AWHP superheat & sub-cooling

[Gary]

^ The super-heater coil row is generally placed in the path of hottest air stream - ie. at air inlet. The air leaving super-heater is then used to evaporate the lower temp liquid.

Another thing to notice is that, in general, the cold air outlet temp is rarely colder than the superheated refrigerant outlet temp. In other words, very little 'temperature cross' occurs.

My commercial evaps do actually have around 1-1.5'C temp cross, but the lab machine can't get there - inefficient coil design.

Each has its advantages.

Placing the superheater up front is more effective at adding superheat, but then the air entering the evap will be a little cooler, reducing the evap capacity gain.

Placing the superheater after the evap maximizes evap capacity, but there may not be enough temp difference to effectively dampen the hunting at low load.

It would be preferable to have the superheater after... if it can do the job.

[desA]

Here's a thought for you:

Hot liquid ex condenser, into front row of an evap coil.

Sub-cools liquid & heats up incoming air-stream. This air can now operate on evap/superheater & may be useful under cooler ambient air conditions. Less issues with defrosts, I'd think.

[Gary]

Here's a thought for you:

Hot liquid ex condenser, into front row of an evap coil.

Sub-cools liquid & heats up incoming air-stream. This air can now operate on evap/superheater & may be useful under cooler ambient air conditions. Less issues with defrosts, I'd think.

This would transfer the heat from the liquid to the refrigerant flow in the evap coil. I'm not sure this would raise the coil capacity... but I could be wrong.

The key to all of these configurations is the dT,air. If the delta-T is higher, then more heat is being removed from the air, therefore the coil capacity is increased.

[desA]

VIC / SGHX re-visited

I have listed a few simulations run in Coolpack, to re-test the concept of VIC, or SGHX, as I was finding that all the gains in the liquid sub-cooling, were essentially just forcing the evap to work harder, by reducing entry vapour fraction at entry.


No SGHX


30% thermal efficiency


50% thermal efficiency


75% thermal efficiency


100% thermal efficiency

[desA]

The above calculations assume no line losses, or compressor external heat loss.

It can be seen that with more heat interchange from liquid line to suction line, that the COP does improve, but with astronomical increase in compressor discharge temperature.

[Gary]

VIC / SGHX re-visited

I have listed a few simulations run in Coolpack, to re-test the concept of VIC, or SGHX, as I was finding that all the gains in the liquid sub-cooling, were essentially just forcing the evap to work harder, by reducing entry vapour fraction at entry.

Is your simulator broken?

[desA]

Is your simulator broken?

Not at all. I thought that Coolpack was usually received as being an authority amongst many hvac folks, It is easy to show up the results graphically, with little fuss. The log(p)-h diagram is an easy background to discuss the effect of any cycle changes, as well.

My simulator turns out some 10 pages for each such run - right down to individual fluid properties at each point in the system.

[Gary]

Are we making any progress?

[Gary]

VIC / SGHX re-visited

I don't understand why everyone wants to equate the VIC with the SGHX. They are NOT the same thing.

[Gary]

There we go. Evap>VIC>s/h

I'll run up some tests tomorrow to see what comes out...

How did these tests go?

[Gary]

Another thing to notice is that, in general, the cold air outlet temp is rarely colder than the superheated refrigerant outlet temp. In other words, very little 'temperature cross' occurs.

My commercial evaps do actually have around 1-1.5'C temp cross, but the lab machine can't get there - inefficient coil design.

What is the approach temp on your commercial evaps?

[desA]

Originally Posted by desA

VIC / SGHX re-visited

Gary:
I don't understand why everyone wants to equate the VIC with the SGHX. They are NOT the same thing.

Can we please review the essential differences between the VIC & SGHX?

[Gary]

The VIC effectively increases the coil capacity without the high discharge temp.

Seems like a big difference to me.

[desA]

Are we making any progress?

Back in the saddle. Had to pop over to a neighbouring country for the day. Apologies for not reporting back earlier - playing catch-up at the moment.

[Gary]

Hmmm... the load would have to drop sufficiently to deplete the surplus liquid before the VIC could go 2-phase.

[desA]

How did these tests go?

I'll finish these runs in the morning. Have done some preliminary runs - but not yet complete.

[Gary]

We have a variety of solutions available for resolving any hunting/floodback issues.

