FIRST COMBUSTIBLE METAL-DECK ROOF FATALITIES

BY FRANCIS L. BRANNIGAN, SFPE (FELLOW)

I have been writing on the subject of firefighter safety in building fires since 1966 and have come to realize that some in the fire service are so hung up on experience that they dismiss analysis of the hazard. In 1971 in the first edition of Building Construction for the Fire Service (BCFS), I wrote “Beware the truss” without any experience to cite. We now have a large body of experience with wood truss failures bought at the cost of lives and devastated families.

(1) To the Building Department, this is a noncombustible void. No attention is paid to the CMDR above. (Photos by author unless otherwise noted.)

I defined the hazard of the combustible metal-deck roof (CMDR) fire in 1946 and its tactical solution in 1948 (see BCFS3, 302-308). Highly protected risks (HPR)-trained engineers scoffed at my anecdotal evidence because the roofs were “Approved.” (When someone tells you something is approved, demand to know who approved it for what. Get the report. In the case of roofs, the approval refers to the hazard of making or receiving embers.)

(2) The layered construction of a CMDR. (Source: The former Factory Insurance Association report on General Motors Livonia Transmission plant fire.)

The 1953 General Motors Livonia Transmission plant fire, which resulted in a $75 million total loss from a CMDR fire, got the attention of the HPR insurance industry. The problem has never been considered in any building code of which I am aware (I would be delighted to stand corrected), and buildings can be classified as noncombustible regardless of the roof covering.

(3) A fire delivers 800°F heat to the underside of the roof. (Photo by Chris Brannigan.) Edit this paragraph Assign Image

The firefighter hazard experience has been a couple of close calls that did not provide fatal experiences. Six Dallas, Texas, firefighters were dumped into the fire when elongating bar joists pushed down a concrete block wall. They were close to the edge and were recovered immediately. Twenty St. Paul (MN) firefighters were evacuated 60 seconds before the CMDR collapsed (March 2001 Ol’ Professor). In neither case was the CMDR situation recognized at the time.

(4) The heated tar adhesive/vapor seal liquefies and emits flammable gases. They cannot escape upward because of the layers of roofing. They push downward, ignite, and start a self-sustained fire independent of the contents fire. There is high heat and heavy black smoke. Just �putting the wet stuff on the red stuff� is not enough. The steel roof must be cooled and kept cooled to cut off the gas. (Photo by Chris Brannigan.)

Now we have suffered the line-of-duty death experience, which may cause the fire service to pay more attention to this hazard. I received the following letter from long-time friend, David White, publisher of Industrial Fire World:

(5) Kensington (MD) firefighters cooled the overhead and froze the tar icicles.

I finally got to talk with the Memphis (TN) Fire Department. The building the firefighters were killed in was in a strip shopping center. It was at least 20 years old. The building’s roof was a built-up roof with tar and gravel. The roof was held up by lightweight bar joists. The store had suspended ceiling tiles over the public area of the store. In the rear, where the storeroom was, there were no ceiling tiles.

(6) Ventilation will not help and would be hazardous. This fire is in a test building. Only the roof is involved, no contents. A 50-square-foot vent was required Note: The white structure is a vent house; the size of the vent could be varied. (Photo courtesy of Factory Mutual Research Corporation.)

The fire started in the rear of the building. When the first firefighters arrived on-scene, there was light smoke in the store at the ceiling level.

(7) These Dallas (TX) firefighters were dropped into the fire when elongating bar joists pushed the wall down. (Photo courtesy of the Dallas Fire Department.)

The firefighters were advancing a line into the store and were about 50 percent of the way into the store when the ceiling and the roof came down on them.

(8) This Winter Park (FL) mall was built with an open interior court. It had a CMDR. A concrete one-story addition was erected in the open court. The CMDR fire elongated bar joists, which pushed on a concrete column and caused a roof section to fall, narrowly missing Orlando firefighters on mutual aid. Assistant Chief Mike Kelly of Orlando pulled the ceiling in the entryway, saw the CMDR fire, and killed it with cooling streams.

The fire department thinks that the fire was burning in the rear of the store, flashed over in the ceiling area, and the roof and ceiling came down on them before they could get out.

This information came from the Fire Department’s Training Division.

Comments

Based on my experience and study of combustible metal-deck roof fires since 1946, I believe that the fire developed as follows:

(9) When Mike Kelly became chief of New Smyrna Beach, Florida, he brought this second-story restaurant addition to my attention: The CMDR of the original building is still in place, hidden in the void.

The fire in the rear was sufficient to heat the corrugated steel of the underside of the roof (800°F for five minutes) and start an independent CMDR fire fed by gases generated by the heated tar, which came down through the joints in the steel. The fire was rolling along in the ceiling void, heating the bar joists. When heated enough, the bar joists failed.

