The Interaction of Roof Vents and Draft Curtains with Standard Sprinklers, Part 2
By Richard Schulte
Schulte & Associates, Evanston, Ill.
Discussion
Perhaps the most important point made in the Beyler/Cooper paper is the admission that "current design practices are likely to limit the number of vents operated to one, and vents may in fact not operate at all in very successful sprinkler operations."
While the issue of whether the operation of smoke and heat vents will affect the activation of sprinklers is still the subject of debate, there is no debate about the significant impact that sprinkler activation has on the operation of automatic roof vents. This point was clearly demonstrated in the research sponsored by the National Fire Protection Research Foundation (NFPRF) at Underwriters Laboratories (UL) in 1997 and 1998 and also in a fire that occurred in a bulk retail store in Tempe, AZ, on March 19, 1998.
In the NFPRF fire tests, an automatic roof vent located directly over a fire failed to operate. The failure of the roof vent was attributed to the "cooling effect" of the sprinkler spray discussed by Golinveaux and Hankins in the excerpts above. Droplets of water from the sprinkler spray collected on the vent fusible links and prevented the link from operating. (The same "cooling effect" would also occur with plastic smoke and heat vents that operate by shrinking and then falling out.)
In the bulk retail store fire in Tempe, the sprinkler protection was inadequate for the hazard protected and the system was clearly failing to gain control of the fire. Although three of the 29 automatic roof vents provided had operated at the time of the arrival of the fire department, the building, which was 100,000 square feet in floor area with a height varying from 24 to 29 feet, was filled with smoke from the floor to the roof. A ladder company had to be dispatched to the roof to manually open the other vents that had not operated. Obviously, sending firefighters to the roof to open the vents manually violated the precautions contained in the NIOSH publication. Interestingly enough, the fusible links for the roof vents had a temperature rating of 165 F, while the temperature rating of the sprinklers was 286 F.
Among the "positive claims" included in the Beyler/Cooper paper are that the use of roof vents in sprinklered buildings allows the venting of the combustion products, smoke, heat, toxic gases and combustible gases generated by the fire. The issue that Beyler and Cooper fail to address is how this venting will occur if no vents operate, or whether the venting will be sufficient if only one vent operates. Logic tells us that if the vents don't operate, due to the activation of sprinklers, then venting will not occur.
The Beyler/Cooper paper also makes frequent reference to the positive effects of draft curtains, which will contain the combustion products. What Beyler and Cooper don't say in their paper is that the IBC/IFC provisions for roof vents (that were under development in 1999) only require draft curtains to be provided in industrial and storage buildings without high-piled storage and that, even when draft curtains are required, the IBC permits the curtained area to be 50,000 square feet in floor area. In other words, all of the benefits attributed to draft curtains by Beyler and Cooper don't apply to buildings constructed to comply with the IBC/IFC provisions for roof vents. So much for the containment of the combustion products by draft curtains.
The Beyler/Cooper paper makes reference to fire tests conducted by Factory Mutual Research Corporation (FMRC) in 1994. Two tests included a combination of sprinklers and draft curtains (without roof vents). Based upon the results of these tests, FMRC concluded that "the presence of curtain boards can cause increases in sprinkler operation, smoke production and fire damage (i.e., sprinklers opened well away from the fire)." In response to FMRC's conclusion, Beyler and Cooper state that "since draft curtains are not normally used in the manner tested [without roof vents], the goal of the testing is unclear." Once again, the logic of the Beyler/Cooper response is flawed, based upon the statement that "current design practices are likely to limit the number of vents operated to one and vents may in fact not operate at all in very successful sprinkler operations." If no vents operate due to the activation of sprinklers, a test of sprinklers and draft curtains without roof vents should yield the same results as a test of sprinklers, roof vents and draft curtains. It certainly doesn't take a Ph.D. to figure that one out. (Interestingly enough, at the ICC code change hearings in Orlando in late September 2006, an engineer retained by the manufacturers of roof vents, William Koffel, of Koffel Associates, a former president of the Society of Fire Protection Engineers (SFPE) and a member of the NFPA 13 and NFPA 204 committees, made the exact same point as Beyler and Cooper about FMRC's conclusions from their tests with draft curtains.)
