Minimum discharge pressure for ammonia systems

[US Iceman]

To all ammonia guys, this ones for you...

I am interested in finding out how low you can operate your discharge pressures in the winter. What is the minimum pressure you can run?

Any problems with liquid supply to the system?

How many of you are using liquid injection for oil cooling for screw compressors?

Or even, what type of oil cooling are you using? What have been the problems you have had with your oil cooling?

How about hot gas defrost problems?

Has anyone had experience running at 10C condensing with ammonia systems?

[I LUV NH3]

I don't like to see the head pressure drop below 110 psig.

You loose the ability to hot gas defrost. Especially on evaporators that are at the farthest from the hot gas source. Typically defrost back pressure regulators are set to 70-75 psig. You can have up to 20 psig in pressure drop through the piping and valves not to mention the evaporator. This can add up to no defrosts.
Liquid injection? I design my systems with thermosyphon oil cooling. Most compressor manufactures agree that you can loose up to 11% of the compressor's capacity by cooling via liquid injection. Not to mention the reduced life of the bearings.

[NH3LVR]

I am a bit old fashioned and like to see the Head up a bit.
But that is not the way we do things anymore.
The head required to operate the Liquid Injection varies from Compressor to Compressor. Even in the same engine room between similar machines. Older FES and Fricks seem to have trouble lower than lbs (9BAR) In some cases even that is not enough.
One customer found he could keep his head down to lbs (11.8BAR) in the Winter. Of course his coils would not defrost and his TX valves could not keep up.
Another customer had a abundance of Condenser. He kept the head as low as it he possibly could.
I changed liquid line solenoids on a regular basis for him. He finally overfed a Intercooler and defeated the high level safety one night, to the tune of $10,000. Vilters do not pump liquid well.
Low head pressures are fine if the system is designed with that in mind.
We are quoting a Liquid Pump for one plant now, to be installed in the Liquid Injection line for the screws.
I have seen the Head as low as 96lbs (12.2 BAR) during a defrost this fall. This occasionaly shuts down on of his high stage machines.

[aawood1]

Hi, This year under orders from the people who pay the running bills we had to lower the Fan starts on the 4 water/ air condensers we us on one plant. The new start points are now set at 1st. condenser 8.5 Bar from 9.5 Bar, 2nd. condenser 9 Bar from 10 Bar, 3ed. condenser 9.5 Bar from 10.5 Bar. and the last one to 10 Bar from 11 Bar. The old settings have been used for the last 24 years with no problems, and we had a hot gas defrost system fitted to a freezer 8 years ago.

Will have to get over this Winter to find out what problems we will get from the new settings. The only one at this time is a bad hot gas defrost when we have a low load on the plant.
Arthur

[US Iceman]

I originally posted this over a year ago.

There are some good ideas and concerns being listed so far. What I think this comes down to is, the system has to be designed to operate at low discharge pressures.

Liquid feed systems and hot gas defrost are the two big items. TXV's tend to crap out below about 125 psig (8.6 bar g), hand expansion valves on liquid make-up lines have to be opened further.

The hot gas systems require some thought. As mentioned earlier, it is common to have a high pressure loss in the defrost header. Why not make the pipe larger?

Use a drainer to keep condensate out of the header, so that it heats up faster. Besides, if you drain the condensate you have less opportunity to have liquid hammer during the start of the defrost cycle.

These can be done very easily on new systems.

On old systems, use a dedicated recip. compressor to supply the hot gas. If you operate a single compressor to supply the hot gas, you can operate it with higher discharge pressure, while the other compressors are operating at a much lower discharge pressure.

I've seen some owners install an abundance of condenser capacity to run 150 psig (10.3 bar g) in the summer.

I have also worked on some jobs where we were running at 80 psig (5.5 bar g) in the winter.

Another item to watch carefully is the minimum discharge pressure before it gets too low. Oil separators tend to get fickle around 140 psig (9.6 bar g).

