Improving Aerospace System Engineering and Manufacturing Activities through Semantic Integration of

1
Improving Aerospace System Engineering and
Manufacturing Activities through Semantic
Integration of Activities and Standards
Concepts toward a proposal to the NSF Information
Technology Research Program

• J. Scott Hawker
• Assistant Professor of Computer Science
• Gary P. Moynihan
• Professor of Industrial Engineering
• University of Alabama

2
Summary

• Standards provide proven knowledge on engineering
  and manufacturing products and processes
• It is difficult to find and use relevant
  standards
• Standards knowledge is locked in documents meant
  for manual processing by humans
• There are tens of thousands of standards from
  hundreds of standards development organizations
• For a given system design, there are tens to
  hundreds of cross-referenced, tailored, revised,
  and often conflicting standards that may or may
  not be relevant to multiple engineering and
  manufacturing disciplines
• The proposed effort will provide
• Formal, software-based representation of
  engineering and manufacturing activities and
  standards
• Techniques for automated discovery and use of
  standards
• Significant base toward theory of engineering
  knowledge
• Approach
• Semantic web-based representation of activities
  and standards
• Knowledge-based tools for semi-automatic
  extraction of semantic content
• Agents for search, discovery, and delivery of
  standards to activities
• Smart standards that deliver their content as
  task-specific tools, aides, experts, etc., for
  integration into CAE/CAD/CAM tools

3
Product Engineering and Manufacturing Activities
Constraints Experience

• Concurrent, iterative activities in integrated
  product and process development
• Multiple levels of detail
• Knowledge-driven

Manufacturing techniques
Engineer Manufacturing System
Process capabilities
Needs
Constraints Experience
Processes Tools
IDEFØ notation
Design techniques
Engineer Product
Product specifications
People Tools
Needs

• Inputs
• Product needs
• Engineering Knowledge
• Techniques
• Constraints
• Experience
• Materials

People Tools
Manufacture Product
Standards
Products
Materials
People
Processes Tools
4
Standards Improve Engineering and Manufacturing

• Standards provide engineering knowledge
• Requirements, constraints
• Engineering and manufacturing processes
• Guidance, techniques, best practices, lessons
  learned,
• Reusable assets templates, frameworks, patterns,
  reference architectures, tools, algorithms,
  qualified parts and flows, etc.
• Standards
• Improve product quality
• Reduce program cost, time, risk
• Improve engineering capability
• Help assure system interoperability
• Help assure contractors deliver correct
  sub-systems

5
It is Difficult to Identify and Use Relevant
Standards

• Tens of thousands of standards from hundreds of
  standards development organizations
• Which are relevant to a given engineering or
  manufacturing task?
• Most standards must be purchased before they can
  be read
• U.S. Federal agencies mandated to use industry
  standards
• Standards are highly cross-referenced, tailored,
  specialized, revised, etc., and often conflicting
• Standards are documents for human consumption
• Paper form (electronic PDF slowly coming)
• Manual processing to incorporate into product and
  process designs
• Complicated calculations, logic, technologies,
  processes, etc., are hard to capture and
  interpret in text

6
Aerospace Systems are Standards-Driven

• Improving product engineering and manufacturing
  in the aerospace industry will have broad impact
  in a huge, critical industry
• Standards form the single largest source of
  technical data used by Engineering and Operations
  (39 of Engineering Data, 38 of Operations
  Data) at The Boeing Company
• Linda Duschl, Manager of Engineering Standards

• Boeing
• 7,000 Military and Federal Specs Standards
• 400 ASTM material specs and test methods
• 1,000 AIA/NAS part standards
• 2,500 SAE standards

• NASA Technical Standards Program
• Agency-wide preferred technical standards
• 1,000 standards from 40 non-government
  standards development organizations
• 600 MIL standards and specifications
• 80 NASA-specific standards
• Provides access to 60,000 specifications and
  standards
• 35,000 non-government
• NASA personnel involved in over 145 national and
  international standards developing organizations,
  committees and working groups

7
Problem Summary Overwhelming Impact of Standards
Requirements
Standard Tailoring Industry?MIL?Agency?Center?Grou
p
Environmental Requirements (vehicle, carrier,
rack, facility, etc.)
Program/Project Requirements
Aerospace Systems Developer
Lessons Learned
Best Practices
Experience
8
Conceptual Approach
Space system designer, manufacturer, user
System requirements, components, materials,
processes, etc.
Relevant standards and experience

• Web-Based Organization and Search
• Activity-Centered Information Portals
• Semantic Web
• Knowledge-Based Systems

Standards Advisor

• Technical Specifications and Standards
• Experience Lessons learned, application notes,
  etc.

9
Approach Summary

• Use semantic web technologies for
  computer-aided standards discovery and use
• Develop domain models/ontologies to characterize
  engineering and manufacturing tasks and to
  facilitate the delivery of standards to those
  tasks
• Discipline and task-based ontology elements (key
  words and relationships, basic processes, named
  techniques, etc.)
• Basis for task-centered information discovery and
  delivery
• Develop and use an ontology to capture standards
  metadata and content
• General structure and use of standards
• Metadata and data describing intent, content, and
  relationships of specific standards
• Task ontologies provide core vocabulary
• Provide information search techniques to
  integrate tasks and standards
• Task models seek out related standards
• Standards content models seek out related
  standards and lessons learned
• Tools for metadata and content extraction and
  browsing
• Develop an architecture for smart, interactive
  standards
• Deliver standards content as task-specific tools,
  aides, experts, etc.

