Temp Change in Server Room

[Sean_]

So I'm trying to figure out what seems like a relatively straightforward calculation here but I'm having trouble believing the results I'm finding and I'm wondering if you all would be able to provide any feedback.

I'm trying to determine the temperature increase that would occur in a Server Room if all the HVAC systems were to go offline for various periods of time. Here are the specifications:
-The room is 50"x70"x10" and begins at a temperature of 70ºF.
-A Heat Load of 100kw can be thought to be spread evenly throughout the room.
-What will the delta in temperature be after 10 minutes?

My approach was to use
Q = (M * C * (Tf - Ti)) / Time

Where:
Q = Heat Transfer Rate (Kw) = 100KW
M = Mass (kg) = 35,000ft^3 -> (991m^3) * (1.19956kg/m^3) = 1188kg
C = Specific Heat (kJ/kg ºC) = 0.7208kJ/kg ºC
Tf = Temperature Final = ?
Ti = Temperature Initial = 21.1ºC

But when I calculate it all out I'm getting 90.5ºC, which seems ridiculous for such a large room. Can that be right?
I did some research and this thread thread403-100404 seemed to be the closest thing to what I was requesting but I still have trouble believing it, somebody slap some sense into me please!

 

[MintJulep]

Your approach assumes that the walls, floor and ceiling of the server room are perfectly insulated. Is that true?

Your approach assumes that there is nothing with mass inside the room except air. Is that true?

Check your units. Especially the time ones.

 

[AdrienneGC]

Yikes, I just worked through your math and came up with 91^oC.

I wonder if your Q is too high? Are you using nameplate data?

Here is an excerpt from ASHRAE Fundamentals:

Office Equipment
Computers, printers, copiers, etc., can generate very significant heat gains, sometimes greater than all other gains combined. ASHRAE research project RP-822 developed a method to measure the actual heat gain from equipment in buildings and the radiant/convective percentages (Hosni et al. 1998; Jones et al. 1998). This methodology was then incorporated into ASHRAE research project RP-1055 and applied to a wide range of equipment (Hosni et al. 1999) as a follow-up to independent research by Wilkins and McGaffin (1994) and Wilkins et al. (1991). Komor (1997) found similar results. Analysis of measured data showed that results for office equipment could be generalized, but results from laboratory and hospital equipment proved too diverse. The following general guidelines for office equipment are a result of these studies

Nameplate Versus Measured Energy Use. Nameplate data rarely reflect the actual power consumption of office equipment. Actual power consumption is assumed to equal total (radiant plus convective) heat gain, but its ratio to the nameplate value varies widely. ASHRAE research project RP-1055 (Hosni et al. 1999) found that, for general office equipment with nameplate power consumption of less than 1000 W, the actual ratio of total heat gain to nameplate ranged from 25 to 50%, but when all tested equipment is considered, the range is broader. Generally, if the nameplate value is the only information known and no actual heat gain data are available for similar equipment, it is conservative to use 50% of nameplate as heat gain and more nearly correct if 25% of nameplate is used. Much better results can be obtained, however, by considering heat gain to be predictable based on the type of equipment. However, if the device has a mainly resistive internal electric load (e.g., a space heater), the nameplate rating may be a good estimate of its peak energy dissipation.

Adrienne Gould-Choquette, P.E.

http://www.adicotengineering.com

 

[SAK9]

This is not such a straightforward calculation.You have a whole heap of stuff in that room at 21 C. Once the airconditioning goes off,the building fabric,metal racks,equipment all absorb heat and they have a much higher specific heat than air.So the amount of heat that goes into heating air is only a fraction of that 100Kw. It will not get to that 90C mark because once the temperature goes above the surrounding temperature heat will be transmitted through the fabric to the surrounding zones as well and will reach a steady state at some point

 

[EnergyProfessional]

You have to determine your values and what you are actually looking for. I assume you want 70°F space temp, then you need to deliver colder air. Looks like you have a given air flow, so you need to determine the supply temp, which will be below 70°F.

And for the love of science, stay in one unit system. Don't mix metric and IP units. And your room is a small cardboard box, or did you mean to use the ' instead of "? Did you just turn the room volume into a flow rate?

Assuming the servers run 24/7, we can assume steady stete, so the mass isn't important. but any loads from outside will play a role as well (probably minor in high density servers)

 

[BronYrAur]

Imagine a 100kW heater blowing in the space. Would heat up rather quickly. As others have said, there is more in the room than just air, but consider your 35,000 cubic feet volume passing over this 100kW heater in 10 minutes. So take 3,500 CFM as an average airflow (even though you don't have an airflow). Your 100kW is roughly 341,000 BTUH.

341,000 = 3,500 * 1.08 * Delta-T (sorry, but I have long forgotten metric)

delta-T is roughly 90 deg F.

The actual heat transfer is much more complicated since you have essentially have a bunch of radiators that may or may not have internal fans. You will have localized hot spots, but I think the takeaway is that you can't survive long without cooling.

 

[IRstuff]

How much cooling capacity is required keep the room at temperature? That's a check on your heat load.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm