Unusual minisplit application

njbatt

Ahoy Wallies! Strange one for you: I am the plant engineer for a facility that has some electrical equipment which dissipates a lot of heat in a small space that needs to be kept cool. The quality and creativity of the responses on this site often outstrip that of the HVAC contractors we have work on this stuff, which is generally not their fault for reasons I wont get into. I'd like to get some ideas for potential options for how to deal with this situation so that I can be more prescriptive with the contractors and hopefully actually solve the problem.

Here are the details. This is a room that is a fireproof steel vault about the size of a 20 foot shipping container (8x8x25 feet). It contains electrical equipment that is capable of rejecting upwards of 50 kW of heat (170k BTU) but closer to 5-10 kW most of the time (17-34kBTU). Currently there are two Daikin minisplit wall units, one mounted at each end of the room up at the ceiling, but they are inadequate. We have the thermostats set at 70 degrees but the returning air temperature is typically 90-100 degrees Fahrenheit.

The electrical equipment is arranged in two rows against the walls with a narrow aisle in the middle (like a raised floor server room, but minus the raised floor). Cool air is drawn in at the bottom and hot air blown out the top with high velocity blowers. There are small vents going to an adjacent equipment room that shut if the fire suppression system is activated and two others that go outside and have blowers that maintain a fairly strong negative pressure in the space.

There really isn't any space inside or outside the room to duct anything or add more equipment easily. These wall units are connected along with two other fan coil units (ducted types) in another room to an outdoor compressor unit that is rated at 12 tons according to its data sheet.

What (if anything) can I ask someone theoretically familiar with this Daikin equipment to do to remove more heat from this space without swapping out the wall units? With that 12 ton outdoor unit and moderate cooling load on the other two air handlers, what is the theoretical maximum amount of heat these wall units can extract at 95*F ambient? Thanks in advance for your creative thoughts!

pecmsg

How many BTU's can these heads handle?
12-Ton outdoor unit?

ChrisJ

Is there much latent load on the units? Humidity coming from anywhere or is the air usually very dry?

njbatt

That is a very good question. None of the facility documentation references those wall units - the ducted units in the other room yes (12k BTU each) but not these. I can't see any nameplate on them, but there also isn't really any clearance above or to the sides to look.

By their size I would guess the wall units are in the 18k range.

njbatt

ChrisJ: good question. The humidity is usually in the 20-30% range. The only way for moisture to get in is through the other area that is under negative pressure which has some poorly sealed conduit penetrations that go outdoors.

njbatt

Looking more closely at the humidity, would that measurement spiking indicate the units satisfying of the thermostat? This happens with the return air temp at around 83, so maybe the corner where the thermostat is mounted is actually seeing 70 when this happens. Is this part of my problem: the wall itself and location are keeping the thermostat from driving the system to cool as hard as it should be?

njbatt

That image came out tiny, I'm not sure how to make it bigger but the yellow trace is the humidity and the scale is on the right Y axis, the green trace is temperature and scaled on the left Y axis.

pecmsg

Can these be set for using the head thermistor as the reading?

Ironman

What size is the outdoor unit(s) that the ductless units are connected to? Is it 1 on 1 or 2 on 1?

Usually, the largest ductless wall unit that's offered is 3 - 3.5 tons, depending upon manufacturer. You're obviously well below the 170k btus that you stated is needed.

There's no way to make equipment produce more capacity than what the manufacture has designed it to do.

njbatt

All 4 units (the two wall units in the hot room and the two ducted units in a different room) are all connected to a single RXYQ72PBTJ, which has a nominal capacity of 12 tons.

pecmsg may be onto something getting the thermostat to use the head unit thermistors - that's where you'll see the heat, not so much down on the wall. These units basically need to run flat out all of the time so any interruption (as appears to be seen in the chart I posted) is something we can eliminate and improve the situation.

