Printable View
A client has asked for a running cost and payback comparison between Energy saving DX system with Glycol filled dry cooler circuit and additional free cooling coil and a standard DX system with remote condenser :eek:
The system is for a Data centre and would run 24/7
Does anybody know of a site or program that could assist with this comparison?
Cheers
Jasper
Energy Saving DX system?
I've seen a data centre using CCu's with dx system and watercooled condensor and second coil to use the water or glycol direct for cooling if temperature is low enough.
But the amount of money you save largly depends on th size off the system.
If it's a couple of houndred kW's I think it's cheaper to have two chillers, water cooled and drycoolers with a heat exchanger for free cooling.
Cost if the installation is higher, but maintaining it is much cheaper.
From your (brief) description, I assume that you customer is looking at the Emerson/Lieberth concept?
I have worked on a site with that system installed, 4 downflow close control units with twin DX circuits + additional "free cooling" coil in each unit.
Before I went to that site, I had read the sales bumf from Emerson and thought these systems would be the dogs shiny bits.
In real life?
Forget about it!
During normal working conditions, the compressors are pumping heat in to the drycooler loop, then when the loop gets below ~10 C, the "free cooling" coil will kick in and take over from the compressors.
Think about it, what ambient conditions do you need to get the drycooler loop down to 10C whilst the compressors are pumping in heat in to it? (assuming that you don't want to fill the roof or car park with drycoolers...)
I have another idea, that I believe would have the possibility to pay for itself well within it's lifespan.
Install normal CW close control units piped up to the most energy efficient chiller you can find.
A bog standard system BUT before the CW returns to the chiller, install a diversion valve feeding the biggest drycooler you can fit. As soon as the ambient temp drops below the return water temp, divert the water through the drycooler before it goes in to the chiller.
That way, as soon as your ambient is below ~12C your customer would start saving money and when the ambient hits 0C, the cooling should be "free".
ther ar two ways to work with energie saving ,both with water/glycol (we work with uniflair)
1**with a drycooler you can make condensor water/glycol 28c or wen the outsite temp is lower then 15 c you can switz over to energie saving (full fans ) to make 7 c water/glycol .
the computair has one sensor at te water/glycol temp entry ,as the delta temp (glycol - airtemp room)+/- 5 c energie saving modulating valve can open on request.
also can energie saving valve and mecanical (compressors)work both.
best energie savings with oversize drycoolers en computairs
2**you kan also work with chilled water computairs (inside)and on the roof chillers with free-cooling coils to the condensor coils ,as the out site temp is low you can start the fc pump that pumps the water/glycol trouth the free-cooling coils and later trouth the chiller coil .
also both systems can work.at the same time.
the 1** methode works with outsite temp < 15 c
the 2** methode works with outsite temp < 8 C
also think about redundensi of the water/glycol system ,then it wil cost extrymly more dan dx units that can work stand alone .
for running and cost comparison y wil ask the quis at the office .
sorry for english.
use chilled water "liebert" ccu's with ec fans, running at temps of 12degC / 16degC flow / return.
chillers with integrated free-cooling coils, again liebert can offer these, with same flow/return temps - high energy savings and happy client.
for running and cost comparison y wil ask the quis at the office .
APACI1
I look forward to the running and cost comparison.
Many thanks for your reply
Jasper
@ APAC1: Hé weer een kaaskop op het forum!;)
Hi
Reading these post's with great interest.
I have just started an energy audit on a serieys of switch sites, 24/7 high density heat load, (750watts m2) Pressurised rooms using, AHU for 2.5 Air changes/hr (100% Fresh air, heater battery,cooling coil, humidfier,) The contained switch area is controled with 15 down-flow units (in built DX and dry cooler coils) Dx takes control after external of 12Deg C,
I propose as an energy saving exercise to,
1) Re-define the set point criteria Temp Set point criteria from 21degC 45%-50% RH to 24Deg c Humidity - 20-60%RH
Does any body have any information of the effects of low humidity in computers rooms and the affect upon static???
2) Limit Pre heat and humidifier on AHU and effectively use as free cooling unit, increaing the air flow to maximise the effect, down to temp of 7 deg c (no lower because of the need to humidfy)
3)!!! This is were I get stuck, I want the dry cooling system to achieve so much more, what is a resonable temperature difference to expect?
