AWHP superheat & sub-cooling

[desA]

Do I get a free sample when we're done?

Yes, I'd think that would definitely be in order... Compressor would have to be 60Hz, though.

When we're set, I'll e-mail you a specific design schedule for you to play with on your end. You can then tweek further until this animal becomes a world-beater. We can then communicate per e-mail at that level...

I did also promise Magoo a general spec, so that he can build a 'retirement heat-pump' as a hobby.

At this point, I think Mad_Fridgie can also chip in with some of his bits...

[Gary]

^ It's covered in layers of lagging... sticky stuff was all I could lay my hands on, up here.

It is exactly the same construction as the 3x above, except with only 2 layers of coils. The interconnects, manifolds & so forth are precisely the same.

Does the liquid refrigerant come off the bottom?

[desA]

^ Nope - top as well...

Exactly the same as the 3x picture - exactly the same... I feel that this is a design flaw.

[mad fridgie]

How do you intend to protect your works?

[desA]

How do you intend to protect your works?

A very, very valid question. Most patents are useless & impossible to enforce.

Agreements with builders follow an NDA (non-disclosure agreement) - also difficult to enforce.

The trick, I've found, is to have certain 'key elements' hidden in the design itself, which, without them, the concept will not perform.

How do you protect your IP over there?

[mad fridgie]

I have to say this is the problem i have at moment, The concepts are world beating, not the specifics. In NZ the industry is small so reputation is important, Hopefully this reduces the chances of being screwed? (I hope)

[Gary]

^ Nope - top as well...

Exactly the same as the 3x picture - exactly the same... I feel that this is a design flaw.

I wonder if we shouldn't be experimenting with the machine you tested in Bangkok instead?... or putting the same type of condenser in this one?

[desA]

Originally Posted by desA
^ Nope - top as well...

Exactly the same as the 3x picture - exactly the same... I feel that this is a design flaw.

Gary:
I wonder if we shouldn't be experimenting with the machine you tested in Bangkok instead?... or putting the same type of condenser in this one?

I'll keep the BKK machine out of the internet limelight, for now - I think that would be best - it also has a different compressor, piping & a whole bunch more - than the lab machine. That will probably have to remain proprietary for now, I'd guess. Sorry about that.

The reason I have been testing the current lab machine, as it is, was that it doesn't cost a great deal to hack & tweek. I also wanted to find out just how far we could push a tube-in-tube condenser coil machine - & also where the control weaknesses lie. These tube-in-tube coils are also very effective for low-cost budget heat-pumps, where size is not a huge issue.

(For instance, the 2nd generation 'BKK' machine is now 40% of the volume of the original lab machine, at higher performance - something like a +207% power density improvement over this lab machine.)

I'm happy to for instance, re-route the condensate flow off the bottom of the 2x, or 3x coil, if necessary, in order to test manifold configurations.

I have 2 more spare lab machines of similar configuration - all piped up with 3x coils in place. I just need to swap out the compressor (3-phase) & re-wire the control panel. The box shape etc is identical.

[Gary]

The reason I have been testing the current lab machine, as it is, was that it doesn't cost a great deal to hack & tweek. I also wanted to find out just how far we could push a tube-in-tube condenser coil machine - & also where the control weaknesses lie. These tube-in-tube coils are also very effective for low-cost budget heat-pumps, where size is not a huge issue.

I'm happy to for instance, re-route the condensate flow off the bottom of the 2x, or 3x coil, if necessary, in order to test manifold configurations.

Fair enough.

Both the 'condensate out' and the 'water in' should be piped off the bottom of their respective headers.

[desA]

Fair enough.

Thanks Gary.

Both the 'condensate out' and the 'water in' should be piped off the bottom of their respective headers.

Excellent comments. The waterside bottom entry is critical - you are correct. This will help with the natural convection draw at the water heats up.

The condensate exit should always be at a low point, to prevent liquid backup & blinding of the lowest coil.