Regarding getting more capacity out of equipment than is designed, there is sometimes a way. I ran a facility years ago that had two 180 ton McQuay chillers each with two turbocore compressors on them. We had intense heat from electrical equipment there that was sometimes more than the chillers could handle and I found that by playing with the cooling tower set points to keep the condenser water hotter the chillers would get up to 110% of their nameplate capacity. I asked the McQuay tech about this when he was on site for something unrelated and he said that those compressors turn so fast that getting the refrigerant really hot and less dense the way I was would actually allow for a meaningful increase in compressor speed, effectively increasing the capacity of the unit. This required a delicate balance to maintain without tripping out on high pressure but it worked for a while until we got a third chiller.

pecmsg

njbatt said:

Regarding getting more capacity out of equipment than is designed, there is sometimes a way. I ran a facility years ago that had two 180 ton McQuay chillers each with two turbocore compressors on them. We had intense heat from electrical equipment there that was sometimes more than the chillers could handle and I found that by playing with the cooling tower set points to keep the condenser water hotter the chillers would get up to 110% of their nameplate capacity. I asked the McQuay tech about this when he was on site for something unrelated and he said that those compressors turn so fast that getting the refrigerant really hot and less dense the way I was would actually allow for a meaningful increase in compressor speed, effectively increasing the capacity of the unit. This required a delicate balance to maintain without tripping out on high pressure but it worked for a while until we got a third chiller.

All you succeed in raising the Condenser water temperature was to make the chiller use More electric for the same work!

Ironman

My statement about capacity was generic, not absolute. You can spray water on an air cooled condenser and cause an increase in capacity, but that will have unintended consequences because the equipment wasn't designed for it.

I thinks Paul's suggestion is good and may help, but 144k btus is still short of 170k and you may have less than that if the indoor units don't add up to that.

njbatt

pecmsg said:

njbatt said:

Regarding getting more capacity out of equipment than is designed, there is sometimes a way. I ran a facility years ago that had two 180 ton McQuay chillers each with two turbocore compressors on them. We had intense heat from electrical equipment there that was sometimes more than the chillers could handle and I found that by playing with the cooling tower set points to keep the condenser water hotter the chillers would get up to 110% of their nameplate capacity. I asked the McQuay tech about this when he was on site for something unrelated and he said that those compressors turn so fast that getting the refrigerant really hot and less dense the way I was would actually allow for a meaningful increase in compressor speed, effectively increasing the capacity of the unit. This required a delicate balance to maintain without tripping out on high pressure but it worked for a while until we got a third chiller.

All you succeed in raising the Condenser water temperature was to make the chiller use More electric for the same work!

Actually no. The data bore out that we were removing more heat from the plant in the high condenser temp scenario than with the temperature lower. Doing this literally saved statistically significant facility downtime as high temp alarms were no longer being tripped and we could empirically verify that we were extracting more heat.

ChrisJ

njbatt said:

ChrisJ: good question. The humidity is usually in the 20-30% range. The only way for moisture to get in is through the other area that is under negative pressure which has some poorly sealed conduit penetrations that go outdoors.

30% at 100F is a dew point of 60F. To me, that's a pretty significant latent load on the typical mini split. I would expect it to want to run at 40F or so and that's keeping it much higher.

I actually don't know how that will effect your decisions, but wanted the info out there for others. It may help them help you in the long run.

njbatt

Thanks ChrisJ, that's an interesting point. Other than sealing the leaky conduit protrusions to stop outside air infiltration, how else can I get the humidity down without adding yet more heat? A desiccant?

pecmsg: I don't mean to be contradictory but I am very data driven. I try hard not to draw conclusions without having data to back up or refute a hypothesis. That was the case with the chiller experience I referenced and is the case here. ChrisJ's comment makes sense and he wouldn't have been able to draw the conclusion without the data I collected and presented.

pecmsg

njbatt said:

pecmsg said:

njbatt said:

Regarding getting more capacity out of equipment than is designed, there is sometimes a way. I ran a facility years ago that had two 180 ton McQuay chillers each with two turbocore compressors on them. We had intense heat from electrical equipment there that was sometimes more than the chillers could handle and I found that by playing with the cooling tower set points to keep the condenser water hotter the chillers would get up to 110% of their nameplate capacity. I asked the McQuay tech about this when he was on site for something unrelated and he said that those compressors turn so fast that getting the refrigerant really hot and less dense the way I was would actually allow for a meaningful increase in compressor speed, effectively increasing the capacity of the unit. This required a delicate balance to maintain without tripping out on high pressure but it worked for a while until we got a third chiller.

