Chiller dehumidifier 1

[ruggedscot]

We have a chiller packaged unit operating as a dehumidifier, ran into a problem and wanting some advice here on this one.

when the humidity level becomes satisfied the chiller stops, of course once stopped it shoots back up and then we have to wait for a set time for the motor protection softwear to count down before it permits a start, by this time we have the humidity shooting up.

Questions are ? How can this be avoided....

Can a soft starter be used to limit the motor stress when starting to allow more starts on the system ?

Can a different type of chiller be used that can take rapid starts with minimal rest times between starts ?

Hope that you can advise me on this one.....

Rugged

 

[Skogsgurra]

Rugged,

Sounds like a rather huge motor. A soft starter will not help you much since the rotor losses over time will be about the same with or without soft starter. (I am talking about a soft starter with reduced stator voltage). If you use a frequency inverter instead, you can avoid the stopping and starting altogether. Just let the motor run with the speed needed to keep RH just right. An RH controller should be added to make things fully automatic. Starting with a frequency inverter also avoids most of the rotor losses during start (nominal slip during ramping up).

 

[busbar]

Is the indoor-air fan run continuously, or cycled by the humidistat, or something else?

 

[ruggedscot]

Right its a hall that contains servers mostly the air is recirculated but we have fresh air going in through two fresh air handlers. It is this that we need to watch, the humidity levels should be kept around 50 % and we are having trouble maintaining this if the outside humidity rises due to the problem described before. the humidity goes out of spec to quickly for the chiller to cope with if the dehum function stops.

Rugged

 

[quark]

Generally the dehumidification system design depends upon a constant air dehumidification temperature called Apparatus Dew Point(ADP). Based on the inside moisture load, you should always cool your air to this ADP. At this point the air becomes totally saturated(100% RH) and as the air moves inside the duct work and in the room, it picks up sensible heat and the RH reduces to the required parameter. This is optimum design for a cooling dehumidification cycle.

However, as designers don't want to take chances with critical designs, air is cooled down to a generous ADP and hot water coils are used to reheat.

When you start and stop a chiller, inspite of the fastness with which you get the temperature, the inertia of the entire system makes your job difficult.

You can have two options.

1. Introduce a chilled fluid storage tank and maintain a constant chilled fluid temperature in it. Now you can start and stop chiller based on chilled fluid temperature.

2. Go for a desiccant dehumidifier for fresh air AHUs. Don't expect much savings with desiccant dehumidifiers. Your chiller load remains same. Moreover, you have to spend extra energy to regenerate the desiccant medium.

Regards,

 

[ruggedscot]

The system is mainly a recirc system using AHUs pulling air from the area being served and feeding back to the same area. We have fresh air supplies brought in through two AHU's and this feeds the other AHU's to provide a feed of fresh air. If the outside humidity rises and the system monitoring the humidity becomes satisfied then it shuts off. what happens then is that the humidity rises sharply and we have to wait a delay before the dehum comes back in this delay being about 5 mins allows the air supply humidity to rise right up.

Were talking large AHU's here and its to maintain air in an equipment hall. Im still working on some ideas and the water one looks particularly intersting indeed. Just now its a simple refrigerant circuit into the AHU that provides the dehimid function. if we were to go the route of a water chilled circuit then this might be the answer.....

 

[peebee]

Regarding the chiller start timer -- are they there just to limit the number of motor starts? I thought the refrigerant had to condense or something before you could restart a chiller, and that's what the timer's for?

Even my little window AC unit at home will wait a couple minutes before it will allow a restart -- I wouldn't think motor temperature would be an issue there. . . .

Can anyone confirm this?

 

[busbar]

Used to be that single-phase hermetic reciprocating compressors would not start until refrigerant pressures equalized, taking several minutes—stalling with early reattempts. That was never a problem with three-phase compressors.

Nowadays, the mechanical design of single-phase hermetic “scroll” compressors does not need equalizing off time.