Good points... I'll get those implemented on one of the other lab machines (same evap, TXV (need to change orifice), piping - need to change compressor - have one size up in house, this will fit well to the 3x coil)

[Gary]

Excellent comments. The waterside bottom entry is critical - you are correct. This will help with the natural convection draw at the water heats up.

It will also equalize the flow through the coils and push any air bubbles through.

[Gary]

Yes, I'd think that would definitely be in order... Compressor would have to be 60Hz, though.

And single phase, too.

[desA]

Originally Posted by desA
Excellent comments. The waterside bottom entry is critical - you are correct. This will help with the natural convection draw at the water heats up.

Gary:
It will also equalize the flow through the coils and push any air bubbles through.

Very true on the flow equalisation part.

The air bubble clearance is always an issue - many of us forget about that one, unfortunately.

[Gary]

A note on Tcomp,d control of fan:
On a circulating water loop, where Tw actually holds Tc,sat up, trying to tweek fan speeds & so forth off Tcomp,d helps nothing - in terms of cutting the process, that is. If the fan speed is dropped, Te,sat reduces, raising the instantaneous pressure ratio & driving Tcomp,d up further for a while, until the heat-pump performance degrades enough to begin to steady out. This is a waste of energy - better to switch the compressor off, in my view.

With the CPR holding the Pcomp,in at 76psi, the compression ratio remains the same so long as the Te,sat is above 15C/76psi.

Possibly we need to control fan speed off Tcomp,in instead of Tcomp,d... or maybe both.

[desA]

With the CPR holding the Pcomp,in at 76psi, the compression ratio remains the same so long as the Te,sat is above 15C/76psi.

True...

Possibly we need to control fan speed off Tcomp,in instead of Tcomp,d... or maybe both.

A thought - why not control off Te,sat (or P,LO) in the evap itself? What happens is that if there is an instantaneous mismatch between the TXV feed into the evap, & the allowed outlet (relatively locked) through the orifice/CPR/valve, the Te,sat value in the evap will move - the amount of movement subject to the TXV superheat adjustment.

So, tighten SH & Te,sat will rise - loosen SH, Te,sat will lower.

Controlling fan speed off the Te,sat signal can allow the evap internal balance to be well managed, by adjusting heat input from the incoming air.

[mad fridgie]

At present you are using a 4 row coil,
Have you looked into a coil with lesser row and increased face area, using a larger diameter fan. this maybe? gives a better option for fan speed control.
you also need to consider the moterinput of a new fan and the different air pressure drop through the coil

[desA]

^ I've seen some folks using large face-area, thin coils, with two fans. They switch off the one fan, under certain circumstances - I had understood that was for noise reduction, rather than evap control.

My current focus has been in the other direction - for small, compact evaps with small fans.

[mad fridgie]

Noise is an issue in it own right!
Fan speed, tip speed, velocity. We will leave this for now.
Another simple way of controlling load is a air by-pass, a simple damper that that opens when Te rises, many control scenerios. Damper blade cheap, damper motor cheap, No CPR, No Fan speed control

[desA]

Noise is an issue in it own right!
Fan speed, tip speed, velocity. We will leave this for now.

You are 100% correct. Fan noise is a real issue & something pretty difficult to overcome, at decent cost & fan draw.

Another simple way of controlling load is a air by-pass, a simple damper that that opens when Te rises, many control scenerios. Damper blade cheap, damper motor cheap, No CPR, No Fan speed control

This idea, I like very much. I'd had a more than a few thoughts on this, to be honest. I'd love to see examples of how some folks have solved this.

Would this perhaps be a small servo-motor, or actuator? I have space to implement this & would be very keen to try. Excellent idea, excellent.

Thanks for this.

[Gary]

A thought - why not control off Te,sat (or P,LO) in the evap itself? What happens is that if there is an instantaneous mismatch between the TXV feed into the evap, & the allowed outlet (relatively locked) through the orifice/CPR/valve, the Te,sat value in the evap will move - the amount of movement subject to the TXV superheat adjustment.

