AWHP superheat & sub-cooling

[desA]

Hi Nike123, I see you lurking... welcome to comment...

[desA]

Update - water calibration curve:

What I have done is to add an additional thermocouple probe onto the condenser water outlet pipe. I will calibrate the two Tw,out probes against the T,tank probe tomorrow & adjust the experimental results accordingly.

Practically, though, I'd expect the useful data for the Approach to be for (Approach) to the left of the temp crossing point & for (Approach)' to the right of the temp crossing point. Let's see what skewing the calibration curves bring out.

[Gary]

We have been assuming that T,tank is the same as Tw,in.

It is not possible for Tw,in (T,tank) to be higher than Tw,out.

Nor is it possible for Tw,in (T,tank) to be higher than Tc,exit.

I tend to believe the problem is T,tank.

To my thinking, it would be far preferable to measure Tw,in and assume that it is the same as T,tank than the reverse.

[Gary]

The TXV is working very well despite the increase in liquid pressure. That's very good news.

At the end of cycle the compressor discharge temp is running dangerously high (103.9C). Since we are riding the upper limits of the compressor, we need to provide maximum cooling for it. The compressor is cooled by the superheated vapor. Therefore we need to run the lower limits of acceptable superheat. Going by Magoos rule, at end of cycle the TD is 25.2-14=11.2TD.

11.2*0.6=6.72SH.

The TXV superheat could be lowered to about 6.72K from its current 7.9K. This would lower the compressor discharge temp.

[desA]

^Good point. Thanks for that. I'll drop that down a tad to suit.

What maximum discharge temp would you be comfortable with off a scroll compressor?

The Copeland document AE-1263-R3 states that discharge temp measured 6" from the compressor discharge should be max ~ 250'F (121.1'C), with internal compressor exit maxima no higher than 300'F (148.89'C).

On that basis, we still have a little spare room.

The other limits along the way are as follows:
1. Motor winding limit : 325'F (162.78'C)
2. Oil breakdown limit : 325'F (162.78'C)
3. Compressor oil film limit : 310-325'F (154.4-162.78'C)
4. Sump temperature limit : 200'F (93.3'C)

[desA]

We have been assuming that T,tank is the same as Tw,in.

True... thanks for the reminder - was getting lost in there for a second.

It is not possible for Tw,in (T,tank) to be higher than Tw,out.

Nor is it possible for Tw,in (T,tank) to be higher than Tc,exit.

I tend to believe the problem is T,tank.

To my thinking, it would be far preferable to measure Tw,in and assume that it is the same as T,tank than the reverse.

Point taken.

The problem with these type K temp probes is that the accepted experimental uncertainty is +- 2.2'C. Add to that positioning inaccuracy, tank mixing & so forth - & we have a quagmire.

I plan to run up a calibration trial on all 3 probes tomorrow. May have to start with ice & go on up to boiling to try & find out the bias of each... headaches.

[Gary]

^Good point. Thanks for that. I'll drop that down a tad to suit.

What maximum discharge temp would you be comfortable with off a scroll compressor?

Being from the old school, I was taught that if you spit on the discharge line and it boils you have a problem. But it seems that compressors are more tolerant these days.

Still, I like to keep the discharge a little cooler to ensure the compressor lives a long and happy life.

And a lower superheat picks up more heat in the evap, thus increasing COP.

[desA]

^ Points well taken.

I'll set the superheat down in the morning. That's sound advice.

The scroll compressor concept seems to be doing fairly well, it seems. Copeland seem to be continually revising their operational envelopes, based on field feedback.

[desA]

http://tinypic.com/r/250ou8n/3

What would be your comments on this chart?

(-Cross)=Tc,exit-Tw,out

From my heat-exchanger design experience, I would say that the condenser performance is beginning to drop off drastically after Tc,sat = 50'C, based on (-Cross), although the Approach drop-off seems to occur around Tc,sat = 55'C. The SC shows a rapid rise after Tc,sat = 55'C as well.

