AFTER the smoke clears from a major fire, questions about what happened and what can be done to avoid future conflagrations often linger. That's when Jonathan Barnett's know-how comes in handy. Dr. Barnett, a fire-safety instructor at Worcester (Mass.) Polytechnic Institute (WPI), is an expert in using computer models to reconstruct and study fires. His next project may be a study of the Scandanavian Star tragedy, the North Sea ferry fire that claimed 161 lives in April. He is negotiating with a consultant to the boat's owner about doing a computer simulation of the fire.
``There are several scenarios of what happened,'' Barnett explains. ``We can take those and feed them into the computer and see if the computer gives us the results we expect.''
Once the appropriate scenario is identified, ``what if'' games are possible. That may mean looking at the under trained crew to see if an improved firefighting response would have made a difference.
``Maybe the fire was so large that irrespective of what the crew did, it wouldn't have mattered,'' he says. ``And what if there had been more fire doors? Or what if the interior finish in the corridor had been changed? Would these have made a difference? This is important not only for litigation purposes, but also for the design of new ships.''
The potential uses for this advanced technology include any designed structure - commercial, residential, or public. In fact, Barnett is working with two large subway authorities (he can't divulge the cities) that want to determine how effectively smoke and toxic gases could be vented if a fire occurred.
For small personal-computer type projects, Barnett uses the HAZARD I program developed by the National Institute of Standards and Technology in Gaithersburg, Md. By further developing such programs, says Richard Bukowski, the agency's manager of technology transfer, designers will be able to ``look at fires in buildings that are still on the drawing board. [The architect] can set fires in the building to determine the response of that design, the materials, the layout, arrangements, and so on.''
With the existing program, computers can handle several thousand heat-transfer calculations per second of simulation time, determine a fire's spread, and see what the effect will be upon a building's occupants.
This design tool is possible because of the amount of sophisticated knowledge that exists about the nature of fire. But putting this knowledge to practical use has occurred slowly.
``While the United States is near the forefront in the development of advanced fire technology, it appears to lag behind in its application,'' says David Lucht, director of WPI's Center for Firesafety Studies.
This conclusion, included in a written assessment of the problem, will catalyze discussion at a conference on fire-safety design in the 21st century, to be held next May at WPI. The school offers the country's only advanced-degree program in fire protection engineering. (The University of Maryland will launch a graduate program in September.)
Mr. Lucht says that because so much is known about the properties of building materials and fire dynamics, builders, engineers, and architects can bypass costly and time-consuming lab testing.
A case in point, he explains, involves the steel girders used in high-rise buildings. Since steel beams lose their strength at 1200 degrees F, they frequently are encased in concrete or gypsum to prevent their collapse. To determine how much protective material is needed in a given situation, it's still common practice to actually burn the building materials in a laboratory.
``That's a very expensive way to do business and very uncertain in terms of how well you're doing it,'' Lucht says. ``The technology is pretty well established to use computer models. It's more efficient, accurate, and precise to do it that way. That's the way we design other parts of buildings. We don't go out and see if they fall down.''
Determining what some of the institutional and attitudinal hurdles have been to bridging the gap between theory and practice in fire safety is a major reason for next year's conference.
Lucht surmises that many designers, builders, and public officials are skeptical of new technology. For years, structural fire safety has been dictated by consensus building codes, which have grown out of experience in the field. These are well established and widely heeded, even if not scientifically supported.
``I know if the building code says something and I do it, my liability is much reduced ... even if I'm convinced the code is wrong,'' notes Philip DiNenno of the Society of Fire Protection Engineers. ``The educational problem here is how to get the building or fire department to accept calculations that they don't particularly understand or have never seen before.''
What is now known about fire - from ignition and fire growth to structural behavior under fire conditions - has begun to achieve the status of a science, a trend that is expected to accelerate.
In one effort to break the technical knowledge logjam, the Society of Fire Protection Engineers has published an 800-page handbook that pulls together much of the profession's theoretical and analytical findings. This first-of-a-kind volume makes facts about safe-building design more readily available, says Mr. DiNenno, the book's editor.
Requests for code waivers are one means of nudging the building community ahead. Another, which has been undertaken by Dr. Bukowski's office, is direct education through training sessions for code officials.
ARCHITECTURAL and engineering students learn about the technology at schools such as WPI, the University of Maryland, and the University of California, Berkeley, which offer fire-safety courses.
If these individuals can bridge the knowledge gap, building costs could be shaved by 3 to 5 percent, DiNenno estimates, and with no loss, and perhaps some gain, in a building's fire safety.
``That's really what engineering is about - cost effectiveness,'' Barnett says. Pursuing it, he contends, would lead from ``relatively arbitrary building-code requirements'' to ``much better, performance-specific designs.''
In fire-related designs, however, there's often more than meets the eye.
``During a fire, things change second by second,'' Lucht explains. ``From an engineering standpoint, you have to account for that. It can require very powerful computers.''
Barnett anticipates that studying the Scandanavian Star fire would take two years and employ several computer models.
These are not video games. ``You can't take somebody with no knowledge of fire at all and sit them down at the computer and have them run realistic and proper sorts of analyses,'' says the National Institute's Bukowski. ``It takes a certain amount of understanding.''