[NH3LVR]

USIceman
A lot of good thought in this thread.
We size our HG Lines adequately, but there is always the problem of pressure difference between the Head and Evap. We sometimes go to 90lbs (6.2BAR) on defrost regulates. 100lbs (6.9BAR is not a lot of difference.
We almost always install drainers on the lines except when the runs are very short. It helps a lot.
Are you suggesting we run two head pressures/Condensers on older systems? I think I misunderstood.
We have a new Screw Machine waiting to be installed.
It is a Variable Speed and I understand it has a new type of valve (Not TX) for the Liquid Injection. It is supposed to overcome the Head Pressure Requirement for Liquid Injection. It is waiting for some Seismic Issues to be resolved before installation.

I had not considered a Oil separator problem before. I would have assumed they would work better under the lower Head/Temp conditions.

[US Iceman]

Hi NH3LVR,

It seems there are getting to be more of us ammonia guys here. That's great!

I think part of the problem of sizing hot gas defrost lines is; how do we estimate the total mass flow of refrigerant needed for defrosts? I've seen some systems where a 2" pipe is used, others install a 4" pipe. I would like to know how they sized this.

Small pipes make the problem more critical. Larger pipes help a lot.

Are you suggesting we run two head pressures/Condensers on older systems?

Not quite... What I'm suggesting is to have a spare/other recip. compressor piped into the hot gas header (instead of the main discharge line) to supply the evaporator coils with hot gas during defrost.

The other compressors could operate at the lower discharge pressure, while the dedicated recip. compressor (for hot gas) could operate at 150 psig (10.3 bar g) just to supply the hot gas for defrost.

The discharge temperatures are a lot higher than on screws, so the extra temperature helps to keep the hot gas header warm. If we use a drainer to keep the liquid out of the header, then it starts warm. You also don't have to wait for the header push all of the liquid out of it during defrost. This liquid normally flows into the coils at the start of defrost, which does no good either.

...I understand it has a new type of valve (Not TX) for the Liquid Injection. It is supposed to overcome the Head Pressure Requirement for Liquid Injection.

Is it a Jordan valve?

Low discharge pressures help to create higher gas velocities in the separator. Due to this, the carryover rate can increase, which is not good.

[Andy]

Hi

I set up all my plants for a 10.5bar head and when I am designing I use a mulitplier of 1.5 time coil duty for each hot gas line

Also it is better to fit a pump out system when not running the hot gas line, saves having to push large quantities of condensed liquid through valves not designed for it slowing up the defrost.

Liquid injection is a poor way of cooling, better with an oil cooler Oil coolers reject the heat out of the refrigeration circuit increasing COP, liquid injection adds to the compressor load. Oil coolers work better with low ambient, injection doesnot. Plants designed for economy use low heads and oil cooling

Kind Regards Andy

[US Iceman]

...I use a mulitplier of 1.5 time coil duty for each hot gas line.

I use a similar logic, except I tend to use a larger multiplier.

The way I look at this follows...

When the coil is ready for defrost there is a very large temperature difference between the cold coil and the saturation temperature of the hot gas defrost supply.

When defrost starts, there is a large quantity of condensate being formed. This means you need to be able to supply a large quantity (high mass flow) to the coil to defrost it quickly. To do this with low discharge pressures means you need a larger pipe ID, for the high mass flow with a low pressure drop.

The mass flow is very high when defrost starts, and decreases as the temperature difference (between coil & sat. temperature of gas supply) decreases.

Basically, I use a multiplier of 3 to account for this spread in mass flow requirements.

This results in a larger pipe ID, lower velocity in the hot gas header, and lower pressure drop. It works pretty nice.

Also it is better to fit a pump out system when not running the hot gas line, saves having to push large quantities of condensed liquid through valves not designed for it slowing up the defrost.

I prefer to use a high-side float valve to just drain the condensate. This keeps the liquid out, and also helps to keep the pipe warm.

In principle we are both doing the same thing. The condensate in the headers is bad!

[Andy]

Hi Iceman

what about the main hot gas line, by that I mean before you branch off to the coolers/valve stations, how do you determine the size for this

I use a 25% rule, by that I mean 25% of the coolers will be on defrost at anytime.

Kind Regards Andy

[US Iceman]

Hi Andy,

The main discharge line (to the condensers) would be sized for all compressors in normal operation as if the hot gas is not being used.

The hot gas defrost header would be sized for 3X the total coil capacity that could be in defrost at any one time. I think your 25% value is pretty close.

So, in effect I would use 25% of the total coil capacity (that this defrost header serves) and multiply that by 3. This makes the hot gas defrost header large enough to have a high capacity at low pressure loss.