[Gary]

I would attribute the dropoff in performance to the excessive subcooling. Liquid expands when heated. The liquid has no receiver to expand into, therefore it backs up into the condenser, taking up valuable space.

[Gary]

At Tc,sat 75C, we can remove refrigerant to reduce the subcooling. If the superheat does not rise as a result, then the system has sufficient liquid refrigerant to feed the evap.

If reducing subcooling results in high superheat then the system needs a receiver.

[desA]

A few thoughts on condenser size & liquid back-up & sub-cooling in the longer term.

The current condenser in this AWHP is a set of two tube-in-tube condenser coils placed in parallel. What if the condenser surface area were to be enlarged by adding a third coil in parallel?

My logic here is that is should be better to err on over-surface for the condenser & so maximise the heat-output from the system. According to the compressor performance charts, the system is capable of pushing around 10-15% more performance than the current heat-pump can deliver.

(We can discuss system balances evap/compressor/condenser a little further down the track)

[desA]

I would attribute the dropoff in performance to the excessive subcooling. Liquid expands when heated. The liquid has no receiver to expand into, therefore it backs up into the condenser, taking up valuable space.

A few observations on these heat-pump dynamics as the heating cycle advances:

1. At startup condition, the evaporator has inlet quality of x=0.12 kg/kg.
2. At hot condition, the evaporator has inlet quality of 0.40 kg/kg.
3. As the heating cycle advances, & the evaporator inlet quality rises, the refrigerant mass charge begins to migrate from the evaporator towards the condenser - purely to attempt to maintain thermodynamic & mass balance equilibrium.
4. This mass migration then causes the condenser to become 'flooded' as it fills up & acts as a receiver.
5. If the level of flooding (as evidenced by the amount of sub-cooling) becomes excessive, then the sub-cooling area begins to encroach on the area required for condensation & the overall condenser performance begins to drop off.
6. If the condenser were to be enlarged sufficiently to cope with the flooding effect i.e. act as a receiver, then sufficient area would still remain for the condensing phase to occur unimpeded.

Interestingly, the mass migration effects can be observed during a heating cycle, by observing the compressor amperage behaviour. At times, the compressor seems to take on more load & can be heard to 'dig in' & pull slightly harder. During this time, small swings in condenser Tc,exit temp can be seen. Once this event settles, then the system stabilises at the new equilibrium point, settles down & all is smooth again. These events occur throughout the cycle & can be seen on the amperage-time trace as small waves.

This is a consequence of real-time system dynamics (unsteady) imposed on a thermodynamic system that presupposes & is designed for, a 'steady' operating regime.

This is why I prefer a thermal bulb to drive the TXV in such situations & not use an electronic variant. Thermal bulbs have intrinsic thermal inertia & a slow response - this helps to smooth the refrigerant waves in the system. If the wrong controller (fast response) were to be used, some very interesting wave dynamics could result. Sometimes, slower is better - in my view, at least.

[Gary]

A few thoughts on condenser size & liquid back-up & sub-cooling in the longer term.

The current condenser in this AWHP is a set of two tube-in-tube condenser coils placed in parallel. What if the condenser surface area were to be enlarged by adding a third coil in parallel?

My logic here is that is should be better to err on over-surface for the condenser & so maximise the heat-output from the system. According to the compressor performance charts, the system is capable of pushing around 10-15% more performance than the current heat-pump can deliver.

No doubt... bigger is better.

But we haven't seen what it will do with 35C incoming air yet.

[Gary]

A few observations on these heat-pump dynamics as the heating cycle advances:

1. At startup condition, the evaporator has inlet quality of x=0.12 kg/kg.
2. At hot condition, the evaporator has inlet quality of 0.40 kg/kg.
3. As the heating cycle advances, & the evaporator inlet quality rises, the refrigerant mass charge begins to migrate from the evaporator towards the condenser - purely to attempt to maintain thermodynamic & mass balance equilibrium.
4. This mass migration then causes the condenser to become 'flooded' as it fills up & acts as a receiver.
5. If the level of flooding (as evidenced by the amount of sub-cooling) becomes excessive, then the sub-cooling area begins to encroach on the area required for condensation & the overall condenser performance begins to drop off.
6. If the condenser were to be enlarged sufficiently to cope with the flooding effect i.e. act as a receiver, then sufficient area would still remain for the condensing phase to occur unimpeded.

