The fiery collapse of three World Trade Center skyscrapers has led scientists, engineers and building officials to question the century-old practice of relying on tests done in hulking, low-tech furnaces to determine if building materials will survive out-of-control blazes.
The furnace tests, conducted at places like Underwriters Laboratories here, focus on the ability of separate building components -- a steel column or a concrete roof support -- to survive temperatures as high as 2,000 degrees.
The test has many defenders, but many scientists say that even before Sept. 11, they had concluded that the furnace tests were out of touch with a computer-age understanding of the devastation caused by the fury of a real fire. At least when it comes to buildings, that research shows, the whole often acts differently from a simple-minded sum of the parts.
The super-baked, seared furnaces here are the seething heart of the most trusted and advanced testing laboratory in the United States for building structures, materials and products. But within those same furnaces, Underwriters and other labs are using testing methods that have changed little since they were devised when the first Model T Fords were rolling off the assembly lines.
A 12-foot steel-and-masonry door slides shut, closing off a brick-lined furnace, its red-painted exterior corroding around a maze of pipes, flanges and valve handles. Inside, a vertical steel column is braced with concrete supports, a web of wires and sensors strung between its surface and the furnace wall like clotheslines. A worker lights a kerosene-soaked rag wrapped around the end of a steel bar and shoves it through a small window in the door, igniting 70 yellow billowing flames that lick the column. Within three minutes, the temperature has shot past 1,000 degrees and is still rising.
The test is on, and its fidelity is anything but an academic issue. This and similar tests shape what amounts to the DNA of the nation's infrastructure, what goes into every commercial and large residential building.
The basic test of the fire-resisting capabilities of fireproofing, steel supports and concrete structures -- a furnace procedure called ASTM E-119 -- has many proponents who say that the declining number of fire deaths in the United States since the mid-1970's proves its effectiveness.
But critics call it unrealistic, primitive, even medieval. And other countries, including England, Australia, New Zealand, Sweden and Japan, have modified fire standards and codes to allow designers to use 21st-century computer tools to custom design buildings to withstand unusual or intense fires not anticipated by the traditional regulations.
Tests in this country do not, for example, account for plastic and synthetic combustible materials, which were unknown in the 1910's when the ''standard fire'' used in the test was developed. Nor does the test try to mimic how a real fire sweeps through rooms in a building, creating structural stresses in one place that can lead to failures elsewhere in the building's interconnected skeleton. Within the cramped furnace, the test involves individual beams or columns and does not even check whether, once mounted in a building, the bolted, welded or riveted connections between them are equally resistant to failure in a fire.
''No one who is in the business with any type of technical background believes there is any kind of strong correspondence, in any way, between the performance test and a real fire,'' said Kathleen Almand, a structural engineer who is executive director of the Society of Fire Protection Engineers.
None of those engineers and scientists, and certainly not Ms. Almand, have suggested that shortcomings in the test led to the collapse of the twin towers and 7 World Trade Center, a 47-story skyscraper that burned and collapsed under mysterious circumstances late on Sept. 11.
No modern high-rise had ever completely collapsed in a fire before that day, a fact that along with the steadily decreasing fire death rate has reassured building officials and engineers that the furnace tests, whatever their technical failings, do pretty well in winnowing out dangerous or inadequate materials and structures. But as research and computer modeling of fires advanced in recent decades, scientists and engineers have become acutely aware that the furnace blasts were more fantasy flame than accurate road test of conditions in a burning building. The same researchers say they are concerned that recently developed construction materials like lightweight, high-strength concrete and composites could survive furnace tests but be prone to complex failures.''If I say, 'That product passed the test,' you as John Q. Public may be happy,'' said Dr. James G. Quintiere, a professor in fire protection engineering at the University of Maryland. ''But if we say, 'It passed the test, but we don't know how it will function in your house,' you would not be happy.''
The Science of Fire
Even before Sept. 11, about 20 percent of firefighters who died in building fires were killed in localized collapses of one kind or another. So fire experts say a true scientific understanding of how fires affect large structures could help determine when it is too dangerous for firefighters to be in a burning building.
Dr. Shyam Sunder, a division chief at the Building and Fire Research Laboratory of the National Institute of Standards and Technology, said that developing the science of fire beyond the furnace tests could save lives and have an impact in at least three broad areas: restoring public confidence in tall buildings, protecting firefighters and other emergency personnel and improving safety for building occupants.