 

[peebee]

When you talk "single-phase" and "three-phase" here -- I assume you're talking electrical terms (not mechanical -- like "liquid phase" or something)? Please clarify?

Any idea why my little window AC unit has the timer? Is it then in fact to limit the frequency of motor starts?

And here's a related question then which arises from all this -- in a place with critical cooling requirements (a data center) where the HVAC is fed from a transfer switch which might permit a power interruption of a few seconds -- are we unavoidably stuck with a couple minute delay in refrigeration despite the fact that the power was only out for a couple seconds? I've asked this one to a few mechanical engineers and chiller vendors before and have never received any very confident-sounding answers either way.

 

[ruggedscot]

Hmmm the criticality of a data centre.....

This is usually managed through dual stringing all items. Thus you have the cooling capacity managed over multiple units and then have it done in such a way that there is an inherent redundacy. This will cope with a fail of one string, but if you have a mains fail then this kit is supported through generators and you will have some dead time till the timers time out and the kit starts up again. The cooling is managed by having a good amount of chilled water in the system and then you can rely on that smoothing any bumps out. You just need to have clever operating system linked together and staff on site that know how to deal with a mains fail and on to generator power. Takes a little bit of learning ! remember you go through this again when you drop back onto utility - dont have the wait for generators to sync up and go to bus as the mains is there but you do want to have it under your control so that you dont flick back and forth if the utility network has issues. A call to your local friendly network controller can usually find out if you have any problems to worry about out there. then do your throw back over in your own time. Data Centres are a black art believe me.

 

[busbar]

 
peebee — Strictly electrical on the 1ø vs 3ø issue.

Data-center admins seem to be actual/self-appointed Nervous Nellies—sometimes with managers that have wide-open signature authority. On refrigeration-startup delay, conscientious HVAC manufacturers hawk a conservative stance on equipment reliability {or their perceptions of their clients’ perceptions of reliability needs.} A little randomized staggering of larger-motor starts is an easy solution to this. If it isn’t available already, some salesman will be proud to rollout their newest “fully addressable neural-control” mechnical equipment {soon with autonomous N+1 RS485/ethernet/fiber ports} as the latest advance in restart-control finetuning. Men’s room toilet-paper monitoring with automated refill can’t be far off.

For data centers, it may be that part of the “Black Art” is sales-droid vultures taking advantage of considerable middle-management paranoia.

Window air-conditioner restart timing could be to limit light-flicker complaints.

 

[quark]

Window and Split ACs generally have a time delay from 1 to 3 minutes based on their capacity. The time delay is between the indoor unit fan and the compressor. This is basically to reduce the chances of liquid entering the compressor and subsequent damage.

During offload conditions your AC unit will be stopped and stagnant air may not add the required heat to evaporate the liquid refrigerant in your evaporator. If you immediately start your compressor, there are fair chances for the liquid to go inside.

In other words, Busbar is right to the point. When the pressures are equal, you can ensure that it is gas everywhere.

With industrial chillers, the time delay is in between the start of chilled water pump and the compressor.

Regards,

 

[ruggedscot]

If it can go wrong then it will probably go wrong.

A data centre for a financial institution is a pretty important piece of kit, so it follows that you do what you can to eliminate any risk of any failure from occurring.

Not an easy task and one that means that the shirts and ties throw lots of money at to try and control. If you can remove the risk then you can limit the exposure to risk. Not an easy task.....

You should see the method statements and risk assesments that are required and the 'you do not do anything unless it is clearly written down' way of working that prevails. Get it wrong and well you lose a lot and that hurts.

Some people say that hospitals dont work to the same controls but you know that is one person on the table in a hospital. In a bank it is thousands of lives at stake peoples savings and such, so yes its fraught.

Rugged

 

[peebee]

"Some people say that hospitals dont work to the same controls but you know that is one person on the table in a hospital. In a bank it is thousands of lives at stake peoples savings and such, so yes its fraught."

Interesting perspective. After spending most of my career in the data center area (and previous to that in wafer fab and pharmaceutical manufacturing), I'm now starting to do some hospital work.

