The problem with computer modeling is not the calculation of reasonably accurate results but the sheer amount of reliable data that has to be input. A Building Information Model has the potential to provide the majority of the data required for structural, thermal, fire, lighting and acoustic simulations, although additional site information also needs to be modeled to provide data on external influences. In the past, expert systems have been developed to provide design guidance for insulation. Artificial intelligence has been used to test the effects of fire and fire fighting strategies. This could be extended, using an object orientated approach, to look at the structural effects of fire, terrorism, earthquakes or tsunamis. Combining all of these areas into an intelligent design assistant, familiar with relational databases, that can autonomously access the main BIM database, any databases providing BIM objects and manufacturers specific data and other information found by an internet spider, would allow the assistant to examine the design in detail and take the software’s abilities far beyond simple clash detection. Not only could the expert side of the system offer suggestions for cost cutting and improved construction but virtual actors could be placed within the virtual building and test its layout, signage and comfort. A fire model could take data from BIM and use it to predict fire spread, temperatures reached and fumes produced in each and every space, this could then be countered by a variety of fire fighting strategies from sprinkler systems through to virtual fire fighting personnel. The egress of virtual users of the building could test the provision of adequate fire escapes and the effectiveness of signage, including the obscurance effects of modeled smoke.
|Keywords:||Modeling, Simulation, Artificial Intelligence, Interaction, Calculation, Expert System|
Architecture Technician, School of Architecture, University of Portsmouth, Portsmouth, Hants, UK