The institute, which does research and provides the technical basis for standards but does not impose regulations, is planning to undertake a $16 million follow-up investigation on the trade center disaster after a federal inquiry, led by the Federal Emergency Management Agency and the American Society of Civil Engineers, releases its own report in late April or early May.
The institute has also begun a wider research and development effort with the goal of improving building and fire codes and engineering practices for structures of all kinds.
If the institute's work is successful, Dr. Sunder said, ''we'll have the tools for engineers to design and retrofit for fire safety.''
''Fire and emergency service personnel will have a better understanding of the performance to be expected from buildings,'' he added.
A draft of the FEMA report, which has been obtained by The New York Times, has already placed a sharp focus on the ASTM E-119 test, though the report is careful not to imply that the test is somehow responsible for the collapses. The test ''does not provide tools to determine how long a building component can expect to perform in an actual fire,'' the report says.
An Oversized Barbecue
At the Underwriters Laboratory campus in this northern Chicago suburb, where workers carry out those blazing tests in a yawning building reminiscent of an airplane hangar, forklifts shuttle back and forth, moving huge columns, while cranes suspended from overhead rails slowly lift giant sections of roof into the furnaces and technicians mill about in front of a forest of dials, gauges and controls.
Three huge furnaces, each in a separate corner of the hangar, are designed to handle three different tests: one for steel columns, one for wall assemblies, and one for ceilings and floors. Little about the setup is high-tech.
''Do you need the torch, Bill?'' Eckhard Brodde, an assistant, yelled to the foreman, Bill Joy, as the two prepared the cube-shaped furnace for testing columns last week.
''Let's close the door first,'' Mr. Joy yelled back. Both gave it a heavy push. ''It is not moving,'' Mr. Joy said, before an extra shove was applied. ''There you go,'' he said, as the door clicked into place.
As if lighting an oversized barbecue, Mr. Joy sets fire to a kerosene-soaked rag and sticks it through a small opening in the furnace. ''Three, two, one -- gas on,'' he shouts, as the flames burst into life.
With this inferno of flames racing from dozens of gas jets, these workers are determining just how long the steel column and its fireproofing can hold up under this thermal assault.
The result is a fire rating, expressed in hours, for this column, and eventually for countless different components that make up a building, from sections of roofing to walls for emergency exit stairways to assemblies of glass, brick and plaster that make up a building facade. A column might be given a ''two-hour'' rating, meaning that it has withstood two hours in the furnace at temperatures that research done in the 1910's determined to be typical of office fires of the day.
A history book on U.L.'s testing procedures, published in the mid-1920's and on display in a glass case at a company museum here, describes the procedure Mr. Joy and Mr. Brodde are following, almost as if it were written yesterday.
''The furnace itself is a box of thick masonry, and is heated by four rows of blast burners,'' the book says. ''On the outside of the walls are valves, through which it is possible to control the distribution and intensity of the fire.''
Just as in the 1910's and 1920's, an organization called the American Society for Testing and Materials, or ASTM International, specifies exactly how those tests are done. Ultimately these ratings are incorporated into local and state regulations, or fire codes, which dictate minimum standards for each of the thousands of components in a building.
A Comprehensive Approach
The proponents of using modern science are advocating an extremely different approach using a comprehensive evaluation of how an entire building reacts -- and perhaps fails -- in a fire.
This approach, for example, might identify a critical connection between beams and columns that could become the weakest link in a major fire and therefore has to have much stronger fire protection. It might also find ways for developers to cut costs by reducing insulation where it is less needed.
But any change is likely to come slowly, particularly because furnace tests received little scrutiny outside government research institutes and university labs before Sept. 11. The furnace test ''will tell you if a product does work or doesn't work,'' said Neil Schultz, executive director of VTEC Laboratories in the Bronx. ''You can have all the computer models in the world and they wouldn't tell you that.''
But even at places like Underwriters Labs, ASTM International and the government agencies where fire codes for cities including New York are written, technicians and officials say they are ready to consider the new philosophy.
''Irrespective of 9/11 and the World Trade Center, I think the fire science, the sophistication level, and the understanding is such that the momentum has started to look at these things more precisely,'' said Arthur E. Cote, chief engineer at the National Fire Association, an organization that helps develop codes.''The sort of inevitable march,'' he added, ''is ongoing. The pace will depend on economics and the reward from the effort.''