There is no "area needed for condensation" as such. The vapor fills the available area and the more transfer surface, the more transfer. However the heat transfer per unit of surface diminishes, so there must necessarily be a point where the excess subcooling does less harm than good.

Still, I am not convinced that the added subcooling is needed. The optimum subcooling seems to be about 7K@75C for this system. If, at 7K@75C subcooling, there is sufficient refrigerant flow for the evaporator over the entire cycle, then the excess subcooling is not needed.

It also occurs to me that if the water were regulated such that the Tc,sat were held at 75C, then the subcooling could be stabilized at 7K, without infringement on the condenser.

Yet another point for the heat-it-in-one-pass school of thought.

Interestingly, the mass migration effects can be observed during a heating cycle, by observing the compressor amperage behaviour. At times, the compressor seems to take on more load & can be heard to 'dig in' & pull slightly harder. During this time, small swings in condenser Tc,exit temp can be seen. Once this event settles, then the system stabilises at the new equilibrium point, settles down & all is smooth again. These events occur throughout the cycle & can be seen on the amperage-time trace as small waves.

This is a consequence of real-time system dynamics (unsteady) imposed on a thermodynamic system that presupposes & is designed for, a 'steady' operating regime.

This is why I prefer a thermal bulb to drive the TXV in such situations & not use an electronic variant. Thermal bulbs have intrinsic thermal inertia & a slow response - this helps to smooth the refrigerant waves in the system. If the wrong controller (fast response) were to be used, some very interesting wave dynamics could result. Sometimes, slower is better - in my view, at least.

Not too fast. Not too slow. Not too much. Not too little. Not too big. Not too small. Balance is everything.

[desA]

No doubt... bigger is better.

But we haven't seen what it will do with 35C incoming air yet.

This is true. I'll have to work on that in the lab, or wait for monsoon season to stop... lol

Supposedly, the performance is to go up with increasing incoming air temp. My concern for this particular AWHP is that the evap is specified at maximum refrigeration capacity of around 13.5 kW, with an air entry face velocity of 2.032m/s (I kid not). The current measured air face velocities are in the range of 3-3.3 m/s.

The system thermodynamic balance calls for the following, at say Te,sat=12.5'C:
1. At Tc,sat=40'C :
Q'evap = 6.7kW
Q'cond = 7.3kW

2. At Tc,sat=50'C :
Q'evap = 6.1kW
Q'cond = 6.9kW

3. At Tc,sat=75'C :
Q'evap = 4.2kW
Q'cond = 6.1kW

What would be the effects of an oversize evaporator on this circuit?

[Gary]

The compressor is currently nearing its 15C limit with 25C incoming air. It may in fact exceed this limit with 35C incoming air. With a larger evap it would most certainly exceed it.

[Gary]

The system thermodynamic balance calls for the following, at say Te,sat=12.5'C:
1. At Tc,sat=40'C :
Q'evap = 6.7kW
Q'cond = 7.3kW

2. At Tc,sat=50'C :
Q'evap = 6.1kW
Q'cond = 6.9kW

3. At Tc,sat=75'C :
Q'evap = 4.2kW
Q'cond = 6.1kW

It's easier to make the water warm than it is to make the water hot.

[desA]

But we haven't seen what it will do with 35C incoming air yet.

For the latest AWHP build (own design), I am hoping to be able to test in my builder's environmental test chamber. This is rigged up to be able to manipulate air temps from Asian to European conditions.

I'm trusting that we can go up to a steady inlet of 35'C & test what goes on with the system.

[desA]

The compressor is currently nearing its 15C limit with 25C incoming air. It may in fact exceed this limit with 35C incoming air. With a larger evap it would most certainly exceed it.