You would also have to size the defrost header for a low pressure loss for whatever the total equivalent length is required for the installation.

The individual branch lines to each coil would be sized for 3X the individual coil capacity then also.

I would put a high-side float valve at the end of the header to drain condensate. That way, the header is dry and you would not have to purge liquid through the evaporator coils during defrost.

[Andy]

Hi Iceman

much as I thought you would say either I an under sizing the branch lines or I am selecting the hot gas lines on lower discharge pressures

Thank you

Kind Regards Andy

[US Iceman]

Hi Andy,

My normal preference is to slightly oversize the piping anyway. However, on hot gas defrost headers with my method it seems I usually end up with twice the diameter of what everyone else uses.

The smaller pipes work OK as long as you keep the discharge pressure up. The smaller pipes also let the pipes jump around somewhat when the hot gas solenoids open!

If you sit down and look through some pipe selections for the higher mass flows and low pressure drops you will find the following:

Add the defrost pressure, plus the pipe pressure loss, plus the total valve station pressure loss for the hot gas and you end up with the minimum approximate hot gas supply pressure. If the pipe pressure loss is high, the minimum discharge pressure will be higher.

This is something I developed on my own, since I could not get any reasonable answer that was based on some fact or a little science.

[Andy]

Hi Iceman

The 1.5 comes from so selections Danfoss did years ago for me

The last company I worked for used 3-4 times

I have to say there design on most things was quite good

Some food for thought, we are going to start up a R404a system on hot gas defrost, it should be OK as it is DX (electronic) the suction is the hot gas line line to each cooler, I have already up the mains hot gas line size by twice what I had originally decided to use

I will let you know how it goes

Kind Regards Andy

[US Iceman]

The last company I worked for used 3-4 times

At least we have a consensus! I think the 3-4X number is a good average to size the branch and main headers. I also believe the value is considerably higher when defrost initializes, but drops down to a very low number when defrost is almost complete.

If anyone has a paper written on this I would love to see it.

The next trick is trying to get the defrost cycle to terminate to keep from blowing hot gas down the suction line.

[Josip]

Hi, all

Very nice I have to agree about almost all said, but isolation of hot gas piping is not mentioned

In past times we used one vessel named defrosting vessel to return liquid and gas after defrosting - no problem with return of high pressure gas into suction separator, but now day nobody use it.

Best regards, Josip

[US Iceman]

Hi Josip,

In past times we used one vessel named defrosting vessel to return liquid and gas after defrosting...

Would this vessel be operating at a high suction pressure close to the defrost pressure? Were you using a separate compressor operating at this high suction pressure?

If so, that would be a very efficient way to handle the defrost loads, but it would definitely cost more to install.

That might explain why it is not being used more.

[Josip]

Hi, US Iceman

Hi Josip,
Would this vessel be operating at a high suction pressure close to the defrost pressure? Were you using a separate compressor operating at this high suction pressure?

If so, that would be a very efficient way to handle the defrost loads, but it would definitely cost more to install.

That might explain why it is not being used more.

Defrosting vessel was one high pressure vessel above receiver (and little smaller then receiver) to drain all liquid after defrosting down to receiver. First we close the valve to low pressure separator and then equalize pressures by the valve on gas side between receiver and defrosting vessel, then we drain the liquid through the valve at the bottom of defrosting vessel.

Of course after drain it was necessary to close both valves on gas and liquid pipes between receiver and defrosting vessel. Inside of defrosting vessel remain only high pressure gas what was easy to remove and maintain low pressure by opening one small valve to low pressure separator -10C(14F) which remain open all the time until we refill the vessel.

Defrosting vessel was equipped with sight glass and high level switch to give you signal for drain liquid into receiver. Then we use that sub cooled liquid to charge separators again.

This system is good when you have a lot of low temp rooms, doors are opened often and you need to defrost them regular.

Yes, that was in past time when you have a lot of operators to run complete plant manually but now days all owners like to run their plants without operators because they are too expensive

Best regards, Josip

[US Iceman]

Hi Josip,

OK. I see what you are describing. I've seen something like this before. It's a gravity drain transfer vessel.

I had not thought of using this for defrost condensate.