Any other thoughts would be greatly recieved sorry abou the long post first one!!
I've tryed modeling the room using IES and used the BIN bethod with frequency data to get a firgure for all savings if that helps jasper.
cheers,
I am retired now but when I was working I worked on a number of free cooling sites on data and telecoms rooms. When you consider that these type of rooms generally only require sensible cooling and the number of days in the UK per year that are below 10c then a free cooling option is a good idea. You can have either a water chiller with free cooling coils built in or a seperate drycooler with 3 way valve using a back to back dx/water coil. Talk to companies like Airedale or custom coils for information. I am not certain if retrofitting would always be advised.
Another way of looking at this is to use a refrigerant thermosiphon. When the outdoor ambient is lower than the desired evaporating temperature the system can run by itself without compressors.
Talk to Star Refrigeration. I believe they can do this for you.
Quote:
Originally Posted by The Viking
I work for Emerson & reading your post suggests that you have no idea as to how the "freecooling" strategy works.
It doesn't simply look at the free cooling medium's temperature & then change over operating modes. The parameters should be set up to cross reference the return air temperature & that of the free cooling medium. Should that free cooling medium be at a temperature sufficiently below that of the return air temperature to be of use, it would use free cooling as the primary source of cooling. Should the free cooling then not cope with the room conditions, the free cooling is disabled for an hour where any cooling demand is met by the DX circuit(s). Of course, if that free cooling medium is also the condenser water loop as is sometimes the case, then it follows that if the drycoolers are only sized to give sufficient heat rejection of the DX circuit(s), then the likelyhood is that the water will then be too warm to be of use. I have seen this situation on a number of sites & the problem usually lies with a combination not enough spare cooling capacity in the room to allow the circuits to cycle on/off & the drycoolers being sized to give sufficient heat rejection for the condensers only, resulting in the water temperature being too warm to use for free cooling.
The effectiveness of the freecooling is dependant initially upon the original (usally consultant's) specification & then (more likely unless we have set it up ourselves) poorly set parameters.
We do also use these same type of units in multi tennanted buildings where the client already has landlord supplied chilled water that they already use. The chilled water effectively becomes the free cooling & in this application, the DX circuit(s) would typically, where possible, have air cooled condensers.
Where correctly designed in the first instance, the free cooling concept does work & there are a number of blue chip companies who do save substantial amounts of energy in the winter months due to the free cooling function of our machines (some sites having upwards of 65 units installed with the client installing more units at later dates!).
Regarding the EC fan units, we do a special version of software that uses a combination of evaporator fan speed & water valve to take up the room loads. The fan usually runs at a lower speed & then as the cooling load increases, the fan ultimately speeds up. This is very cost effective for the client & all sites who have this software installed do like it.
For this reason, we are selling EC fan units by the bucketload at the moment.
Tiesse,
On this matter I think we have to agree that we disagree with each other...
The site in question were specified by Emerson sales guys, it was commissioned by Emerson engineers and several Emerson engineers attended site during various faults.
(Note, I/we were called in to help rectify some of the problems, we were not involved with the original install)
During one of the days I attended site, the ambient were ~0C and the system was not utilising free cooling.
That's my experience of it.
But of course you are right.
It might not have been correctly designed.
It might not have been correctly set up.
It might not have been correctly commissioned.
But, unfortunately, that is issues that I have no knowledge of as the customer bought those services from Emerson, the manufacturer.
No point trying to be clever with your post, I was merely trying to help you understand the logic of the Microface free cooling parameter settings.Quote:
Originally Posted by The Viking
Are you Morgan? If you are, we have met a while ago.
Regardless, PM me the details & I'll look into what happened on the site you are discussing. Trust me, give me the info & I will look into it & better than that, give you feedback.
Having said that, just because the ambient is 0degC, it doesn't then follow that the free cooling would be working (as I already posted in my explanation).
Yep Star still build thermosiphon systems for data centers.Quote:
Originally Posted by US Iceman
Its also worth considering the Trox AITCS systems pumped CO2 to evaporators mounted directly on the server rack doors. 20KW of cooling right where its needed - a rack of blade servers can produce 18KW or more of heat.
Thanks for that, I have looked into the star system it may limit our design because of the units positioning requirement. I will look into the direct blade server approach. So what's every-ones opinion on adding a chiller into a dry cooler circuit to deal with extreme external conditions?