applying SW technologies?

Challenges

Lack of knowledge acquisition interfaces that are easy to use by engineers– Trend: use SysML and SysML4Mechatronics as front-end for

acquiring ontologies of engineering models Lack of support for mathematical calculations:

– Trend: Hybrid solutions integrating data mining, statistical analysis and relational constraint solvers

OWA not a natural fit for engineering– Must adopt a CWA style presentation of results at the interface level

Dealing with dynamic engineering data– Trend: applying ongoing research in semantic stream reasoning

Outlook

Ample opportunities for using SW in Industrie 4.0 settings– Only one of the four identified scenarios has been well explored– How about other scenarios?

Engineering Phase






Process Eng.

Electrical Eng.


Electrical Plan

Tool Data

Tool Data


Software Eng.


Tool Data


Control Eng.


Project Manager

Engineering Cockpit

PLC


Operator

SCADATool Data


Diagnosis Analysis

Tool Data

OPC UA ServerConfig

ERP SystemTool Data

Production Planning

Tool Data

Business Manager

Production Manager

Control Eng.






Engineering Cockpit


Business Manager


1

2

3

4

Scenario 1:Engineering

Tool Network


Reuse

Scenario 3:Flexible

Production


Support

Outlook

Ensuring successful SWT uptake by practitioners– Use engineering specific languages as front-ends for the creation of

engineering ontologies (e.g., UML, SysML)– New ontology classification schemes that bridge the needs of

practitioners and SW experts– Better understanding of typical modeling needs and providing

guidelines for solving those, e.g. through Ontology Design Patterns– SW tool evaluation and selection frameworks (e.g., XSL2RDF tools)

Outlook

Extensions to current SW technologies:– High-performance tools that can deal with large, diverse and rapidly

changing datasets Knowledge change management on integrated data sources Managing dynamic engineering data (e.g., stream reasoning)

– Data integration Automatic identification of semantic overlaps between

engineering models More expressive languages to declare mappings between

engineering models– Evaluation of software architectures taking into account the needs of

Industrie4.0 specific applications Further investigating the use of Linked Data technologies in

engineering scenarios

42

Source: By Leonardo da Vinci - Bortolon, The Life and Times of Leonardo, Paul Hamlyn, Public Domain, https://commons.wikimedia.org/w/index.php?curid=1647253

Be part of the current revolution in engineering!

Use Semantic Web technologies to improve engineering!

A Final Word