So, tighten SH & Te,sat will rise - loosen SH, Te,sat will lower.

Controlling fan speed off the Te,sat signal can allow the evap internal balance to be well managed, by adjusting heat input from the incoming air.

I was hoping to control off the discharge temp to compensate for voltage drops in the area, guarding directly against heat rise in the compressor, but we may not be able to do this.

But yes, controlling off Te,sat would be another alternative.

Another low cost strategy which we haven't tested yet is the AEV, and this may turn out to be ideal for your immediate area, where the ambient rarely drops below 20C.

The AEV locks the Te,sat at a steady 15C, eliminating the need for the CPR. As the ambient temp rises the superheat rises. The fan control could then bring the superheat back down. As the compression ratio is not effected, we may be able to control off T,comp,disch.

The downside is that if the ambient drops below the AEV setting, heat transfer stops. If we set the AEV lower for colder ambients, then we get less performance at warmer ambients. We can get around this with dual metering devices and solenoid valves, but we are adding to complexity and costs.

[Gary]

The downside is that if the ambient drops below the AEV setting, heat transfer stops. If we set the AEV lower for colder ambients, then we get less performance at warmer ambients. We can get around this with dual metering devices and solenoid valves, but we are adding to complexity and costs.

Hmmm... we may be able to get around this by placing a fixed orifice device (cap tube, flowrater, etc.) in series with the AEV... an idea worth exploring.

[desA]

I was hoping to control off the discharge temp to compensate for voltage drops in the area, guarding directly against heat rise in the compressor, but we may not be able to do this.

I was hoping for the same thing, but, it does look like it will be difficult in some cases.

Prolonged under-voltage shows up in an amperage rise to the compressor & consequent compressor shell heating. This is one area which can be monitored. The compressor manufacturer has a recommended limit on this shell value as well.

But yes, controlling off Te,sat would be another alternative.

At least here, the control mechanism could be off evap TD, or approach, for instance - or just Te,sat itself.

Another low cost strategy which we haven't tested yet is the AEV, and this may turn out to be ideal for your immediate area, where the ambient rarely drops below 20C.

The AEV locks the Te,sat at a steady 15C, eliminating the need for the CPR. As the ambient temp rises the superheat rises. The fan control could then bring the superheat back down. As the compression ratio is not effected, we may be able to control off T,comp,disch.

The downside is that if the ambient drops below the AEV setting, heat transfer stops. If we set the AEV lower for colder ambients, then we get less performance at warmer ambients. We can get around this with dual metering devices and solenoid valves, but we are adding to complexity and costs.

Do variable AEV's exist? Meaning, AEV setting that is drawn from another relevant control temperature? Or, taken another way, why not move the TXV bulb to a more suitable location, where it can manage the evap more in line with this new control variable?

[Gary]

Do variable AEV's exist? Meaning, AEV setting that is drawn from another relevant control temperature?

A variable AEV is called a TXV. The whole point in using an AEV is that it is not variable, giving us precise Te,sat control.

[Gary]

Hmmm... we may be able to get around this by placing a fixed orifice device (cap tube, flowrater, etc.) in series with the AEV... an idea worth exploring.

On further thought, the same effect could be had by simply limiting the charge, which we would want to do anyway. The AEV may be a very good way to go.

[desA]

^ Ok, so basically, use an AEV instead of a TXV, & fine-tune for prevailing surrounding atmospheric conditions?

What about moving the TXV bulb to the downstream side of the CPR, or restrictor orifice?

[Gary]

^ Ok, so basically, use an AEV instead of a TXV, & fine-tune for prevailing surrounding atmospheric conditions?

Adjust the charge for lowest allowable superheat at the compressor inlet @ the lowest expected Ta,in with max fan speed. (Te,sat should drop below 15C at low ambient)

Adjust the AEV for Te,sat=15C at higher ambient.

