1. Towards a Reference Ontology for Manufacturing
Professor Bob Young
Contact:

2. Points to cover
A view of manufacturing ecosystems and the need for a Manufacturing Reference Ontology
Our recent work:
Interoperable Manufacturing Knowledge Systems (IMKS) –( )
Related PhD projects:
Chungoora N, 2010 “A Framework to Support Semantic Interoperability in Product Design and Manufacture”
Usman Z 2012 “A Manufacturing Core Concepts Ontology to Support Knowledge Sharing”
Imran M, 2013 “ Towards an Assembly Reference Ontology for Assembly Knowledge Sharing”
Hastilow N, 2013 “Manufacturing Systems Interoperability in Dynamic Change Environments”
Intelligent Systems Configuration Services for Flexible Dynamic Global Production Networks (FLEXINET) - (2013 – 2016)
Summary of progress and issues

3. Semantic Interoperability supports information understanding
Drive competitiveness through high quality decision making
Decisions
Decisions
Decisions
Issues:
Clarify understanding across decision domains
Ability to flexibly share information between systems
Ability to readily configure “common” Knowledge Bases
Decisions
Information
Decision makers need access to high quality information to make high quality decisions
Information isn’t all in one place. Not necessarily easily accessible. A lot of it probably in peoples heads. It keeps changing
Different internal groups
Different customers
Different factories
Different suppliers
Different systems
Different Global Constraints

4. All Manufacturing Industry sectors have the same sorts of problems with information

5. Elements of a Manufacturing Ecosystem
Business
Elements of a Manufacturing Ecosystem
Manufacturing
Business
Manufacturing
Business
Factory
Factory
Timescales
Systems
Factory
Strategic Management Activities
Years
Project
Planning
Months
PLM
ERP
SCM
Product Design/Tactical Planning
Activities
Operations Management Activities
MES
Hours - Days
Shop Floor Activities
Control Systems
Minutes -Seconds

6. A simpler view of manufacturing activities
Develop new products
Produce products
Service products
Dispose of Products
Schedule production
Make and assemble parts
Manage supply networks
Satisfy Customers
Supply Products
Make Products
Use resources

7. The Loughborough Viewpoint
Manufacturing is complex and understanding relationships is critical
Domain ontologies are too specific and foundation ontologies are too generic – manufacturing reference ontologies sit between the two
Build our understanding through a range of specific research investigations to support manufacturing industry
Use the most expressive ontology development environment that we can access, to most effectively model the complexity of manufacturing interactions.
Use a commercial environment as this gives us the only realistic route to manufacturing business uptake.
Our work is based on the use of HIGHFLEET Inc.'s Common Logic based IODE and XKS environments

8. Specific IMKS investigations
Product design / manufacture (machining)
Design for assembly/ assembly planning
Manufacturing intelligence systems interoperability
This approach does not ensure interoperability
it provides understanding of interoperability potential
it allows informed choices to be made

9. An example from aerospace manufacture
Stage 4, High Pressure Compressor Disc

10. There are many categories of manufacturing concepts

11. a Manufacturing Core Concepts Ontology illustration – (Usman 2012)

12. IODE - Manufacturing Core Ontology - Classes

13. IODE - Manufacturing Core Ontology – Example Relations
executedIn
hasCapabilityFor
holds
produces

14. IODE - Manufacturing Core Ontology - Integrity Constraints
hasCapabilityFor

15. IODE - Design and Manufacturing Feature Specialisations

16. Linking production knowledge back to product design
Design View
Production View
Production Knowledge
OP Number
Description
Category 10
Outer Dia Turning F 20
Web Profile Machining R 30
Turn Circumferential Groove
200
Perform Final CMM Inspection I
210
Final Clearance Inspection
Rule 1.1:
The groove width should be greater than
tool size + clearance
( => (and (CuttingTool ?ct)
(Parameter ?toolclearance)
(Parameter ?toolwidth)
(hasDimension ?ct ?toolclearance (mm ?real1))
(hasDimension ?ct ?toolwidth (mm ?real2))
(NeckWidth ?nw)
(hasValue ?nw (mm ?real3))
(gteNum (numPlus ?real1 ?real2) ?real3)
( CannotbeUsed ?ct)
: IC Soft "The groove width is less than the tool width + clearance”
Production Knowledge

17. Now - an assembly perspective – key reference concepts in design for assembly and assembly planning (Imran 2013)

18. Imran (2013) reference ontology extensions to suit assembly
Equivalent extensions also needed for other manufacturing processes e.g. forging, casting, injection moulding, etc., etc.

19. A manufacturing intelligence systems interoperability perspective (Hastilow 2013)
Manufacturing Intelligence (MI) aims to leverage ICT to collect data from and monitor production systems in real time. Systems need to be upgraded or replaced over time and need to remain interoperable with existing systems

20. Hastilow’s (2013) work introduces a time dimension and a “system’s” perspective

21. System Core Ontology Concepts (Hastilow, 2013)

22. Manufacturing Intelligence Systems Core Concepts (Hastilow, 2013)

23. A key result of Hastilow’s research
The population based on the ontology flagged 83 errors in the system upgrade requirements.
The errors generated by the ontology relating to the system upgrades were consistent with the real world issues with the system.
if this approach had been used a number of issues could have been resolved at the system definition stage.

24. FLEXINET: an EU project involving 3 manufacturing & 6 software companies plus 4 academic institutions
Takes the ‘systems’ concepts forward into supply network systems
Targets strategic and tactical decisions on business and global supply network decisions
Reference ontology provides basis for software application interactions and KB development for manufacturing companies

25. Reference ontology supports KB construction and interoperation across software services

26. Ontology Layers
Level 0
CORE (ULO)
(Highfleet – based on Common Logic)
Specialisation from generic to progressively more specific
Level 1
PSL – ISO 18629
Systems (functions)
Level 2
Designed Systems
Natural Systems
Level 3
Manufacturing Business Systems
(context)
Healthcare Systems
Banking Systems
Level 4
Marketing
Product-Service Lifecycle Systems
Finance
Scope
Design
Produce
Operate
Dispose/
End of life
Level 5
Enterprise Specific
Everything from level 0-4 should be standard and be used to readily configure enterprise specific ontologies
The enterprise specific ontologies form the basis for the knowledge base that can support ‘what-if” queries on the potential global production system configurations.

27. The level 1 concepts and relationships

28. Main categories of concepts at levels 2 & 4 in the reference ontology
The ontology currently comprises some 600 concepts along with 55 constraints and 26 rules
The FLEXINET system is currently being evaluated by 3 manufacturing companies from the pumps, white goods and food & drink industries
The ontology will be freely available this summer

29. Some detail issues from my perspective
Why aren’t there more Common Logic based environments available? Most applications work seems to be done using OWL.
Closed world versus open world – manufacturing decisions are usually based on wanting to know what is possible, not what might be possible, so we use HIGHFLEET’s closed world assumption. 
To use or not to use Clabjects (objects that are also classes)? There are situations in manufacturing where they are very useful, but limit flexibility.   
There are times when open ended names need to be used rather than formally defined concepts. Offers flexibility in one application area but goes against the whole argument for reference ontologies.

30. A final comment using a rail systems analogy
There is a lot to do but it seems there is a lot of potential Ideally we need investment in standard manufacturing reference ontologies to support effective information sharing across manufacturing.
Standard rail networks enable trains to run “interoperably”