Modeling Typical IT Rooms AC Units

9 posts / 0 new
Last post

Hey all,

I'm working on a LEED energy model for an office building that has a couple standard sized data rooms for server storage. They will be served by a couple air cooled Liebert units. I started inputting information about the units in the proposed design model and requested EER information from our vendor. The EER value was provided but also pointed me towards ASHRAE 90.1-2007 Section 2.3 as the reason why EER values aren't typically provided for these types of units (he claims the cooling of the server rooms would be considered process cooling):

ASHRAE 90.1-2007 Section 2.3 - The provisions of this standard do not apply to:

a. Single-family houses, multi-family structures of three stories or fewer above grade, manufactured houses (mobile homes), and manufactured houses (modular)

b. Buildings that do not use either electricity or fossil fuel, or

c. Equipment and portions of building systems that use energy primarily to provide for industrial, manufacturing, or commercial processes.

I've typically model the scheduled units on the proposed design model and System Type 3 or 4 on baseline model. However, if the above Section 2.3 applies, should we be modeling the systems that serve IT rooms in an office building as "process energy" and be equal?

William Mak, LEED AP BD+C

Will Mak's picture
Offline
Joined: 2011-09-30
Reputation: 200

Hi Will!

We were almost burned hard specifying and approving Liebert for a LEED project under similar circumstances. Despite their large presence for a niche market, their efficiencies aren't entirely up to snuff with today's commercial energy codes.

We ultimately avoided having to send the equipment back (and foot the bill) in that case, but only because we were able to reason with both the local 90.1-compliance, LEED reviewers and other parties in play that this was a reality of the industry we had to live with. We did not attempt to "dodge" the code per the advice you were given below, but I do feel we were lucky. It takes no stretch of the imagination to picture an uncooperative AHJ/reviewer striking this down. The building altogether was an impressive performer, to be sure, but these isolated data room units left a sour taste. I will personally think hard about compliant alternatives before falling back to designing around Liebert for future LEED work.

That said, the approach given below is a new one to me. I also tend to model the proposed systems as they are, and apply a baseline system to the corresponding baseline zones. I haven't tried calling data room systems process loads before, but I suppose it could work if you're consistent in applying/documenting it that way. You might still find headaches trying to simultaneously document prescriptive compliance.

I'd be interested to hear others share any related experience =)!

~Nick

NICK CATON, P.E.

Nick-Caton's picture
Offline
Joined: 2011-09-30
Reputation: 805

Ironically, I'm in almost the same case. The Liebert units on this job have efficiencies that are below the ones listed in ASHRAE 90.1-2007.

The approach to model these systems as process loads is great in theory but you and I know it's a pain in eQuest, especially since you are now trying to mimic the same energy consumptions, when other factors like lighting are different...

William Mak, LEED AP BD+C

Will Mak's picture
Offline
Joined: 2011-09-30
Reputation: 200

Liebert is no less efficient than any of the other computer grade self-contained units out there. The problem is that the units have to fit the niche that has been created for them. They need to be no more than 3' deep to work with typical server room design, they have to generally have high static fans to pressurize under floor systems, and they have to be fully serviced from the front of the unit. All of these have implications on the evaporator / condenser coil layout, and thus efficiency.

There are certainly things that you can do to improve computer grade AC unit design (like using the ECM motors that have been popping up for a lot of manufacturers, and playing around with higher space temperatures), but ultimately, you're never going to have a Liebert (or Stulz, etc) unit with an EER of 18.

Also, in almost all cases, the efficiencies you are getting from the manufacturer or manufacturer's rep are probably based on the abnormal operating conditions of a server room (68, 70, 72 degrees) - make sure you are properly accounting for this in eQuest, as these are not ARI conditions.

James Hansen, PE, LEED AP

James Hansen's picture
Offline
Joined: 2011-09-30
Reputation: 200

Will:

I couldn't agree more with you on Section 2.3. However, what matters is not what I think, but the LEED/USGBC review team.