All you succeed in raising the Condenser water temperature was to make the chiller use More electric for the same work!

Actually no. The data bore out that we were removing more heat from the plant in the high condenser temp scenario than with the temperature lower. Doing this literally saved statistically significant facility downtime as high temp alarms were no longer being tripped and we could empirically verify that we were extracting more heat.

I don't understand how a Centrifugal Chiller uses "Less Energy" by making it overcome a higher head pressure.

Look at the Performance Curve for any compressor...…...As the ambient goes up so does the amperage.

My Carrier 19XL series were at PEEK Efficiency at 86% of load, any higher and the KW per ton went up!

ChrisJ

pecmsg said:

njbatt said:

pecmsg said:

njbatt said:

Regarding getting more capacity out of equipment than is designed, there is sometimes a way. I ran a facility years ago that had two 180 ton McQuay chillers each with two turbocore compressors on them. We had intense heat from electrical equipment there that was sometimes more than the chillers could handle and I found that by playing with the cooling tower set points to keep the condenser water hotter the chillers would get up to 110% of their nameplate capacity. I asked the McQuay tech about this when he was on site for something unrelated and he said that those compressors turn so fast that getting the refrigerant really hot and less dense the way I was would actually allow for a meaningful increase in compressor speed, effectively increasing the capacity of the unit. This required a delicate balance to maintain without tripping out on high pressure but it worked for a while until we got a third chiller.

All you succeed in raising the Condenser water temperature was to make the chiller use More electric for the same work!

Actually no. The data bore out that we were removing more heat from the plant in the high condenser temp scenario than with the temperature lower. Doing this literally saved statistically significant facility downtime as high temp alarms were no longer being tripped and we could empirically verify that we were extracting more heat.

I don't understand how a Centrifugal Chiller uses "Less Energy" by making it overcome a higher head pressure.

Look at the Performance Curve for any compressor...…...As the ambient goes up so does the amperage.

My Carrier 19XL series were at PEEK Efficiency at 86% of load, any higher and the KW per ton went up!

I know nothing about centrifugal chillers, but my guess would be the metering device is too small and the higher head pressure allowed for greater refrigerant flow.

This is one of the reasons, if not the main reason air conditioners do poorly when the ambient is too low.

njbatt

pecmsg: aha, we are talking about different metrics. Your objective function is kW per ton and your cooling load is probably driven by the weather, so it's variable and rarely if ever exceeds your plant capacity, yes?

My McQuay situation was cooling semiconductors that threw off a more or less constant amount of heat year round with a chilled water plant that was undersized to begin with. I didn't care how many kW per ton it took, I just cared that the equipment didn't overheat. I discovered the hot condenser phenomenon by accident. One day after yet another shutdown due to overheating and running the plant partially in hand mode I noticed that the chillers could actually keep up with the load when the towers got abnormally warm. We did some experiments and looked closely at the data and sure enough, there was a narrow band where we could get > 100% out of the chillers without them tripping, just barely keeping things cool. We eventually got a third chiller and were able to run things in a more normal way.

EBEBRATT-Ed

@njbatt

Something is wrong with your idea of raising the head pressure to gain capacity, that is not possible as @pecmsg mentioned.

The other issue is that you have the wrong equipment for the job. Typical ac units designed for comfort cooking are designed for a typical split of 80% sensible cooling and 20% latent cooling.

What you are cooling more closely resembles a computer room which is mostly all sensible heat and very low latent heat removal.

What you are doing is process cooling not comfort cooling. It requires a system designed specifically for the load you have. Applying standard comfort cooling equipment to a process load seldom works.

Leibert equipment $$$$would work. Or you can install a split system designed with the proper evaporator coil and the right air flow for the job.

Job 1 is to calculate the load accurately. Is there any external heat load surrounding the container or is it all internal load??

Harvey Ramer

njbatt said:

pecmsg: aha, we are talking about different metrics. Your objective function is kW per ton and your cooling load is probably driven by the weather, so it's variable and rarely if ever exceeds your plant capacity, yes?