Where I've seen these, the vessel at a higher elevation is used only for an accumulator. When the liquid level rises to a certain height, the level switches open and close solenoid valves for equalizing and venting so the liquid runs into the receiver.

Yes, that was in past time when you have a lot of operators to run complete plant manually but now days all owners like to run their plants without operators because they are too expensive

Yep, you are correct. I had a discussion with someone the other night about this. After everyone gets dependent on control systems to operate their plant, they will forget how to run it manually.

Thanks for the description on the defrost vessel.

[TXiceman]

Some of the newer liquid injection valves are electronic Danfoss valves. Much wider range of control than a TXV.

Ken

[Mark Sanchez]

Gentlemen,

Great information!

Typically what I've found is that the system should be designed for the lower head pressure.
ie:- hot gas line sizing for lower pressure differentials.
- Drainers should definitely be incorporated.
-Oil seperator sized for lower head/ if the seperators are not oversized for lower disch press. the coalescers will blow out due to the higher density of the lower pressure gas.

Another thing to look for is that the customer may be using more electricity to run the lower head , than the savings being recognized by the compressors. Check the HP of the fans and pumps being utilized vs the amp draw of the compressors at the lower heads vs say 125# disch. Longer defrost times will also complicate this as the hot gas relief has to be recompressed. Ineficiencies are worse on systems that relieve hg def to the low stage/temp vs relieving to the hig temp.


Hope this adds some food for thought

Mark

[US Iceman]

Some of the newer liquid injection valves are electronic Danfoss valves. Much wider range of control than a TXV.

I have not tried any of those, although an opportunity to may arise shortly.

Ken, Have you used any of the Jordan temperature control valves? These are the same ones Sullair used for liquid injection on their screws. These seem to work pretty well and are simple.

[TXiceman]

I have used a few Jordan valves. They do Ok in a system that pretty stable and the compressor unloads and loads slowly. The valves are pretty slow acting and can over and under feed with a rapid load swing for the compressor. In that case, the oil temp may swing out of range.

Ken

[nh3wizard]

Frick installed the Danfoss valves on the liquid injection on thier new packages, as long as you get the Quantum paraimeters set right the valve works great.

[NH3LVR]

Just looked at a new FES Variable Speed Screw a couple of days ago. They are also using the Danfoss Valve in place of the TX on the Liquid Injection.

[Andy]

Hi

the AKVA valve as Danfoss calls it is the only really good way of close control of superheat on DX ammonia that I know of

Kind Regards Andy

[NH3LVR]

Andy;
Sorry, I may have left the wrong impression. The Danfoss valve that FES is using is a motorized Valve.
I was in the plant to take some measurements for the installation and unwrapped the valve.
I noted the Model Number down carefully. Only when I got home did I realize I only had the Actuator Number.
Will punish myself with some Real Ale.

[Andy]

Hi NH3LVR

Thats the new ICM motorized vlave module for the ICV range that came out to replace the PM valves. V seats on the ICV make for real control. I have seen these same valves fitted as Co2 expansion valves on a chiller at the IKK show. Should be a good valve

Kind Regards Andy

[NH3LVR]

Thank you for the Info, Andy.
Although Danfoss has been selling TX Valves here for years, the Industrial use is only now becoming common. RS is our most used supplier.

[Andy]

Thank you for the Info, Andy.
Although Danfoss has been selling TX Valves here for years, the Industrial use is only now becoming common. RS is our most used supplier.

No Problem

Danfoss have also a new valve system for NH3 control, a bit like a hydralic block, with all the controls on one main block. This would be good for packaged NH3 equipment, I'm thinking scraped surface systems and possibly chillers
http://www.danfoss.com/NewsAndEvents...rigeration.htm

Kind Regards Andy

[Bradford]

Another thing to look for is that the customer may be using more electricity to run the lower head , than the savings being recognized by the compressors. Check the HP of the fans and pumps being utilized vs the amp draw of the compressors at the lower heads vs say 125# disch.

Good point. We applied floating head pressure to one plant that runs about 5,000 hp during the week for process and 500 hp on the weekends for cold storage. Because the head pressure setpoint was fixed, every condenser device could run trying to achieve setpoint even though only 10% of the compressors were on. During our tests, our control system caused the compressors to consume about 8 kva more but the condensers went down by about 300 kva! All by raising the head pressure setpoint slightly.

