Letters: Reader Feedback
Data center fire protection I just read your article entitled “Data Center Technology Past, Present and Future” (CSE 07/07, p. 38). Very interesting. I am a fire protection engineer with more than 30 years of experience in the field. I hope that I am up-to-date on extinguishing agents. As such, I would like to comment on a paragraph found in your informative article.
Data center fire protection
I just read your article entitled “Data Center Technology Past, Present and Future” ( CSE 07/07, p. 38). Very interesting.
I am a fire protection engineer with more than 30 years of experience in the field. I hope that I am up-to-date on extinguishing agents. As such, I would like to comment on a paragraph found in your informative article.
The paragraph (p. 39) reads: “Data centers often employ a common supplemental fire protection system—a clean agent—that releases a gaseous substance that captures the oxygen from the room, and therefore, extinguishes the fire. The gas then turns into a powder, leaving a dust residue behind that is easy to clean up and doesn't damage the equipment.”
I believe the above paragraph contains errors. There are several clean agents on the market today. All work in one of three ways. Some use two of the methods:
Inert gases that displace the air (oxygen) dropping the concentration, such that a fire cannot burn.
Agents that absorb the heat of combustion, thus cooling a fire below a self-sustaining temperature.
Agents that break the chemical chain reaction involved in combustion.
None “captures the oxygen from the room.” I do not know of any oxygen-scavenger agents.
Second to be a “clean agent,” there cannot be any residue at all. The inert gases (argon, nitrogen, CO 2 ) are just that. None of these leave a residue or turn to a powder. The other agents are gases or a volatile liquid that vaporizes quickly and completely. I have not heard of any of them leaving a residue. One agent is rather “oily” as a liquid but it does vaporize completely. I personally have not worked with this agent, so it may leave a film of some kind.
I also would like to take issue with the statement in a statement in an earlier paragraph (p. 39) about “leaking sprinkler heads.” Sprinklers are highly reliable, thoroughly pressure-tested and have an exceedingly low leak rate. They just do not leak enough to be of a concern in a data center.
While what the author says about sprinklers operating at 165°F is true—only the sprinkler in the vicinity of the fire will operate. Sprinklers operate independently and only when they (an individual head) reaches 165°F. To obtain a ceiling temperature of 165°F takes a rather large fire. The operators have to be sound asleep to not sense the fire themselves or see the indications of a malfunctioning piece of equipment. This will be a blazing fire with a lot of smoke, and the piece of equipment involved will already be lost. The sprinklers will prevent the fire from spreading to the rest of the room.
One common solution is to install smoke detection in the room, in addition to the sprinklers. The detection system will sense the fire when it is in a very small phase. An alarm will sound, allowing the operators to find the problem and deal with it. Modern detection systems can tell the operator exactly which detector has alarmed, leading them right to the problem. If the operators cannot deal with the problem, the extinguishing system is allowed to do its thing. Eric A. Dorbeck , P.E.Consumers Energy Jackson, Mich.
Thank you for your feedback, as I agree with many of your comments. The purpose of the article was to draw attention to the cyclical nature of the data center industry and apply across all systems rather than to be an in-depth article for each specialized field. I struggled with the wording of the section leading to several of your comments, for example, through edits to reach a wider audience, the words “captures the oxygen” rather than displacement or other terms that would be preferred in an article written for the NFPA.
After working in data centers for 25 years, first with IBM and now as a consultant, I find that sprinklers often leak, not because the devices fail, but due to regular contact by ladders and personnel during the installation of power or cabling. The ideal data center seldom exists and the vast majority of the operating data centers are conversions or retrofits. Michael Kuppinger , P.E.Senior Vice President, Environmental Systems Design, Inc. Chicago
It is an interesting dilemma that Michael Ivanovich, editor-in-chief, points out in his Editor's Viewpoint, “Engineering Shortage, Employers to Blame?” ( CSE 05/07 p.09). The shortage of engineers has more to do with the field not being sexy—and if you throw not very lucrative on top—why bother?
In its history, the engineering field has been one that holds tremendous integrity and decent pay. That identity seems to be changing. As an engineer, I see other fields that are paying much higher salaries, such as the financial and medical professions, and think that if I were young, I would go into these fields. Because they have more money, they can hire as much top talent as they want.
In a young person's eyes, there has to be a carrot in front of the horse. Unless the salaries increase in the engineering field, I see a continuous shortage with more and more work put on the remaining engineers, thus reducing quality. John Milenius