Agreed. This is my contention with 'one evap fits range' philosophy.

For the new machines, the evap is critically sized & not more. I'm aiming for better air speed control as well.

[desA]

Hi Chef, please feel free to contribute...

[Gary]

How are we doing with the compressor current draw? Still within the specs?

[desA]

How are we doing with the compressor current draw? Still within the specs?

Good point. Bingo!!!

According to the compressor specs at Te,sat=12.5'C:

1. At Tc,sat=40'C : Current@230V = 5.83A - measured 4.9A @ 210V;
2. At Tc,sat=50'C : Current@230V = 6.86A - measured 6.25A @ 201V;
3. At Tc,sat=55'C : Current@230V = 7.49A - measured 7.2A @ 195V

4. At Tc,sat=60'C : Current@230V = 8.25A - measured 8.4A @ 192.5V
5. At Tc,sat=65'C : Current@230V = 9.19A - measured 9.6A @ 189.5V
6. At Tc,sat=70'C : Current@230V = 10.33A - measured 11.1A @ 191.5V
7. At Tc,sat=75'C : Current@230V = 11.72A - measured 13.25A @ 189.5V

Now that becomes more clear, doesn't it?

1. There is some level of current-voltage trade due to poor local power supply (known - a constant source of irritation);
2. The system starts to choke up after Tc,sat ~ 55'C.

[Chef]

Hi Chef, please feel free to contribute...

Your thread with Gary does not need any interferance and the data you post is very useful, certainly the most detailed for a long while - I am just watching to see how it finally pans out.

However I would like to hear more about the fluctuations you see in the compressor current versus time trace.

Just sitting on the porch.

Chef

[desA]

However I would like to hear more about the fluctuations you see in the compressor current versus time trace.

I'll detail that aspect a bit further into the study, with a fine current-time plot. The refrigerant 'wave' events certainly become fairly clear in the region where the sub-cooling has become dominant.

Just sitting on the porch.

Take care.

[Gary]

Have you checked the voltage at the mains? You could have an undersized wiring problem in addition to the known power supply problems.

Is the compressor 50hz or 60hz? Is the local power supply 50hz or 60hz? Single phase or three phase? Does the compressor have multiple power ratings?

[desA]

Have you checked the voltage at the mains? You could have an undersized wiring problem in addition to the known power supply problems.

I'll check the mains voltage again tomorrow at the main incoming feed.

In this part of the world, the main issue is that the power infrastructure has not kept abreast of the housing development & so some areas seem to be on the power borderline. What happens here is that during peak periods the available voltage drops off - it can go as low as 187V, on the odd occasion. I've even had so bad that a single airconditioner & computer cannot be operated at the same time, before the UPS begins to cry foul... lol

I tend to perform my testing at times of the day where the incoming power is close to 220V at start of test.

For the long test of the other day, it took many hours spread over the day due to my testing technique where I bring up the heat-pump load near to a test point & stabilise at the point for a period, before taking readings.

The typical power cycle over a day follows a sine wave, with lows in morning & evening.

The local power supplier is aware of this & attributes it to a critically-loaded transformer. The advice was - "Why don't you look for another house in another area where we have better power?"... LOL

My answer is - "No problem. I'll move back to my home country." Hence my planned move in around 4-5 months.

Is the compressor 50hz or 60hz? Is the local power supply 50hz or 60hz? Single phase or three phase? Does the compressor have multiple power ratings?

Compressor = 50Hz
Local power supply = 50Hz

Single phase
Single power rating for this compressor motor, as far as I know. I'm led to understand that the motor is changed to suit difference supply voltages & frequency, as required. A three-phase variant of this compressor is also available, but, frankly I've been loathe to go to 3-phase out of concern for phase balance problems - it's bad-enough on single phase.

[Gary]

As I'm sure you are well aware, the key to accurate testing is controlling the variables... and in this you seem to be currently at a disadvantage.