Adjust the fan speed to maintain acceptable Tcomp,disch (and/or acceptable Tcomp,in) at highest expected Ta,in.

What about moving the TXV bulb to the downstream side of the CPR, or restrictor orifice?

TXV bulbs are slow acting. Moving the bulb further downstream could endanger the compressor (possibility of flooding), especially at low ambients where the valve may hunt.

[Gary]

I can now confirm that if the 50Hz operating conditions are moved up to Tc,sat=75'C, that similar compressor mass-flow conditions exist as for the compressor running at 60Hz, Te,sat=10'C, Tc,sat=65'C. This calculation also allows for efficiency variation (small).

I wonder how far this could be taken.

What would the safe operating envelope be if the compressor were run at 40hz?... or 30hz... or 20hz?

If this compressor were used for AWHP in America (at 60hz), how could it possibly produce water at 65C?

[desA]

I wonder how far this could be taken.

What would the safe operating envelope be if the compressor were run at 40hz?... or 30hz... or 20hz?

That's an interesting thought. Are you thinking here of using an inverter control on the compressor, for instance?

If this compressor were used for AWHP in America (at 60hz), how could it possibly produce water at 65C?

By using a low-approach condenser (based on Tc,sat-Tw), with low water flow-rate through the unit, you could get slightly past 65'C, at a squeeze.

If you wanted to push a little harder, then a dedicated desuperheater would need to be used, leaving the condenser space to condense at full performance.

[Gary]

That's an interesting thought. Are you thinking here of using an inverter control on the compressor, for instance?

Just exploring the nature of compressor envelopes.

By using a low-approach condenser (based on Tc,sat-Tw), with low water flow-rate through the unit, you could get slightly past 65'C, at a squeeze.

If you wanted to push a little harder, then a dedicated desuperheater would need to be used, leaving the condenser space to condense at full performance.

If the Tc,sat is limited to 65C, then getting 65C water would require a zero approach condenser.

A dedicated desuperheater might get a trickle of water to 65C.

I'm thinking an AWHP designed for 60hz would need a different compressor or inverter or something. How is it done?

[desA]

Just exploring the nature of compressor envelopes.

All good stuff. I'd like to see how an inverter-based heat-pump would perform & how the envelop would interact. Very, very interesting line of thought, here - very interesting.

If the Tc,sat is limited to 65C, then getting 65C water would require a zero approach condenser.

Not really. The approach as defined, is really describing the condensing section approach (=Tc,sat-Tw', where Tw'<Tw,out)), even though we commonly use it as Approach=Tc,sat-Tw,out.

By slowing down the water flow, the unit can be pushed to allow a small amount of additional temp, past the Tc,sat mark. It's not a lot, but is possible, with the right condenser.

A dedicated desuperheater might get a trickle of water to 65C.

Sure - function of the water flowrate.

I'm thinking an AWHP designed for 60hz would need a different compressor or inverter or something. How is it done?

I'd like to explore the inverter idea a little more.

Another option could be a cascade system, with the top fluid being able to manage higher Tc,sat, so pushing up the water allowable exit temp. This is something of interest to me, under conventional technologies - CO2 presently excluded.

[Gary]

Another option could be a cascade system, with the top fluid being able to manage higher Tc,sat, so pushing up the water allowable exit temp. This is something of interest to me, under conventional technologies - CO2 presently excluded.

Seems it isn't a matter of top fluid. It's the compressor pumping that top fluid.

What fluid could this compressor pump that would allow a higher Tc,sat?

[Gary]

Another option would be to increase the number of poles in the compressor motor. This would slow the motor just like reducing the hz.

[Gary]

The more interesting of the options is the inverter as the mass flow can be maximized and the limit raised.

But that doesn't seem to do much for the COP. From what you have said in previous posts, COP has nothing to do with mass flow and everything to do with Tc,sat. Or did I misunderstand your previous posts?

I'm having a problem wrapping my brain around this.

