Spring is coming. Will your data center be a spring chicken or an old bird? As the weather begins to warm (hopefully) this is a great opportunity to take stock of your data center and it’s economizer system and operation.
We have this sensor, we call it the million dollar sensor. Now how can a single sensor be worth a million dollars? The answer is simple, really, it’s the outside air temperature and humidity sensor, and we have selected it such that many of the problems with other sensors, particularly humidity, are eliminated. By having a highly accurate sensor, we can reduce the margins for the economizer to engage ( for instance, return air temperature plus 2 degrees instead of 4) and have knowledge that we won’t be what I call “anti-economizing” where the air we bring in has more enthalpy than the air exiting the building.
By having accurate humidity measurement, we also have accurate enthalpy measurements. Depending on the region of the country, enthalpy economizer control can add a few hundred hours, but in many regions a comparative dry bulb will work very well with a simpler control scheme and less component failures.
So make sure your outside air sensors, return air sensors, and check that economizer sequence of operation.
But let’s talk about that economizer logic. Now, hopefully everyone is using the ASHRAE temperature bands, from what I have seen the recommended band has the most universal acceptance, and with those bands you have set the economizer aggressively. The important thing about the recommended band is that it is an operating band and not a limit. You should use every corner of the band to maximize energy savings: let humidity ride up to the upper limit rh and dew point and allow temperature at the server inlet to approach 81 degrees. The energy savings which can be realized are very substantial. If you can satisfy your environment with 70 degree air in the supply plenum, then any time the outside air is less than 70 degrees you can be fully economizing. In Chicago, that is 7000 hours a year with no mechanical cooling. And if you have your server inlet at 75-77 degrees, and your air volume is set up for a nice 25 degree delta across the servers, than you should have a partial economizer available whenever it’s less than 98 degrees out which in Chicago is about 2 hours a year (in reality it’s less but that’s for the 95-99 degree temperature bin). The only real lockouts for partial economizer should be high humidity in the data center environment. Using the warmer air at 75-77 at server inlet or more, you won’t be worried about relative humidity, you will be approaching the max dew point ASHRAE rating first. So you just need to protect your raised floor from getting to humid. One method is monitoring the dew point at the server inlet and as the upper limit of the ASHRAE band is met, bet reduce the maximum economizer percent, and if dew point goes outside of the and we can disable the economizer. There’s lots of ways to do it, but really pushing those savings requires aggressive use of the ASHRAE bands.
Be aware of the word of caution on Enthalpy economizers. Enthalpy Economizer Issues
You can push things further, especially in dry climates, with direct evaporative cooling which cools the air without mechanical cooling. But a lot of climates can get a great portion of the economizer hours with the aggressive use above.
And you may even be able to re-evaluate your bus loading, alluded to in previous posts, if you never need to engage the full capacity of the cooling system (as well as the fact that the cooling will work more efficiently when the return air is warmer).
A word of caution: R-22 is not a high temperature and pressure refrigerant. That is to say, when that return air starts passing 85 degrees, R-22 based machines can start to experience reliability issues as they stop cooling the compressor properly and lots of other effects can happen. So you may want to reach out to your vendors to discuss. Conversely, by using the expanded band, you may so reduce compressor hours though that this impact is overshadowed by low run hours and your equipment actually ends up lasting longer. But, it is a consideration to keep in mind.
Anyway, let’s all be spring chickens! Don’t leave these gobs of money on the table.
In the hot and humid tropical climate, the hours of economization that can be achieved is very limited, making free cooling economically unfeasible. There are products on the market to improve efficiency of DX cooling units by using water mist to adiabatically lower the ambient temperature surrounding the outdoor condenser units. To maximize the adiabatic effect, some products make use of high pressure system (up to 50bars) together with nozzles to create ultra-fine droplets. However this increases the cost of such system as high quality filtration and stainless steel components are required to avoid clogging of the nozzles. I wonder if this had been used in any site in the US. It would be interesting to pilot this in one of the site (ideally in tropical region with the most challenging climate) to see how this performs.
LikeLiked by 1 person
Mak, we are installing some evaporative assist products on chillers in our New Jersey market as well as our Dallas market. Also, we are looking at it for RTUs for new units when we replace old units at End of Life.
The benefit is great in hot humid environments where mechanical cooling is used frequently, and also great in dry environments, especially where we have air cooled chillers with no economization. The drier the air, the more cooling can be realized from evaporation. We also currently employ direct evaporative cooling, where the economizer brings in warm air that might not be cool enough for the data center, and evaporate moisture directly into that air stream to cool it. As long as we maintain the upper limits for humidity of the ASHRAE band, this can generate up to 15 degrees of cooling in most of our infrastructures (complete saturation of air could be in excess of 25F, but we can’t completely saturate server inlet air).
This is accomplished via misting systems (MeeFog, Humifog by Carel) and by wetted fill media (Munters Glasspak).
We have evaluated Munters for air cooled chillers, but the air velocity is too high for that process. MeeFog or Humifog may be good options if ample absorption distance is available, and also a product called “Evaporcool”, which uses a wetted media (almost like a greeny pad material) instead of direct misting, which significantly mitigates coil re-wetting.
LikeLiked by 1 person
Evaporative cooling product like Evaporcool and Munters is not that popular here yet although its benefit for our climate is obvious. Maybe people do not fully understand the benefit and often stopped by the capex involved. I have seen some self-retrofit evaporative system using water pump delivering water through holes drilled on PVC pipes, basically just spraying a lot of water onto the condenser coil. I wonder how efficient that could work, all the water wasted could easily outweigh the small efficiency gain by the condenser, plus the drainage needed to drain off that amount of water from the roof and corrosion problems with moisture collected on the coil.
LikeLiked by 1 person
The comment about R22 systems is absolutely correct.
R410A Systems, used in new equipment since about 2010, can even be MORE sensitive to higher entering coil temperatures than R22. End users may want to verify the operational range with the manufacturer before proceeding with higher entering refrigerant coil temperatures.
LikeLiked by 1 person