I feel confident that we can shift both flash elimination and superheating duties to the VIC, resulting in very substantial evaporator capacity increases.

[desA]

By mutual consent, the VIC technical discussions will now be continued in an alternative area.

The AWHP thread will continue with development & testing of the remaining areas still left to explore on the current lab heat-pump, with specific reference to superheat & sub-cooling as referenced at the start of this thread. There are still a number of open questions left to explore.

If anyone is still specifically interested in the VIC concept, please contact the authors Gary, or desA for further details, per private mail.

[desA]

Originally Posted by desA
Another thing to notice is that, in general, the cold air outlet temp is rarely colder than the superheated refrigerant outlet temp. In other words, very little 'temperature cross' occurs.

My commercial evaps do actually have around 1-1.5'C temp cross, but the lab machine can't get there - inefficient coil design.

Gary:
What is the approach temp on your commercial evaps?

TD = 20-25'K;
Approach = 7.05-7.6K
Tcross ~ -0.8 to -1.4K (ie. air leaves evap Ta,out colder than superheated refrigerant outlet Te,sup)

[Gary]

Has the second lab unit, with the 3-tiered condenser, been modified for testing yet?

[desA]

^ Not yet... I've continued on the 2-tiered unit, running it for protracted runs to test the stability of recent changes over time & under varying atmospheric conditions.

On that 3-tier heat-pump, I'll have to also re-wire back to single-phase - so. I'll have to get my head together for that series of mods.

I'll plan do that after a trade show I need to attend in BKK during early October. This may be a good time to introduce a bunch more changes as well. I also have a spare plate HE on hand that will knock the socks off the 3-tiered condenser.

Welcome VIC...

[desA]

TXV hunting

I've been doing a study around Magoo's rule for selecting evap superheat as a function of TD.

The suggested rule for evaps is : SH = (0.6 -> 0.7)*TD

I let the multiplier, here 0.6 -> 0.7 be a variable (let's call it beta), & observed that Ta,out is generally Te,sup (evap outlet temp) plus, or minus another variable - lets call it alfa.

Plug all this into the relationship for evap heat-transfer:

Q'e = (UA)*dTlm

and out comes something very interesting - when the value of beta is determined. I expected a single value to pop out & land somewhere in the range of 0.6 - 0.7 .

In all cases, the relationships show that beta can take on two (so far) values - in other words, there is a low value of SH, or a high value of SH, that will give the same answers for the heat-transfer equation. Some of this must stem from the non-linearity of the dTlm term. This is termed as 'bi-stability' in mathematical terms.

Now, can I ask the question again:
What is TXV valve hunting & what are the observed symptoms?

[Gary]

Keep in mind that a TXV is a proportional control, as opposed to PI or PID. There is a reaction time involved as well as the magnitude of reaction (orifice size). If the orifice is undersized for the current load, the reaction will be insufficient to close the offset. If the orifice is oversized for the current load, the reaction will be too great for the bulb to sense and moderate within the time constraint thus the mass flow will oscillate beyond the TXV's ability to progressively decrease the oscillations, driving the oscillations to the maximum.

hunting = maximum oscillation

I'm thinking the above is probably true of all proportional controls.

[desA]

^ Thanks, Gary. This makes sense.

Now, I wonder if these minimum & maximum oscillation bounds would correspond to the two beta (=SH/TD) values that an evaporator naturally can swing to?

I've done further research on this bi-stability phenomenon for the evap characteristic & it does seem to be one of the things that contributed to early CO2 system bi-stability problems.

What seems to be showing up is that there are two scenarios:
1. Low fan speed, large dTa => lower beta value;
2. High fan speed, small dTa => higher beta value.

So, then, it seems that the SH selection for a particular evaporator configuration & at a particular operating point, will have to be selected carefully.

I have a feeling that the beta=0.6->0.7 rule probably works well in cold-room situations, or indoor air-conditioning applications where TD is more moderate. For heat-pumps, this value may very well move to a more appropriate range, suitable for the large TD's experienced in AWHP applications.

Let's see where this all goes.

[mad fridgie]

Keep in mind that a TXV is a proportional control, as opposed to PI or PID. There is a reaction time involved as well as the magnitude of reaction (orifice size). If the orifice is undersized for the current load, the reaction will be insufficient to close the offset. If the orifice is oversized for the current load, the reaction will be too great for the bulb to sense and moderate within the time constraint thus the mass flow will oscillate beyond the TXV's ability to progressively decrease the oscillations, driving the oscillations to the maximum.

hunting = maximum oscillation

I'm thinking the above is probably true of all proportional controls.