In my experience, I've modeled data rooms in offices as a System 3/4 for the baseline and as scheduled for the design case. I've modeled small rooms with a mini-split DX unit, and also for larger rooms with CRAC style units (Liebert, DataAire, etc). I haven't had any issues with this in the review process. In addition, from some internet research, it appears that architects/engineers are getting large-scale data centers certified, so they must be wanting us to split data center HVAC usage from server/process usage.

Sorry that I do not have any more insight than this, but I do understand your frustration with data rooms/centers and applicability to 90.1.

I would also point you to some addenda for 90.1 that does mention data centers and outlines some requirements (addenda BU). It looks like these Liebert units would meet the requirements for efficiency in this addenda. For the purposes of LEED certification, I would caution to read all addenda, since LEED doesn't let you pick and choose addenda, there may be some addenda out there that might hurt your energy savings percentage.

Andy

Andy Phelps's picture
Offline
Joined: 2011-09-30
Reputation: 0

I should be quick to emphasize: I'm only relating some personal experience here for others' benefit and to constructively add to the discussion. I appreciate James's perspective and accept there are probably a few good reasons for the way things are in the world of DX data center cooling, considering the environment and physical limitations. It seems obvious to me that 90.1 was not built to consider minimum efficiencies for this sort of equipment/environment, but that won't stop others from trying to enforce the standard in spite of my opinions. The section 2.3 reference is starting to make a lot of sense from this angle...

That said, I don't think it's entirely unreasonable to expect some progress from the commercial industry either. For smallish IT closets these days, perhaps a residential split with a high EER/SEER is really the way to go. For larger cases requiring high-static distribution or heavier capacities, I'm only advising that it never hurts to take stock of who/what's out there and what might fit the budget/performance/efficiency requirements for the situation at hand. If nothing truly "fits the bill," at least we can rest easy knowing that's the case and not wonder whether we missed a better option.

An extra bit for Will and others considering HVAC systems as "process loads:" this isn't always so tricky as it might seem =)! Rather than construe a custom fractional schedule to approximate the system's energy consumption (like a plug load), there's another approach: submeter these systems and report separately on the LEED template... Report PS-B will show the monthly consumptions/demands for the submeters after setting it all up. Following are some quick/rough screengrabs illustrating the concept/setup:

~Nick

NICK CATON, P.E.

Nick-Caton's picture
Offline
Joined: 2011-09-30
Reputation: 805

To add to all the fun and confusion, i had this same conversation a few
weeks ago with other LEED people in my office. My interpretation is that
equipment rack room and IT closets fall into an exception of 90.1, more
specifically Section G3.1.1 Baseline HVAC System Type and Description:

Exception b. If the baseline HVAC system is type 5, 6, 7 or 8, use separate
single zone systems ... for any spaces that have occupancy or process loads
or schedules that differ significantly. ... Examples where this exception
may be applicable include, but are not limited to, computer server rooms,
...

Per this statement, you need to model the IT HVAC equipment as its own
system type 3/4 and give it the appropriate efficiencies. Unfortunately
this means that most LEED models with IT spaces that are small get penalties
due to the equipment, as everyone has suggested.

Hope this sheds more light!

--Rob

rdh4176's picture
Offline
Joined: 2011-02-14
Reputation: 0

Yeah, it seems if you take the 2 interpretations we've discussed here, you will model the equipment scheduled on the proposed design model and system 3/4 for the baseline model, and essentially take a hit for the associated energy consumption.

If the review team questions the efficiencies, just reference Section 2.3.

William Mak, LEED AP BD+C

Will Mak's picture
Offline
Joined: 2011-09-30
Reputation: 200

To muddy the waters even more: Recently I'm also seeing a huge penalty on the proposed case with systems 3 and 4 being required to use airside economizers. If one cannot lump the computer, elec/idf room AC equipment into process loads and it is not feasible to place airside economizers on the proposed equipment is there a way around this? Perhaps the answer is modeling active humidity control in these spaces which would in turn penalize the base case systems.

Jansen, Connor's picture
Offline
Joined: 2011-09-30
Reputation: 0