[desA]

Seems it isn't a matter of top fluid. It's the compressor pumping that top fluid.

What fluid could this compressor pump that would allow a higher Tc,sat?

The problem here is that there are few 'friendly' refrigerants with a T,critical much above 101'C. If we want to drive the temperature higher, a suitable refrigerant has to be used.

I have some products in mind & will pull up some info when I get back into the office (eta tomorrow). We can then bat these around in terms of suitability, availability, cost, or otherwise.

[desA]

The more interesting of the options is the inverter as the mass flow can be maximized and the limit raised.

But that doesn't seem to do much for the COP. From what you have said in previous posts, COP has nothing to do with mass flow and everything to do with Tc,sat. Or did I misunderstand your previous posts?

I'm having a problem wrapping my brain around this.

The COP for the compressor, at balance point of TXV-compressor operation, is a function of Tc,sat ; Te,sat & compressor isentropic efficiency. This is a pure thermodynamic balance, that has to be met.

As Tc,sat shifts upwards, for Te,sat fixed - the COP will drop off.

The thermodynamic cycles are worked out in kJ/kg & so are applicable across any mass-flow - this is a generic thermodynamic cycle - we say in 'specific terms' ie. divided by mass. When you need to know actual kW, you simply multiply by the mass flowrate for that design.

In this way, the cycle SCALES across the range. This is a useful design feature.

[Gary]

As Tc,sat shifts upwards, for Te,sat fixed - the COP will drop off.

So... if Te,sat and Tc,sat shift upwards in tandem, then COP remains the same?

[desA]

^ Not in direct proportion, as I understand it. It's easy-enough to test, though.

Give me a range of (Te,sat; Tc,sat) couples to calculate & I'll run these off on the simulator & feed back the results. An additional complication is that for a compressor, isentropic efficiency changes over the operating range as well - this adjusts the figures a little as well.

The alternative to this is to try & fix a COP & then look for Te,sat - Tc,sat couples that will match this - subject to isentropic efficiency variation.

[mad fridgie]

So... if Te,sat and Tc,sat shift upwards in tandem, then COP remains the same?

As a fairly good rule of thumb!
If you start a fixed two points "Te" "Tc"
If Te is constant and Tc is raised by 1C then you will consume 3.5% more power
If Tc is a constant and Te is reduced by 1C then you will loose 2.5% cooling capacity.
Whay really happen is when one changes, it generally effect the second

[mad fridgie]

Efficiency of the compressor can change dramatical at different compression ratios, both mechanically and electrically.
This can been seen quite well with recip comps , they are not 100% volumetric efficient, the higher the compression ratio, the less refrigerant leaves/enters the compression chamber

[mad fridgie]

Inverters, again your have to be careful on speed, you need to ensure that there is enough lubrication,
also you may find the the electrical wattage decreases, but have large changes in power factor and amperage. (if ia client is paying for KVA things may not look so rosey, and the power distribution companies certainly do not like poor power factor

[desA]

^ Some excellent points in the above 3 posts, mad_fridgie... thanks for those.

[desA]

Inverters, again your have to be careful on speed, you need to ensure that there is enough lubrication,

Are you referring here to the oil conveyance in the pipes themselves, or the compressor itself?

The line conveyance is certainly a huge issue, if managed carelessly. There would either have to be a minimum system turn-down ratio, or design for over-speed at the upper ranges of operation.

Seems very much to be a trade-off game, balanced around the nominal design point of operation. Very interesting...

[Chef]

DesA - Danfoss do a nice range of variable speed scroll compressors and range from 30Hz to 90Hz.

If you are using fan control and a CPR to control suction (throttling) of the compressor you will see from a simulation you get an increase in COP by varying the compressor speed instead. The main advantage is a system that operates at the best isentropic effeciency across the whole range from start-up to final temperature.

Introducing fan speed control and the CPR will reduce the systems effeciency when they kick in. COP might be generally independant of the mass flow rate but its not independant of system pressure drops however induced.