I suppose for the time being we will just have to play the cards we are dealt, and do the best we can under the circumstances.

[Gary]

Hmmm... maybe you can run the 40-55C tests again at a time of day when the voltage is down to about 190V to see if the performance is dropping as a result of the lower voltage, or is in fact attributable to the subcooling.

[desA]

As I'm sure you are well aware, the key to accurate testing is controlling the variables... and in this you seem to be currently at a disadvantage.

I suppose for the time being we will just have to play the cards we are dealt, and do the best we can under the circumstances.

What I did check was the power (W'=V*I) draw by the compressor (calculated) during the run & this is almost bang-on to the quoted power for the compressor at Te,sat=12.5'C, at each point across the test range.

So, it seems that the compressor input power is actually close to the supplier's estimates. Seemed odd at first, but I did double-check my calcs. I'll plot these separately & post the plot.

[desA]

Hmmm... maybe you can run the 40-55C tests again at a time of day when the voltage is down to about 190V to see if the performance is dropping as a result of the lower voltage, or is in fact attributable to the subcooling.

This is easy to do - I'll run these off in early peak power draw-off - probably tomorrow. Then we'll know for sure...

[desA]

http://tinypic.com/r/2zrkcq8/3

Here's the plot. The manufacturer's curve is the lumpy one...

W'c,in,m = manufacturer's compressor power estimate [kW]
W'c,in,c = calculated V*I power from experiment [kW]

[desA]

Oversized evaporator:

Today I tested the logic of the over-sized evaporator raising the Te,sat, by applying an insulation blank over around 40% of the evap inlet surface.

The Te,sat value dropped from 15'C (today) down to 12.5'C & settled there. I let the system stabilise for around 15 minutes & then removed the blank. Te,sat then gradually returned to 15'C.

I'm wondering if it would be worth doing a test with the evap blanked off to see the effect on the overall system performance, more out of a learning experience, than much else.

[desA]

I'd like to consolidate the control strategy so far. Please correct me, or add in the missing information.

Strategy review for riding compressor limits:

1. Lower limit : Tcomp,inlet <= 15'C
1.1 Limit suction pressure such that Tcomp,inlet < 15'C;
and/or
1.2 Control evap fan speed such that Tcomp,inlet < 15'C.

2. Lower limit : Tc,sat <= 75'C (or desired Tc,sat,op < 75'C)
2.1 Control condenser water flow such that Tc,sat < 75'C.

[Gary]

The upper limit for the suction pressure at the compressor inlet (CPR valve outlet) should be the saturation pressure corresponding to 15C.

The fan should be controlled by a temperature sensor on the discharge line (6 inches from the compressor). The ideal setting is yet to be determined. That setting would be the discharge temp at which Tc,sat is 75C, Te,sat is 15C, and superheat is 7K.

The water regulating valve should control the flow such that the Tc,sat is 75C.

[desA]

Super, Gary. Good, now we've got that settled. Thanks so much.

[desA]

A few more thoughts on the evap fan speed control.

Theoretical development
I've developed an algorithm linking the various factors that affect the Te,sat value in the evap. They are as follows:
1. Ta,in = air inlet temp;
2. U = overall heat-transfer coefficient;
3. A = heat-exchanger surface area;
4. m'a = air mass flowrate;
5. Q'e = evap heat load.

We can infer things like m'a, from a knowledge of Ta,in & Ta,out (air outlet temp) & Q'e (evap heat load).

We can also derive Q'e from the refrigerant side properties, if required, & entry/exit conditions, or we can work this out from other relationships.

How is the Tcomp,inlet - fan speed control issue managed in practice?

--------------
The main issue here is that the thermodynamic requirement for Q'e (evap) reduces from the start-up load, to the final hot load. This needs to be accommodated into a control strategy - if fine fan speed control is required.

--------------
Practical implementation
An alternative thought is to simply just use the Tcomp,inlet value as the marker & force the fan speed to reduce gradually so that Tcomp,inlet <= 15'C. How would this control logic be implemented in practice? The 'internal variables' can take care of themselves, with this approach, as long as the rate of change of fan speed is gentle.