Excellent description "well done"

With the 2 biggest process variables being Ta(rh) and Tw (water in) both underfeed and overfeed are likley to occur on a unit this size (money limited)
I would suggets that you simulate at your design ambient/water limitations, ensuring that at these points your are within what you feel is acceptable liitations on the refrigeration machinery.
I then would look at system optomisation at a set of conditions, that are accepted as a standard in your area. I know this is a commercial issue, but at the end of the day you are developing a commercial product (finacially speaking) You then have one less variable to worry about

[Gary]

Excellent description "well done"

Thanks, Mad.

[Gary]

To better understand proportional controls, we need to start with the concept of offset.

If we adjust the TXV for optimum superheat under optimum conditions, there will always be an offset under non-optimum conditions.

If the heat load exceeds optimum, the bulb temperature increases which in turn opens the valve and increases the mass flow. But there is a catch22.

In order to match the increased load, the valve must remain open, but the valve cannot remain open without higher superheat as it is the superheat that opens the valve.

This 'tug of war' ends up in a compromise with the superheat higher than the optimum value. This difference between the actual superheat value and the optimum superheat value is the offset. The further from optimum load conditions, the greater the offset.

For the same reasons, there is always a negative offset under lower than optimum heat load conditions.

[desA]

Originally Posted by Gary
Keep in mind that a TXV is a proportional control, as opposed to PI or PID. There is a reaction time involved as well as the magnitude of reaction (orifice size). If the orifice is undersized for the current load, the reaction will be insufficient to close the offset. If the orifice is oversized for the current load, the reaction will be too great for the bulb to sense and moderate within the time constraint thus the mass flow will oscillate beyond the TXV's ability to progressively decrease the oscillations, driving the oscillations to the maximum.

hunting = maximum oscillation

I'm thinking the above is probably true of all proportional controls.

mad_fridgie:
Excellent description "well done"

With the 2 biggest process variables being Ta(rh) and Tw (water in) both underfeed and overfeed are likley to occur on a unit this size (money limited)
I would suggets that you simulate at your design ambient/water limitations, ensuring that at these points your are within what you feel is acceptable liitations on the refrigeration machinery.
I then would look at system optomisation at a set of conditions, that are accepted as a standard in your area. I know this is a commercial issue, but at the end of the day you are developing a commercial product (finacially speaking) You then have one less variable to worry about

Fair-enough comments. I've go that aspect pretty well sorted now, under our conditions. Gets hotter & more humid by the day at this time of year. The lab system is running sweetly with all her spanks & tickles.

The main issue I am raising here, in terms of the evap superheat issue, is that the evaporator is itself an unstable device - with the ability to 'select' a preferred operating state, given the correct set of prevailing conditions at the time. This is a symptom of bi-stability.

Now, we aim for a low(ish) SH to either suit the Magoo rule (0.6-0.7)*TD, or a slightly lower multiplier, or 7K.

Now, I've checked the (0.6-0.7) rule against low TD (10K) evap figures & it fits perfectly. In this case, there is almost NO bi-stability evident as the low TD drives the two beta (=SH/TD) values close together.

It is an entirely different matter when you start moving into high TD territory - here the bi-stability really begins to kick in.

The issue here is how to manage this in a safe, sane manner - while still keeping the compressor within safe bounds. In other words:
What should the appropriate SH (or beta=SH/TD) value be for high TD conditions?

This some new ground we are exploring here, of this I'm sure. Remember, Japan's local conditions are far more moderate than SE Asia.

[Gary]

People who calibrate proportion thermostats are all too familiar with offset. You calibrate the stat perfectly one day and the next day it is not maintaining the setpoint temp, because the weather changed. There is an offset between actual temp and setpoint temp.

[desA]

To better understand proportional controls, we need to start with the concept of offset.

If we adjust the TXV for optimum superheat under optimum conditions, there will always be an offset under non-optimum conditions.

If the heat load exceeds optimum, the bulb temperature increases which in turn opens the valve and increases the mass flow. But there is a catch22.

In order to match the increased load, the valve must remain open, but the valve cannot remain open without higher superheat as it is the superheat that opens the valve.