10
Use semantic web technology to bridge the gap
between engineering/manufacturing activities and
standards
Standards Lessons Learned
Wide gap Weak connections
CAE/CAD/CAM tools Product Data Management Process
Workflow Engines
Engineer Manufacturer
11
Standards Lessons Learned

• Document Metadata
• Title, standards organization,
• Purpose, scope, summary
• Keywords
• Index
• Table of contents
• List of figures, tables

• Document Relations
• References
• Revisions
• Tailorings
• Where used

• Document Content
• Text elements
• Figures, tables
• Internal references

• Smart Content
• Formal specifications
• Computations
• Conformance checks
• Aides, experts integrated with CAE/CAD/CAM tools

Data
Tools Semantic search and discovery
agents Ontology-aware browsing Knowledge-based
metadata creation Information delivery agents
Data

• Activities, sub-activities
• Descriptions
• Info, documents used
• Info, documents produced
• Terminology (glossary)

Conforming designs and processes

• Project organization
• Work breakdown
• Product architecture
• Organizations

CAE/CAD/CAM tools Product Data Management Process
Workflow Engines
Engineer, Manufacturer
12
End-User Tools (Browse, View, Reports)
Software Applications (Query, Use, and Generate
Content)
Decision Support and Automation
Advanced Web Engines
View, Browse and Search Text
View, Browse and Search Data and Structure
Understand Domain-Specific Meaning
Automated Reasoning
Processing
Data
Semantic Web
Logical Assertions
Vocabulary, Taxonomy, Ontology
XML Schema, RDF, RDF Schema
HTML, XML, MIME types (PDF, multimedia, etc.)
based on diagram from Tim Berners-Lee, W3C
13
Multiple Representations of Engineering Knowledge

• Knowledge in one representation can be mapped to
  another representation
• Discover common domain ontology by mapping to
  common core ontology

Computation (Knowledge Representation and
Reasoning)
Visualization and Editing
Interchange
Jambalaya
Inference engines
Agents
Domain Ontologies (PSL, Dublin, etc.)
Protégé
UML notation
Web browser style sheets (e.g. XSL)
RDF, RDF-Schema
IDEF
UML
SQL
Acrobat reader, MSWord, Flash, etc.
XML, XML-Schema
UML metamodel and meta-metamodel
xHTML
Web browser
MIME
KIF (Knowledge Interchange Format)
Text
PDF
Model theory (Set theory, logic, etc.)
14
Distributed Data and Metadata Architecture
Issues in scalability, interoperability, schema
evolution, long-term availability of standards,
information ownership (access, purchase), etc.
Search
XML RDF/DAMLOIL
Indices, Taxonomies
Crawl
Access
Firewall
Metadata Repository
Search
Indices
MatML, PSL, etc.
XML
XML, MIME, JPEG7, etc.
Crawl
Access
Full-text standards (3rd-party provider)
Lessons Learned Databases
Document Library
Special-purpose databases (materials, structures,
processes.)
15
Demonstration Domains

• Avionics design and manufacturing standards
• Software engineering process and product
  standards
• Note
• We are not discovering formal domain models and
  ontologies from scratch
• Rather, we are representing in a
  computer-processable form the knowledge already
  captured in existing domain standards
• Our contribution is in techniques for
  representing and exploiting the activity and
  standard knowledge

16
Results from Work with NASA Technical Standards
Program

• Standards advisor vision
• Generic engineering task model
• Activities, sub-activities, resources
  used/produced
• Software engineering task model
• Based on ISO, IEEE, OMG standards
• Task-specific web portal prototype
• Deliver, search for, and evaluate standards for
  flight software design processes and techniques
• Simple document ontology
• Based on Dublin Core and other library science
  models
• Document review and metadata extraction tools
• Software agents to discover potentially-related
  standards and lessons learned

17
Result

• Answers on
• How to represent engineering knowledge
• How to find engineering knowledge
• How to use engineering knowledge
• Significant content of engineering knowledge
• Make the information contained in engineering
  standards available as computer-based tools
• Proof-of-concept prototypes to improve NASA
  Technical Standards program and other developers
  and users of engineering standards

18
Projects (Brainstorm list)

• define the activity and standards ontologies
• pick a standard and capture its content
• interactive, browse-able standard
• ontology, metadata
• smart contents
• documents repository and metadata creation tool
• store, read, annotate, link, browse, query, etc.
  engineering standards, research papers, etc.
• metadata extraction tools
• define what kind of metadata to extract and why
• acquire or implement methods described in the
  literature or build our own
• tool/concept tester define and perform
  experiments to demonstrate whether the tools
  solve the problem
• tool/view plug-in framework
• plug-in a tool to create/edit a category of
  metadata, and a corresponding viewer to
  browse/edit the metadata
• plug-in framework an easy way to install,
  integrate, and use plug-ins
• an implementation of this exists see
  http//cs.ua.edu/SEL/Publications/JS20Hawker2020
  02-10-0820-20internal.pdf
• XML-native metadata repository with search,
  browse, etc.
• distributed document and metadata architecture
• Tools, methods, and system administration