Brad

[Sergei]

Hi, everybody. I've just signed in.
Bradford, you need PLC with wet wulb approach. This feature will help you to run this plant efficiently at different ouside conditions and at different refrigeration loads.

Regards

Sergei

[US Iceman]

Hi Sergei,

I noticed this was your first post and thought I would say welcome. Please feel free to participate in the on-going discussions.

As a matter of curiosity, how are you measuring wet bulb temperature for the approach?

[Sergei]

Hi Sergei,

I noticed this was your first post and thought I would say welcome. Please feel free to participate in the on-going discussions.

As a matter of curiosity, how are you measuring wet bulb temperature for the approach?

Hi, US Iceman.
I know 2 approaches to measure wet bulb temperature.
1. Measure dry bulb temperature and relative humidity. Calculate the wet bulb temperature.
2. Use wet bulb sensor for measure this temperature.

Sergei

[US Iceman]

I know 2 approaches to measure wet bulb temperature.
1. Measure dry bulb temperature and relative humidity. Calculate the wet bulb temperature.
2. Use wet bulb sensor for measure this temperature.

I have heard the RH sensors are cheaper than wet bulb sensors. Which one are you using? Have you noticed any problems with the type you use?

Thanks.
US Iceman

[Sergei]

I have heard the RH sensors are cheaper than wet bulb sensors. Which one are you using? Have you noticed any problems with the type you use?

Thanks.
US Iceman

I don't have information about the prices. As far as I know Frick is using wet bulb sensor, but Hench is using RH sensor. These sensors should be calibrated regularly, bacause wet bulb temperature is base temperature for tunig up of optimum head pressure.

Sergei

[US Iceman]

OK, thanks for the information Sergei.

[Bradford]

Bradford, you need PLC with wet wulb approach. This feature will help you to run this plant efficiently at different ouside conditions and at different refrigeration loads.

Hi all, been out for a while, good to be back.

For measuring wetbulb, we use a combination temperature / humidity unit. Wetbulb temperature is calculated using the Raphson-Newton method for best accuracy. The current units that we are having built have been in operation for about 5 years with no failures and no need for calibration.

[Sergei]

Good point. We applied floating head pressure to one plant that runs about 5,000 hp during the week for process and 500 hp on the weekends for cold storage. Because the head pressure setpoint was fixed, every condenser device could run trying to achieve setpoint even though only 10% of the compressors were on. During our tests, our control system caused the compressors to consume about 8 kva more but the condensers went down by about 300 kva! All by raising the head pressure setpoint slightly.

Brad

Hi, Brad.
How you control your head pressure? By wet bulb approach or by fixed head pressure set point.

Sergei

[Josip]

Hi all

Check this:

http://www.sabroe.com/fileadmin/file..._R717_0410.pdf
3+ MB

it is Condenser Pressure Optimizer working with ambient temp sensor and RH sensor.....

Best regards, Josip

[nh3wizard]

Does anyone have any experience working with these optimizer's? If so how well do they work?

[US Iceman]

Hi nh3wizard,

I have seen this before on the web, but have no experience with them. I suspect the guys from across the pond will have something to say.

You might be able to get them from M&M Refrigeration, since they are the Sabroe rep here in the US.

If you try one of these let me know about your experiences.

[Bradford]

Sergei

Wetbulb approach. Also, each evaporator has it's own pressure reset for defrost

[Bradford]

Hi US Iceman,

We include this type of floating headpressure in all of our refrigeration control applications. Current installations are accross Canada and in the US. Europe and GB are next.

As I indicated elsewhere in theis thread, I have seen energy reductions of over 300 KVA.

Regards, Brad

[Sergei]

I've looked at this optimiser. Sorry, I don't believe in magic black boxes or in aluminium boxes even from Sabroe. Operating engineers should have information about settings for their refrigeration plant as well as access to change these settings. Sometimes the right settings for one plant are not working for another one. It looks like this optimiser for refrigeration plant with one condenser and condenser fan VFD. Majority of industrial plants have 2 or more condensers. What is the minimum allowable condensing pressure? Every refrigeration plant has it own minimum allowable condensing pressure.

Sergei

[Sergei]

Hi US Iceman,

We include this type of floating headpressure in all of our refrigeration control applications. Current installations are accross Canada and in the US. Europe and GB are next.