While Beyler and Cooper tout the benefits of the use of draft curtains in conjunction with roof vents (and sprinklers), the one issue that is not mentioned in the paper is the interference with "prewetting" caused by draft curtains. This interference with prewetting was demonstrated in the NFPRF's fire tests at UL and also in the previously mentioned fire at the bulk retail facility in Tempe. In the fire in Tempe, a draft curtain provided in an aisle (as recommended by Beyler and Cooper) prevented sprinklers on the side of the draft curtain opposite the fire from operating and prewetting combustibles. This lack of prewetting allowed the fire to spread across an aisle that was 10 feet in width.
Another issue that Beyler and Cooper raise in their paper, that "smoke and heat vents operate effectively when sprinklers do not operate," also appears to challenge logic and common sense. Obviously, a fire that occurs in a building that contains an undivided space that is 50,000 square feet or larger, or in a building that contains high-piled storage, will be difficult for most U.S. fire departments to control without sprinkler protection. (That's why building codes require sprinkler protection in industrial and storage buildings that have fire areas exceeding 12,000 SF.) Given the typical delays in reporting a fire and in fire department response, a fire in these types of buildings would typically be too large for interior manual firefighting to be successful by the time firefighters arrive at the fire. With normal fire department response times, the building structure will have been exposed to fire for more than five minutes by the time firefighters arrive, and the integrity of the structural members supporting the roof will most probably be questionable. Under these circumstances, committing fire personnel to perform interior firefighting would be extremely risky at best.
Per the recommendations contained in NIOSH 2005-132, the failure of the sprinkler system caused by a closed water supply control valve, pump failure or a break in the main supply piping in the sprinkler system should result in a change of tactics from interior firefighting to an exterior attack. There is no need to send firefighters to the roof to ventilate the fire if there is a complete failure of the sprinkler system. The fire itself will perform the ventilation function in short order with a partial or total collapse of the roof structure, and roof vents will do little to prevent the relatively rapid collapse of the roof structure.
Conclusion
In September 1999, the American Architectural Metals Association (AAMA) announced a new research project on the use of smoke and heat vents in sprinklered buildings. Given the timing, the announcement appears to have been in response to the research findings in the sprinkler/vent tests sponsored by the National Fire Protection Research Foundation (published in September 1998) and to code change proposals to delete the requirements for roof vents based upon the NFPRF research.
After a delay of almost seven years, the AAMA has finally announced that Hughes Associates Inc. has been retained to perform a modeling study of the interaction between sprinklers and roof vents. According to the AAMA newsletter (summer 2006), the purpose of the modeling study is "to concretely demonstrate the value of S&HV [smoke and heat vents] in terms of property protection, occupant safety, firefighter safety and firefighter effectiveness". The newsletter further states that the "test results will provide data for an optimized S&HV design approach suitable for inclusion in Chapter 8 of NFPA 204, 'Guide for Smoke and Heat Venting,' or in model building codes." It's rather interesting that the AAMA would announce the results of the Hughes study prior to Hughes even beginning work on the project.
Considering that the results of the Hughes study have already been announced by the AAMA and considering the statement in the Beyler/Cooper paper that "while there have been many attempts to model all or part of the interactions of sprinklers and vents, the issues are more complex than can be dealt with using even the most sophisticated modeling methods available today [February 1999]," hopefully, the AAMA has also allocated funds for a peer review of Hughes's research. One can only wonder what will happen if the Hughes research doesn't "concretely demonstrate the value of S&HV." My guess is that, if this happens, the Hughes study will "never see the light of day."
Richard Schulte is a 1976 graduate of the fire protection engineering program at the Illinois Institute of Technology. After working in various positions within the fire protection field, he formed Schulte & Associates in 1988. His consulting experience includes work on the Sears Tower and numerous other notable structures. He has also acted as an expert witness in the litigation involving the fire at the New Orleans Distribution Center. He can be contacted by sending e-mail to rschulte@plumbingengineer.com.