Chef

[mad fridgie]

Are you referring here to the oil conveyance in the pipes themselves, or the compressor itself?

The line conveyance is certainly a huge issue, if managed carelessly. There would either have to be a minimum system turn-down ratio, or design for over-speed at the upper ranges of operation.

Seems very much to be a trade-off game, balanced around the nominal design point of operation. Very interesting...

Actually I did mean both, I think you will find as the units become larger (capacity) there is a much wider range of options for equipment and controls, which can allow for greater COPs across a larger working range

[desA]

^^ Thanks Chef for the input.

A nice range of variable speed compressors is something I'd like to look into for the upper-end range of machines. I wonder what the comparative price ratio would be for the Danfoss system & its controls, compared to a simple single speed scroll, with a small amount of throttling (loss)?

To my mind, there seems to be a logical migration to two ends of the heat-pump scale - balanced, stable budget systems against high-performance, higher-priced systems.

[desA]

Actually I did mean both, I think you will find as the units become larger (capacity) there is a much wider range of options for equipment and controls, which can allow for greater COPs across a larger working range

Thanks, mad_fridgie, for the good advice.

[desA]

A question regarding Te,sat drift

This Te,sat drift phenomenon whereby the evaporator sat pressure begins at a reasonable value of say 10'C at start of heating cycle, but gradually begins to drift upwards, with an overall drift of sometimes as much as 5'C during a heating cycle, seems to be a consequence of the refrigerant loop itself.

At times, if Te,sat begins at say 10'C, but drifts up to 15'C at cycle end, then compressor integrity is not threatened. On the other hand, for a different combination of Ta,in & RH%, Te,sat can begin around 14.5'C & drift towards 19.5'C by cycle end - definitely causing concern for the compressor.

My question is that I've seen countless manufacturers - many from colder climates, it must be admitted - but none of them seem to even consider the Te,sat drift phenomenon.

When this phenomenon is brought to the attention of the Asian compressor technical experts, they all turn a whiter shade of pale & admit that "it is difficult to find a compressor that can make 55'C water in Asia".

Practically, I've seen some roo-based machines suffer something like 20% compressor failures within an 18 month period.

What gives?
Are many manufacturers in denial?
Have manufacturer's not considered this Te,sat drift phenomenon?

[Chef]

The price and operating cost (and repair costs) need to be evaluated over a 1 year or 2 year period. It may be more expensive kit in the first place turns out to be cheaper to run and less cost to service in the long run. You will have to do the numbers but as you have pointed out the user is concerned with price per hot shower and for your market they probably wont think a week in advance.

It would be interesting to see the cost of the system using variable compressor and savings over a year on power. You will never know until you do the numbers.

Plus you get a PID controller to plug in the other stuff you have.

Chef

[Chef]

This Te,sat drift phenomenon whereby the evaporator sat pressure begins at a reasonable value of say 10'C at start of heating cycle, but gradually begins to drift upwards, with an overall drift of sometimes as much as 5'C during a heating cycle, seems to be a consequence of the refrigerant loop itself.

The drift is part of the cycle and you need to design it into your system and not try to get rid of it. If you look at the position of h4 during the heating cycle you will see it moves to the right. The quality changes and causes the drift, also the charge is repositioned in the system to compensate. One way is to reduce the volume flow rate of the system to balance the new (lesser) cooling effect. As you have said before the system is unstable and controlling from the TXV is also included in the instable part of the loop.

To get a stable loop you will need to control on an element in the loop that does not have a positive value in the feedback loop. One would be volume flow rate.

Chef

[Gary]

So... if Te,sat and Tc,sat shift upwards in tandem, then COP remains the same?

Okay... let's rephrase this:

If Te,sat and Tc,sat shift upwards in the proper proportions, then COP remains the same?

How closely would this relate to compression ratio? Would it be fair to say that if the compression ratio is maintained then the COP is maintained?