Research has shown that audible fire alarm signals that might reliably alert other age groups are less reliable at alerting children and young adults.1 When additional risk factors common to college students are considered, such as stress, sleep deprivation and possible alcohol or other drug use, the design of an effective fire-alerting system becomes challenging.
Increasing the loudness of an audible signal is an option - up to a point. The National Fire Alarm Code, NFPA 72,2 requires that the audible fire alarm signal in sleeping areas be "at least 15 dB above the average ambient sound level or 5 dB above the maximum sound level having a duration of at least 60 seconds or a sound level of at least 75 dBA, whichever is greater, measured at the pillow level." However, the code also limits the maximum sound pressure level produced by the ambient sound combined with all audible notification appliances operating to no more than 110 dBA. Most residential sleeping areas have ambient levels of 40 to 55 dBA or less depending on the presence of window air-conditioning units. A complicating factor is that many college dorms are not used solely for studying and sleeping. Social activities in college residences add large numbers of people and loud music that challenge the effectiveness of the fire-alarm system. So, while a 75 dBA level at the pillow might awaken most persons in a short period of time, the distribution of awakening time may be greater in college dormitories and may exceed the system goals.
How should occupant notification be designed? The answer is both complicated and simple. First, it is necessary to recognize that it is not reasonable to expect 100 percent arousal and alerting from a fire-alarm system within the time frames necessary to achieve the desired fire-safety goal under all fire scenarios. Previous articles in this series have introduced the Fire Safety Concepts Tree from NFPA 550 as a tool to understand different paths to achieve fire-safety goals.3,4,5 (See Figure 1.) The Fire Safety Concepts Tree is an event tree using logical AND gates (multiplication symbol) and OR gates (addition symbol) to relate various combinations of subevents that lead to the top level successful event. An AND gate indicates that all events immediately below that event must occur for the event to be a success. For an AND gate, the probability of success is determined by multiplying the probabilities of the lower events. For an OR gate, the probability of success is determined by adding the probabilities of the lower events.
Examination of Figures 1 and 2 show that an effective fire-detection and alarm system is an important system for safeguarding the exposed people and property. However, Figure 1 also shows that other fire-safety features such as construction and suppression can be combined with detection and alarm through an OR gate to achieve the fire-safety goals. The triad of construction, suppression and detection/alarm has long been the mainstay of a balanced approach to fire protection. Building codes and fire codes rely on some combination of these elements to achieve a minimum level of fire safety for different occupancies or use groups.6 As shown in Figure 3, these three fundamental systems can be thought of as three pillars supporting a fire-safety goal. Depending on the hazard and the resulting risk, some designs may rely predominantly on two of the three elements. In other cases, a designer or code may use all three in some combination. Where one system contributes less to the probability of success, one or more of the others must carry the additional burden.
In part because of the challenges faced by occupant-notification systems in college dormitories and because of the higher degree of risk to life, there has been a move in many jurisdictions to require automatic sprinklers for both new and existing dorm buildings. Strengthening the suppression pillar/event balances the fire-safety goal in situations where occupant arousal and alerting are more difficult. This does not eliminate the need for fire detection to summon emergency forces and to provide occupant notification. But it does allow more time for occupants to be awakened and alerted, and to initiate the desired behavior - escape or defend in place depending on the fire-safety plan for the property.
Strategies to increase the reliability of occupant notification in dormitories include the use of voice alarms with well-crafted automatic and manual announcements, better distribution of audible signals and an increased reliance on visual signals. 7, 8 While most new systems include strobe lights in common areas, the use of strobes in sleeping areas is typically reserved for spaces used by hearing-impaired persons. However, by adding high-intensity strobes to all sleeping spaces, notification effectiveness is enhanced for all occupants. In construction of new dormitories, automatic control of fixed building lighting may also be used as a supplemental alerting method to enhance notification.
Overcoming the "Cry Wolf" Syndrome9 is a major challenge on college campuses. In dormitories, false alarms create complacency. In research facilities, any alarm is an "inconvenience." False and nuisance alarm reduction or elimination is very important. Similarly, the impact of testing occupant notification should be minimized by developing strategies that concentrate audible/visible appliance testing and involve occupants in the testing. It is also important that the college or university have a strict policy regarding vandalism and the initiation of false alarms.
A great way to ensure proper occupant response to an alarm is to have a zero-tolerance policy regarding non-evacuation. On any alarm, the policy should be that the system will not be silenced and the occupants cannot go back into the building until the fire department has verified that the all occupants have evacuated.
One large university implemented such a policy after a new fire-alarm system began experiencing numerous occupant-caused nuisance alarms. The problem was that the new design resulted in a certain type of system-connected smoke detector being in the center room of a dorm suite. As soon as students arrived in the fall, the quiet system began experiencing as many as five nuisance alarms per day due to cooking, which was actually not permitted. The university immediately chose to redesign and modify the system over the next holiday break. Until then, they appealed to the student body to not cook in the rooms, and they implemented the 100 percent evacuation policy. Even if a student or resident advisor met the fire department at the front door and said that they accidentally activated the system and showed them the detector in alarm, the policy was to not shut the system off until everyone was outside and the entire building was swept by the fire department and security. Nuisance alarms immediately dropped to only a few per week. After the redesign, they were virtually eliminated. Several months later, a vandal pulled a manual station in another dorm building in the early morning hours - the night before finals began. Everyone left. An hour or two later, a heat detector was set off using matches. Everyone left. About 45 minutes later, an arsonist started a fire in an unoccupied dorm room. Everyone left. No one was injured, and the fire department contained the fire to the room of origin. Smoke doors closed and contained the majority of damage to part of one floor. A student interviewed the next day stated that she suspected another false alarm - after all, it was the third that night. But she knew she would not get any more studying done or any sleep if she did not leave since the system would not be shut off and she would be standing outside the building.
*Reprinted with permission from NFPA 550-2002, Fire Safety Concepts Tree NFPA 550-2002, Fire Safety Concepts Tree, Copyright 2007, National Fire Protection Association, Quincy, MA 02269. This reprinted material is, Copyright 2007, National Fire Protection Association, Quincy, MA 02269. This reprinted material is not the complete and official position of the NFPA on the referenced subject, which is represented only by the standard in its entirety.
References
1 Bruck, D., and Horansan, M., "Non-arousal and Non-action of Normal Sleepers in Response to a Smoke Detector Alarm", Fire Safety Journal , 25, 1995.
2NFPA 72, National Fire Alarm Code , 2007 edition,National Fire Protection Association, Quincy, MA, 2006.
3 NFPA 550,Guide to the Fire Safety Concepts Tree, 2002 edition, National Fire Protection Association,Quincy, MA, 2007.
4 "SIP & DIP - Stay In Place & Defend In Place," Fire Protection Engineering, Winter 2007.
5 "Fire Alarm Systems and Interior Finish " A Balanced Approach," Fire Protection Engineering, Fall 2004.
6 "Codes and Standards and AHJs " Oh My!," Fire Protection Engineering, Spring 2007.
7 "Messaging and Communication Strategies for Fire Alarm Systems," Fire Protection Engineering, Summer2003.
8 "The Mosquito and the Picket Fence " A Modern-Day Fire Alarm Fable About Broadband versus Narrowband Signaling, Part 3," Fire Protection Engineering, Summer2005.
9 "Fire Alarm Testing Strategies Can Improve Occupant Response and Reduce the Cry Wolf Syndrome," Fire Protection Engineering, Fall 2003.