The only tactic that works is the following:

(10) This Florida mall is being sprinklered as highly protected risk insurers recommend. Note the lower sprinkler head to cover the contents and the upper head to cover the void above and cool the steel deck of the CMDR.

From a safe location, use a solid stream to blow down the ceiling and cool the steel.1 The safe location is outside the building because the trusses are tied together for mutual support and the collapse may extend beyond the fire-affected area. Sprinklers below the ceiling will not cool the steel. It is a myth that water will cause hot steel to fail.2

Fire companies should survey their districts for combustible metal-deck roofs. Ask if it is a “built-up roof” (layers of tar and roofing material). If informed that it is a membrane roof-single-ply roofing membrane roofs use a single sheet of plastic pasted down-such roofs lack the quantity of fuel to act like a CMDR.3 However, dripping flaming adhesive can set additional fires in the contents.4

Membrane roof installation can provide flash fires. Roll roofing has been used on wooden roofs for years and presents no special hazard.

SKYLIGHTS

There have been reports of skylights presenting roof hazards for firefighters. Skylights used to be universally present in one-story commercial buildings and on top floors. Burglars found them an easy entry; therefore, many have been eliminated. The patched area that replaced the skylight is not nearly as strong as the roof itself.5

(11) The collapse of this �noncombustible� plant occurred with no combustible contents exposed, but liquid combustibles were used in the process. A process explosion set fire to the roof and then subsided. A guard/firefighter jumped down a process hole, knowing he was falling 20 feet, to escape the choking smoke. The Atomic Energy Commission spent $20 million to sprinkler acres of similar roofs in vital production plants. (Photo courtesy of the Atomic Energy Commission.)

Ask if skylights have been removed. If so, locate the patch and note it as a hazard. Ironically, U.S. taxpayers suffered two very costly CMDR fires in government properties when skylights were being installed in CMDR roofs to save on electricity. No one recognized the hazard of torch cutting the roof of a noncombustible building.

Get out with your tower ladder or aerial on Sunday and look down on roofs for hazards such as rectangular patches indicating a removed skylight. Note air-conditioning equipment and possibly materials stored on the roof.

Since there are many such roofs, use the term CMDR; all should know what it means.

Do not attempt to ventilate a CMDR fire. Tests on a pure roof fire with no contents showed that a 50-square-foot opening is required to adequately vent it. Do not worry about water damage. Cool the underside of the roof, and keep cooling. The fire can come back. I saw this at my first CMDR fire. This attack did not work in the Plantation (FL) Towne Mall fire because a second rain roof had been erected over the first leaky roof and the fire was between the roofs.6,7

I continue to recommend that identified firefighter hazards be archived at alarm headquarters and be added to the dispatch.

Example

Report of a fire in a store at 710 Maple; Cross street 7th. HAZARD ALERT: CMDR; also machinery on the roof.

Written preplans should have this type of hazard up front. All hands will be aware of the hazard. It takes work to set up such archives, but it beats having to announce firefighters’ line-of-duty deaths.8 ■

Endnotes

1. See photo on page 302 of Building Construction for the Fire Service, Third Edition (BCFS3).

2. See BCFS3, 259-260.

3. See BCFS3, 35-36.

4. Montagna, Frank, “Why Do We Do That?” Fire Engineering, Dec. 2004.

5. See photos in Ol’ Professor, July 2001. A Fire Department of New York (FDNY) firefighter fell through a patch and suffered crippling injuries. FDNY Battalion Chief Jerry Tracy submitted the story and photos.

6. Casey, David, “Plantation (FL) Towne Mall Fire” (1996), Fire Engineering, April 1997.

7. At my second metal deck roof fire, the firefighter operating the deck pipe cooling the roof shifted the stream to hit a burning pile of stock. I ordered him to return to cooling the roof and requested a handline for the pile fire. It turned out to be the personal property of the Commandant of the Marine Corps (USMC). I told his aide that if the Commandant had asked my advice, I would have recommended storing the belongings in the sprinklered Navy warehouse across the street. He said, “What’s a sprinkler?” The Quartermaster General of USMC “knew” that sprinklers did more damage than fire. Every USMC Depot of Supplies suffered a major fire. At Norfolk, we had two.

8. I would appreciate your telling me what you have done about this recommendation, or your reasons for not doing anything. Send comments to FBrannigan@ comcast.net/.

■ FRANCIS L. BRANNIGAN, SFPE (Fellow), the recipient of Fire Engineering’s first Lifetime Achievement Award, has devoted more than half of his 63-year career to the safety of firefighters in building fires. He is well known as the author of Building Construction for the Fire Service, Third Edition (National Fire Protection Association, 1992), and for his lectures and videotapes. Brannigan is an editorial advisory board member of Fire Engineering.