CMMS under the Magnifying Glass

1
CMMS under the Magnifying Glass

• - An Approach to Deal with Substantive
  Interoperability -

Mårten Lundgren M. Sc. Swedish Defence Research
Agency (FOI) Div. of Systems Technology, Dept.
of Systems Modelling 46 70 5561992 marten.lundgre
n_at_foi.se
2
Who we are

• Swedish Ministry of Defence
• Swedish Defence Research Agency (FOI)
• Systems Technology Division
• Department of Systems Modelling
• Methods and technologies for complex modeling and
  simulation.

3
Progress at FOI

• 2001/2002 Usefulness of concept
• 2003 Further investigation
• KA / KE
• Tool support
• Syntax and semantics
• Limitations in CMMS KA / KE
• Ideas on how to solve them

4
CMMS at a glance
Real world
Develop Focused Context
Knowledge Engineering
Gather Information
Formalize Data
Construct CMMS
Knowledge Acquisition
5
Perceived limitations of CMMS

• Process dependent
• Poor methodology support
• Poor modeling abilities

6
Limitation 1 Process dependency

• Too specific ? not reusable.
• Too complex ? not understandable

1 aim
3 fire
2 breathe
Activity
1 aim
Choice
3 fire
2 breathe
7
Solution 1 Activity centric models
process chain of events

• Increased reusability
• Decreased complexity

8
Limitation 2 Poor methodology

• No basic approach to knowledge
• Too general
• Only KE focused.
• Misses many critical questions
• How to delimit scope?
• How to identify SMEs?
• How to automate knowledge refinement?
• How to solve syntax and semantics issues?
• Etc....

9
Solution 2 Activity based KA and KE

• Develop KA to support KE
• Activity centric modeling proved helpful
• Conform to common standards
• Interoperability benefits

10
Limitation 3 Poor modeling abilities
specific
generic
complex
11
Solution 3 Meta-layers

• Meta-layers classify components and add semantics

• Meta-meta layers enable even more semantics

• Further mapping to MOF enables automation of
  mapping
• between notations or even code generation.

3. Reader makes assumptions
2. Knowledge engineer formalizes information
1. Information recieved from SMEs view
12
Summary Limitations

• Process dependent models are poor analogies of
  reality which causes problems in complexity and
  modularity.
• Coarse and incomprehensive CMMS methodology gives
  little or no support to important issues.
• Flat models cannot represent knowledge
  efficiently, or automate mapping of views.
• ? complex and difficult to reutilize knowledge
  models that limit substantive interoperability.

13
Summary Solutions

• Use a basic approach like activity centric
  concepts.
• Develop a detailed methodology based on this
  approach.
• Use a standardized meta-modeling framework that
  enables abstraction extension
• Modularity
• Automation
• Interoperability