As I indicated elsewhere in theirs thread, I have seen energy reductions of over 300 KVA.

Regards, Brad

Hi, Bradford.
Sophisticated PLC is your tools to save energy. You should have the right settings and the right operating strategies to maximize your energy savings.

Regards, Sergei

[US Iceman]

What is the minimum allowable condensing pressure? Every refrigeration plant has it own minimum allowable condensing pressure.

Now we getting to the main problem as I see it.

Each plant is different and does have it's own minimum dicharge pressure. The real problem is finding what particular part of the refrigeration system determines this.

Now, I do not know much about the Sabroe Optimizer if it is for one condenser or many. But I do know that big systems with many condensers can be a challenge to setup properly.

Operating engineers should have information about settings for their refrigeration plant as well as access to change these settings.

Absolutely right. The operators need to be able to change settings as required and also need to be aware of why the setpoints are important.

I am not too fond of control systems that use "fixed setpoints" that are not adjsutable by anyone other than the programmer.

I have seen some control systems that can reduce the total energy by over 40% or more. So I know what is possible. I also believe the operators need to understand what the control system does and what it is capable of, so that the setpoints do not get changed to an old value used before.

Good control systems do not replace good operators.

[Josip]

Hi, all

I've looked at this optimiser. Sorry, I don't believe in magic black boxes or in aluminium boxes even from Sabroe. Operating engineers should have information about settings for their refrigeration plant as well as access to change these settings. Sometimes the right settings for one plant are not working for another one.

Of course operating engineer has info and access to those settings. Each plant has own settings even if they are installed in the same factory due to different design or duty.

It looks like this optimiser for refrigeration plant with one condenser and condenser fan VFD. Majority of industrial plants have 2 or more condensers.

Agree!

But still we control only one i.e. last one started All equipment started earlier we push up to maximum capacity. Trying to control all devices, compressors, condensers, pumps, fans, etc simultaneously we will drive complete system into "hunting" situation and finally to alarm stop.

What is the minimum allowable condensing pressure? Every refrigeration plant has it own minimum allowable condensing pressure.

Having good plant with enough capacity, with good control of each device we can go very low with condensing pressure but then we have to heat something i.e. defrosting cycle, what now?

Switch on manual and come back to 30/35C (speaking about NH3).

It seems to this question is not easy to give simple answer

Coming back to "optimizer" its name is telling us we can try to get what we pay. (is that good expression?)

I am assuming also all PCs, PLCs and other black boxes as devices installed to help-optimize our work. Of course they must be programmed for that. Even the best operator is only human thus capable to make mistake

Best regards, Josip

[US Iceman]

Hi Josip,

You raised some good points in your last reply.

Having good plant with enough capacity, with good control of each device we can go very low with condensing pressure but then we have to heat something i.e. defrosting cycle, what now?

Switch on manual and come back to 30/35C (speaking about NH3).

I think we can design the systems to operate at 10C (50F) as a reasonable condensing temperature, when the wet bulb temperature or ambient conditions allow.

But, we have to determine how not to use components that have a higher minimum inlet pressure (TXV's or liquid feed valves).

Another way to look at the problem is as you suggest; run low discharge pressures when possible and reset to higher pressures for a short time.

Another possibility is to use a separate compressor for the hot gas source, while the other compressors are running low discharge pressures.

I do not think there is a single solution to this, but there are a lot of opportunities!

[Sergei]

[quote=US Iceman;57175]
But, we have to determine how not to use components that have a higher minimum inlet pressure (TXV's or liquid feed valves).
Regarding to TXV. Usually, high temperature evaporators are equipped with ammonia TXV. Major refrigeration load for the coolers with these evaporators depends on outside temperature. Lower condensing temperature we can get at cool ambient conditions. Capacities of TXV's are going down, but refrigeration load is going down as well.
Regarding to liquid feed valves. Usually, we have solenoid and expansion(balancing) valve. Expansion valve, usually, set to 50% of operating time for high condensing pressure. Sometimes this setting is working well for all(low and high) condensing pressures. Sometimes 30% of operating time is working fine for all condensing pressures. Sometimes you need 2 setting(summer and winter). I think it is not difficult to readjust this expansion valve twice per year.

Sergei