This 'tug of war' ends up in a compromise with the superheat higher than the optimum value. This difference between the actual superheat value and the optimum superheat value is the offset. The further from optimum load conditions, the greater the offset.

For the same reasons, there is always a negative offset under lower than optimum heat load conditions.

Thanks so much for that excellent review on proportional control & its downsides. This is very true.

A few thoughts:
1. Acceptable Te,sat for the compressor can be managed, reasonably.
2. What about maximum acceptable compressor suction temperature?

[mad fridgie]

The question is really how can we control the system, if the system is out of control, At present you limiting performance based upon the compressor manufactures envelope.
Take a step back it is a compressor, (larger systems are designed this way)
Say for instance Te is 25C superheat is 20C Tc is 45C, will this cause a problem?
Low compression ratio,
High mass flow
Is the motor over current?
Is the discharge overtemp?
If no, is there a problem? (apart from outside specs)
How long would this situation occur?
At some point a line in the sand has to drawn for best performance, component protection, is covered by your designated limits of your design

[Gary]

And then came electronic controls with proportional/integral (PI) and proportional/integral/derivitive(PID). These are strategies for reducing the offset.

[mad fridgie]

And then came electronic controls with proportional/integral (PI) and proportional/integral/derivitive(PID). These are strategies for reducing the offset.

Can handle, P & PI ,"D" always ****s me up, "thanks" for auto tune

[Gary]

The issue here is how to manage this in a safe, sane manner - while still keeping the compressor within safe bounds. In other words:
What should the appropriate SH (or beta=SH/TD) value be for high TD conditions?

In general, Copeland wants minimum 11K superheat at the compressor inlet (although this may be different for scrolls). So... I guess the correct answer would be, the superheat at the bulb should be as low as possible, but not low enough to result in less than 11K superheat at the compressor inlet nor low enough to cause hunting, under the lowest load conditions (maximum negative offset).

My experience would indicate that the formula is upside down for high TD coils, that the superheat should be .3*TD to .4*TD.

[mad fridgie]

SH, first priority is to ensure "No Liquid return" Max SH is 25C, but not at all conditions, Bigger the compression ratio the smaller allowable suction sh, (lower discharge) smaller comprssion ratio higher allowable SH. All limited by motor current.
I would say if your dicharge temp was under 105C and your moter current was 5% less than the name plate FLA, you would be pretty safe.

[desA]

The question is really how can we control the system, if the system is out of control, At present you limiting performance based upon the compressor manufactures envelope.
Take a step back it is a compressor, (larger systems are designed this way)
Say for instance Te is 25C superheat is 20C Tc is 45C, will this cause a problem?
Low compression ratio,
High mass flow
Is the motor over current?
Is the discharge overtemp?
If no, is there a problem? (apart from outside specs)
How long would this situation occur?
At some point a line in the sand has to drawn for best performance, component protection, is covered by your designated limits of your design

All well & good, but, what does your compressor supplier have to say about this? More importantly, will they try & void warranty claims if you knowingly regularly move outside their operating window?

[desA]

Originally Posted by Gary
And then came electronic controls with proportional/integral (PI) and proportional/integral/derivitive(PID). These are strategies for reducing the offset.

mad_fridgie:
Can handle, P & PI ,"D" always ****s me up, "thanks" for auto tune

A word of wisdom to the wise in terms of controllers trying to control unstable processes, with dead-time (time delays):
1. PID are not always the best option;
2. Autotune functions can have absolutely disastrous consequences - if misapplied. Be well warned.

Unstable processes, with inherent dead-time, are the control engineer's nightmare - if, of course, they understand control philosophy properly.

[desA]

Originally Posted by desA
The issue here is how to manage this in a safe, sane manner - while still keeping the compressor within safe bounds. In other words:
What should the appropriate SH (or beta=SH/TD) value be for high TD conditions?

Gary:
In general, Copeland wants minimum 11K superheat at the compressor inlet (although this may be different for scrolls). So... I guess the correct answer would be, the superheat at the bulb should be as low as possible, but not low enough to result in less than 11K superheat at the compressor inlet nor low enough to cause hunting, under the lowest load conditions (maximum negative offset).

My experience would indicate that the formula is upside down for high TD coils, that the superheat should be .3*TD to .4*TD.

