Towards Realtime nD Modelling

1
Towards Real-time nD Modelling
Joe Tah Faculty of the Built Environment South
Bank University London.
2
Previous research

• Integration of
• 3D-CAD
• Project management
• Using
• UML-based OO modelling
• OODBMS
• KBS Multi-Agent Systems Architecture

3
Object Models

• Integrated project object model
• A project object model
• A structural system object model
• A project task system object model
• A resource system object model
• A productivity object model

4
Knowledge representation

• Rule sets
• Setting structural components material specs
• Elaborating component tasks into units of work
• Determining the amount of work in tasks
• Selecting appropriate concreting methods
• Selecting resources for concrete placement
• Selecting resources for fixing reinforcement
• Selecting resources for erecting formwork

5
Real-time 4D Modelling MicroStation/Microsoft
Project Integration
6
Lessons

• OO modelling provided the basis for a common
  language for database KB
• Systematic knowledge elicitation and
  representation is key
• Distributed computing standards absolutely
  essential for real-time processing

7
Emerging issues

• IFCs appeared to be maturing
• OODBMs failed to gain widespread use
• Timely to revisit relational databases

8
Recent Work

• Build a project-model database on a relational
  database using IFCs
• Explore data exchange between project-model
  database and MS Project
• Make recommendations for future work

9
Methodology

• Identified a relevant subset of IFC classes
• Re-modelled IFCs in UML
• Map classes to the relational model
• Implement relational database
• Develop a connector to MS Project
• Use storey building case study to explore data
  exchanges

10
Implementation

• Map classes into relational tables
• Implement in MS SQL Server 2000
• Use VB.NET to develop connector to MS Project
  2000 based on ADO

11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
Findings

• Structure of IFC suitable for automated data
  management and self-administration.
• Overwhelming difficulties associated with
  implementing databases for use in IFC-compliant
  user-managed databases

15
Recommendations

• IfcProcessExtension unnecessarily complex and
  could be simplified
• Practical requirements for planning, scheduling,
  and control software not fully supported and an
  update is necessary.

16
Conclusions

• Development necessarily evolutionary.
• OO Standards-based common language is key, hence
  role for IFCs.
• The potential for KBS is enormous and yet to be
  fully exploited.
• MAS appear to provide a framework for integration
  in a Web Services future?