Not only do hospitals have lousy power installations compared to banks. The surgical areas are positively disgusting compared to a real pharmaceutical or fab clean room.

I appreciate your point about it being one person vs. thousands. But at the same time, jeez, I sure hope I never have to go under the knife.

 

[ruggedscot]

I appreciate your point about it being one person vs. thousands. But at the same time, jeez, I sure hope I never have to go under the knife.

But hey are we not better and more resiliant than mere machines ?

I served my time in a hospital and I know where you are coming from, can be scary seeing behind the scenes ! But you know hospitals get no where the same money as what a data centre gets - private versus public. This can be seen here in the UK when you look at private hospitals, very well done and look the part till you scrape the surface and look behind the scenes. With a data centre its reversed - especially dark sites - nothing to look at and all done to set spec, but the equipment wow nothing spared - best of the best and no risk taken with any kit what so ever - back ups for the back ups....

Clean rooms are worth seeing understanding principles and how they work, getting to grips with environmental control. Its look like we will be going back to water cooling, going full circle. This was abandoned when chips got smaller and more power effective but now chip densities are getting to the point that they need extra cooling. Put in a localised chiller in the unit and you have to dump that heat in the room adding to the heat load, get the point of dumping out of the room and you reduce the heat load presented so its looking like were going back to chilled water cooling at point.

It is an interesting game to be in and participate in.

Rugged

 

[KKmack]

Somewhere in this thread was a question about the reasons for delayed motor starts. Besides the need for refrigerant pressure equalization, as discussed above, there is also the need to limit a motor's exposure to full line voltage starts.
The larger the motor, the more restrictive, and the fewer allowed starts per hour.
A typical AC motor draws 3 to 6 times its rated full load current during startup, when the motor and its load accelerate from stop to rated speed. This extra current (and increased torque) during startup places tremendous thermal and mechanical stresses on the motor. To deal with the stress, most motors require minutes of cooling or equalizing time after each start; by running uninterrupted at rated speed, coasting or remaining stopped. Again, the larger the motor the worse the starting effect and the longer the wait between starts. Also, heavy loads placed on a motor during start up - due to unequalized refrigerant back pressure, for example - makes the situation even worse.

 

[peebee]

So -- if we got a Multilin-type motor relay, something with solid-state thermal modeling, could we possibly eliminate the delay on restart after a brief power interruption? If the motor's been running for a while, it would seem that it could handle another start after a single brief outage. Or, would unequalized refrigerant back pressure issues keep us from doing an instant restart after an outage? If a few minutes of down time are unavoidable, OK, we can live with that. But if an improved control package will possibly allow us to instantly restart if the motor's up for it, well, lets spend the extra bucks on the better controls. Maybe I should raise this question in the HVAC forum? Unfortunately, most of the ME's and chiller vendors I've spoken to about this issue don't even realize the restart time delay exists.

How about with VFD's? It would seem that with the soft starting afforded by VFD's that we could start the motor repeatedly. But I'm under the impression that the time delay exists even with VFD controls.

Someone mentioned that the time delay is there to let the pumps spin up. But if power has only been dropped for a few seconds, we can't have drained THAT much water from the system. Maybe a flow switch or pressure switch would permit us to get the chiller on-line faster?

And another question, just slightly off topic -- any thoughts in general on the value of putting a Multilin-type relay on a VFD? Is there any value to this or is it just a waste of money?

 

[mc5w]

When I start up my window air conditioner and when turning down the thermostat's setting the thermostat will sometimes have a contact bounce leading to the compressor stalling out. I have about 5 seconds to turn the thermostat back up because the motor overload relay is slower than the 20 amp time delay fuse at the fuse box.

Another possibility is to use an evaporator bypass valve that allows the condenser to cool the compressor while it is idling.

Is your dehumidifier the kind that uses some of the condenser heat to reheat the air like household dehumidifiers do? If so, does it have 2 condensers, 1 for reheat and the other for disposal of excess heat?

Mike Cole, mc5w@earthlink.net