Copeland's literature varies, depending on which design standard you're using:
1. ARI - SH 11.1K ; SC 8.88K
2. EN12900 - SH 10K; SC 0K

They also cut minimum evap temp with 10K suction SH & 25'C suction gas return.

Other suppliers are similar, but have some variations on this general theme.

How would the original concept of SH 7K fit into all of this?

[desA]

SH, first priority is to ensure "No Liquid return"

Fair comment.

Max SH is 25C, but not at all conditions,

Did you mean to say max temp return is 25'C, or really max SH is 25K?

Bigger the compression ratio the smaller allowable suction sh, (lower discharge) smaller comprssion ratio higher allowable SH. All limited by motor current.
I would say if your dicharge temp was under 105C and your moter current was 5% less than the name plate FLA, you would be pretty safe.

Fair comments.

[mad fridgie]

All well & good, but, what does your compressor supplier have to say about this? More importantly, will they try & void warranty claims if you knowingly regularly move outside their operating window?

Another good point, I can say that I have never had this issue arise (blown a few up due to floodback), I see these issues as going through a transision to reach design which should be within limits.
I do attempt to keep my machines within operatinal limits (somewhat easier the bigger you go) but I am also aware that occassionally we are going to move outside, I do not stretch compression ratios, I do not mind reducing compression ratios (within limits) higher Te, as long as Tc is low (motor output limit)
I do not mind high suction super heat as long as I have low compression ratios.

[Gary]

Copeland's literature varies, depending on which design standard you're using:
1. ARI - SH 11.1K ; SC 8.88K
2. EN12900 - SH 10K; SC 0K

They also cut minimum evap temp with 10K suction SH & 25'C suction gas return.

Other suppliers are similar, but have some variations on this general theme.

How would the original concept of SH 7K fit into all of this?

Liquid droplets disappear between 3.5K and 5.5K SH, so 7K is a safe setting, although a little lower might yield a little more performance.

[desA]

Another good point, I can say that I have never had this issue arise (blown a few up due to floodback), I see these issues as going through a transision to reach design which should be within limits.
I do attempt to keep my machines within operatinal limits (somewhat easier the bigger you go) but I am also aware that occassionally we are going to move outside, I do not stretch compression ratios, I do not mind reducing compression ratios (within limits) higher Te, as long as Tc is low (motor output limit)
I do not mind high suction super heat as long as I have low compression ratios.

Thanks, m_f...

[desA]

Liquid droplets disappear between 3.5K and 5.5K SH, so 7K is a safe setting, although a little lower might yield a little more performance.

Thanks, Gary.

[Chef]

To better understand proportional controls, we need to start with the concept of offset.

If we adjust the TXV for optimum superheat under optimum conditions, there will always be an offset under non-optimum conditions.

If the heat load exceeds optimum, the bulb temperature increases which in turn opens the valve and increases the mass flow. But there is a catch22.

In order to match the increased load, the valve must remain open, but the valve cannot remain open without higher superheat as it is the superheat that opens the valve.

This 'tug of war' ends up in a compromise with the superheat higher than the optimum value. This difference between the actual superheat value and the optimum superheat value is the offset. The further from optimum load conditions, the greater the offset.

For the same reasons, there is always a negative offset under lower than optimum heat load conditions.

Gary - It would be really interesting if you could expand upon this offset. Why and how does the offset become either positive or negative under different heat loads when it is the SH itself that is the control parameter? Are you referring to a change in deltaP across the evaporater at differant flow rates or some other phenomena?

Chef

[desA]

I'd like to thank everyone who has been involved in this thread, it has been a privilege being able to debate back & forth with so many members with such extensive knowledge.

One thing that has become apparent through all of these many pages is that there are many different ways to arrive at a solution for any given RHVAC-related system - some better than others, others applicable under the circumstances in which they operate.

I wish the very best to all the contributors & trust that any further readers learn a great deal from the contents of this thread.

I wish you all the very best. My personal questions originally asked, have been answered - as have reams more.

Take care:
Gary, Magoo, Chef, mad_fridgie, & anyone else I may have missed. A huge thanks to Gary for his many hours of patience & practical advice.

I think we're done...

[Brian_UK]

At the request of the original poster, desA, this thread is now closed.

If you have any other viewpoints or questions similar to the original post then do please start a new thread.

Brian