My McQuay situation was cooling semiconductors that threw off a more or less constant amount of heat year round with a chilled water plant that was undersized to begin with. I didn't care how many kW per ton it took, I just cared that the equipment didn't overheat. I discovered the hot condenser phenomenon by accident. One day after yet another shutdown due to overheating and running the plant partially in hand mode I noticed that the chillers could actually keep up with the load when the towers got abnormally warm. We did some experiments and looked closely at the data and sure enough, there was a narrow band where we could get > 100% out of the chillers without them tripping, just barely keeping things cool. We eventually got a third chiller and were able to run things in a more normal way.

Raising the water temperature in the cooling tower increased the evaporation rate and cooling capacity.

Harvey Ramer

If you have 100° return air temp, your evap temp is probably somewhere in the 60° - 65° range. At those low humidity levels you are probably only doing sensible cooling. You have to look at the Daikin engineering data, but the evaps' sensible capacity is probably somewhere around 70% of total capacity.

njbatt

Thanks Harvey, I did indeed find that data on the Daikin literature and confirmed that the head units are 24k BTU (18k sensible). Today we noticed that one of the units in the other (cooler) room was set to auto, which if I am reading the manual correctly means that whenever it goes into heat mode it will interrupt the cooling of the other units (they're all connected to the same compressor unit outside). Is that right? We set them all to cool at the lowest set point and temporarily stopped the negative pressure ventilation to stop any outside air from being pulled in but couldn't get the temperature down any more, nor did the humidity come down appreciably.

EBEBRATT-Ed

@njbatt

Humidity shouldn't be an issue what type of climate is this container located in? Is it sealed tight? If you are cooling this load and it is not occupied there is no need for outside air.

nicholas bonham-carter

What about some outside air, such as an exhaust fan blowing all the heat outside. Naturally some outside air needs to be coming in to replace it. Is the local climate able to provide some air which is cooler than the ambient in the container?—NBC

hot_rod

What type of equipment is in the room? Maybe you pull heat out of the equipment like the do with water cooled servers farms

njbatt

The enclosure is in the Philadelphia area, so it gets reasonably cold outside in the winter but pretty hot and humid in the summer. I considered free cooling but that would require cutting into the enclosure to allow outside air in, which would have considerations for fire suppression (the need for more automatic dampers, sensors to make sure we're not pulling air that's too humid, etc).

hot_rod, The container is full of air cooled racks of high power lithium cobalt manganese batteries. They can pull or push up to a megawatt at a time, with about a 5% loss (hence the 50kW worst case heating load figure).

Thanks for all of the responses, there isn't budget to re-do the HVAC at this time so a creative solution will be required.

ratio

It's possible to under- or over-commit the outdoor unit, capacity wise, to some extent. Double check the capacity of all the units, if you're overcommitted you can free up some capacity by turning off one of the other heads.

ChrisJ

ratio said:

It's possible to under- or over-commit the outdoor unit, capacity wise, to some extent. Double check the capacity of all the units, if you're overcommitted you can free up some capacity by turning off one of the other heads.

I believe he wants to put bigger outdoor units on the same heads due to limited space inside.

njbatt

ChrisJ said:

ratio said:

It's possible to under- or over-commit the outdoor unit, capacity wise, to some extent. Double check the capacity of all the units, if you're overcommitted you can free up some capacity by turning off one of the other heads.

I believe he wants to put bigger outdoor units on the same heads due to limited space inside.

I have enough (almost) outside: all four indoor units are tied to a single 144k BTU compressor/condenser unit (a single RXYQ144PBTJ). The best solution is to find a way to get more evaporator coils onto the system, but the available space and budget will prevent that at least for now on the budget side.

ratio

You're limited by the sizing of both the indoor & outdoor unit: over/under commit is just sizing the outdoor for sightly less/more than the total of the indoor units. It's common around here to overcommit, the thought being that everything isn't going to be calling for 100% at the same time, but with process cooling I'd rather see undercommit so we're guaranteed capacity.

You can play games with the heads a little I think. Try setting the fan to 'on hi'. With the airflow higher than optimal for latent cooling we should do somewhat more sensible cooling.

hot_rod

Battery coolers like this. i always look for a hydronic solution:)

https://avidtp.com/what-is-the-best-cooling-system-for-electric-vehicle-battery-packs/

Harvey Ramer

You can also set one of the units in the server room to priority. That will enhance the cooling capacity for that unit by prioritizing it over the other heads.