[Gary]

The discharge temp is affected by high and low pressures, inlet superheat, heat of compression and motor heat. If we are to ride the upper limits of the compressor, the discharge temp is as good a way as any to do this.

[desA]

Mmmhh... that's an interesting point.

So, measure the compressor discharge temp, but control/adjust what parameters?

[desA]

Practical implementation
An alternative thought is to simply just use the Tcomp,inlet value as the marker & force the fan speed to reduce gradually so that Tcomp,inlet <= 15'C. How would this control logic be implemented in practice? The 'internal variables' can take care of themselves, with this approach, as long as the rate of change of fan speed is gentle.

What I had in mind here is perhaps some sort of simple PLC logic, which tests Tcomp,inlet. If Tcomp,in > 15'C, shut down fan one notch, if Tcomp,in < 15'C speed up one notch.

Is this more easily done in a simpler way, through standard components, rather than a PLC?

[Gary]

By slowing the fan we can keep from exceeding our ideal discharge temp.

[Gary]

What I had in mind here is perhaps some sort of simple PLC logic, which tests Tcomp,inlet. If Tcomp,in > 15'C, shut down fan one notch, if Tcomp,in < 15'C speed up one notch.

That's pretty much the same logic we would use for discharge temp control.

[desA]

By slowing the fan we can keep from exceeding our ideal discharge temp.

Good point.

Now is there any downside under cold inlet air conditions that we should take care of?

[desA]

That's pretty much the same logic we would use for discharge temp control.

Ok, great.

What is the least expensive way to implement this kind of logic, to control fan speed?

For instance, there are various fan-speed controllers on the market - some of which are programmable, with simple logic.

[Gary]

Ok, great.

What is the least expensive way to implement this kind of logic, to control fan speed?

For instance, there are various fan-speed controllers on the market - some of which are programmable, with simple logic.

I have no idea what anything costs.

[desA]

^

Perhaps : What would be the simplest way to implement the logic, that would not require much, if any, use of electronics?

[Gary]

^

Perhaps : What would be the simplest way to implement the logic, that would not require much, if any, use of electronics?

The simplest would be a thermostat which would turn off the fan at discharge line setpoint and then turn it back on as the temp drops.

Or perhaps a two speed fan motor could be used. The thermostat would kick it down to low speed at setpoint and then back to high speed as the temp drops.

[desA]

The simplest would be a thermostat which would turn off the fan at discharge line setpoint and then turn it back on as the temp drops.

Or perhaps a two speed fan motor could be used. The thermostat would kick it down to low speed at setpoint and then back to high speed as the temp drops.

Ok, this is pretty straightforward to implement. Thanks for that - keeps it simple.

I've seen so many electronic control malfunctions over the years, that I like to keep to as simple a system as possible.

[desA]

Can I ask you to re-visit the correct charge selection for a heat-pump.

For instance, say the Tc,sat=35'C charge requirement was 1220g, & the Tc,sat=70'C requirement was 1015g.

The previous discussions seems to prefer loading to the 1015g charge loading option rather than the 1220g option. Why select the lowest value & perhaps not some value part way up the range?

I can understand the 1220g start-up loading causing excessive condenser floodback & sub-cooling.

Would an option be to charge 1015g + add additional gas until say sub-cooling = 7K (or even 8.5K), at Tc,sat=70'C?

This could mean that the sub-cooling at start-up could be quite low. Would this present a problem, at all?

[Gary]

Ok, this is pretty straightforward to implement. Thanks for that - keeps it simple.

I've seen so many electronic control malfunctions over the years, that I like to keep to as simple a system as possible.

As an experimental system, I would go mechanical as much as possible, then when it is perfected consider going electronic.

Hmmm... When the "Cap and Trade" legislation goes through, energy prices will go right through the roof here in the ObamaNation. Maybe I should build one of these systems and put it up in my attic.