How to increase pool heater capacity?

[desA]

If the heat transfer remained the same then the high side pressure would remain the same.

The governing equation for a condenser seems to show that under same heat-transfer condition, as stated above, that the Tsat of the condenser will actually drop slightly, when flow is reduced & Tw,out increased.

The system floats on the inlet temperature of the external cooling medium. This also presumes no fundamental heat-transfer change in that process.

I'd be very interested in why the high side pressure should remain the same, if the Tsat for equilibrium is to reduce slightly - I'd expect it, under equilibrium conditions, to reduce slightly in line with the reducing Tsat. This is in the order of a few degrees celsius, but it is predicted.

[Gary]

When an actual (accurate) measurement shows one result and a calculation shows another result, there is a flaw in the calculation.

A measurement beats a calculation every time.

[Yuri B.]

Drew might have achieved greater "q" due to decrease of "m'v" being less than increase of "dTw".

[Yuri B.]

Dew might have achived greater "q" due to decrease of "m'w" being less than increase of "dTw".

[Gary]

Drew might have achieved greater "q" due to decrease of "m'v" being less than increase of "dTw".

Since q in fact decreased, as evidenced by the increase in high side pressure, I'm thinking that you have this backwards, but you are on the right track.

What we are missing here is the measured values of mw. A decrease in q tells us that the value of mw decreased faster than the value of dTw increased.

[Gary]

The governing equation for a condenser seems to show that under same heat-transfer condition, as stated above, that the Tsat of the condenser will actually drop slightly, when flow is reduced & Tw,out increased.

The system floats on the inlet temperature of the external cooling medium. This also presumes no fundamental heat-transfer change in that process.

I'd be very interested in why the high side pressure should remain the same, if the Tsat for equilibrium is to reduce slightly - I'd expect it, under equilibrium conditions, to reduce slightly in line with the reducing Tsat. This is in the order of a few degrees celsius, but it is predicted.

The flaw in all this is the assumption of fixed heat transfer. The heat transfer is not fixed.

[Gary]

At this point I would strongly suspect that the system is overcharged, with excess liquid refrigerant backing up into the condenser, limiting it's ability to transfer heat. A simple subcooling measurement would confirm or deny this.

[desA]

The flaw in all this is the assumption of fixed heat transfer. The heat transfer is not fixed.

Thanks for that. I'm very interested in understanding why the heat-transfer should be variable, & the high-pressure remain constant.

The equation, of course, also allows for a changing heat-transfer. I'm interested in the physics/thermodynamics & experimental observations for the reasoning behind this.

I bow to your many years of experience in this - no contest - & would be interested in why this occurs.

I've resolved the governing equation in terms of heat-transfer as a function of the remaining variables. The water mass flowrate appears in two places:
1. Linear increasing term;
2. (1-exp(-a)) term;

This shows a competition between two scenarios for heat-transfer as a function of water mass-flow. I'll plot it out.

This has got me intrigued. I'll scurry off to my laboratory heat-pump & test out the real system response over a number of scenarios. It's actually an age-old chestnut & I'd like to get completely to the bottom of it.

[Gary]

Thanks for that. I'm very interested in understanding why the heat-transfer should be variable, & the high-pressure remain constant.

Possibly I was not clear in this. You were saying that the heat transfer must remain constant... and I was responding that if the heat transfer were constant, then the high side pressure would be constant. In fact, neither is constant.

[desA]

Possibly I was not clear in this. You were saying that the heat transfer must remain constant... and I was responding that if the heat transfer were constant, then the high side pressure would be constant. In fact, neither is constant.

Fair enough.

Can I press a little further in, if I may be so bold...
(1) Why is neither constant?
(2) How are heat-transfer & high-side pressure related?

In a system, some component interaction around the refrigeration loop can sometimes showcase some very interesting effects.

[Gary]

The pressure of refrigerant is directly related to its saturation temperature. As the saturation temp rises, the pressure rises. As the saturation temp falls, the pressure falls.

The purpose of the condenser is to transfer heat from the refrigerant to the water. If heat is transferred faster (increased q), the refrigerant is cooled. In other words its saturation temperature falls, as does its pressure.

If the rate of heat transfer is less (reduced q) the refrigerant retains more heat and its saturation temp rises, as does its pressure.

In this case the suppliers decreased the water flow. As a result, the pressure increased, therefore the saturation temp increased, which tells us there is less heat transfer (lower q).

If you factor lowered q into your formula, you will see that water flow and delta-T cannot be an equal trade-off. In order for q to be lower, the decrease in water flow must exceed the increase in dT.

Decreasing the water flow was counter-productive.

[desA]

^ Thanks very much for your explanation, Gary.

[Drew]

Yes , the fact that the high side pressure increases says that lowering the water flow decreases the efficiency of the heat exchanger forcing the compressor to pump to a higher pressure to condense . With the lowered refrigerant flow rate the compressor and system balance at a new point with adjusted THR and system performance.

And visa versa, increasing flow rate would create a more efficient heat exchanger and would lower the head pressure increasing capacity.

[Drew]

This has got me intrigued. I'll scurry off to my laboratory heat-pump & test out the real system response over a number of scenarios.

ha ha thats funny. I can just picture it!

[Drew]

Heat absorbed = Heat transfered + heat of compression = heat rejected

If we alter the 'heat transfered + heat of compression' part of the formula (compressor capacity), by altering the operating pressures the heat 'in' would still equal the heat 'out'.

The compressor will deliver a certain capacity based on the conditions that are created by the 2 heat exchangers. As the oncoil temp increases, the head pressure rises. As the head pressure rises the compressors capacity decreases.As the compressors capacity decreases there is less energy to reject . The compressor and condenser balance at a new pressure which equates to the capacity.The three components individually now have this capacity which becomes the changed system capacity. The system is dynamic.

Again

Heat absorbed (kw , evap)= heat transered + heat of compression (kw ,compressor) = heat rejected (kw, cond)

[Drew]

In this case the suppliers decreased the water flow. As a result, the pressure increased, therefore the saturation temp increased, which tells us there is less heat transfer (lower q).


Decreasing the water flow was counter-productive.

Decreasing the flow rate was counter productive because it forced the compressor to a higher SCT and at this higher point the compressor has less capacity so less heat has to be rejected by the condenser. This is why there is less heat transfer?

[desA]

What seems to come from this is the following:
- decrease water flowrate;
- water outlet temp rises;
- condenser saturation temp required reduces.
- for the SAME heat-transfer rate (kW)

I do apologise - been re-reading my typing here (had a niggling thought today, thought I'd check) & checking my simulation - a typing error had crept in. It should have read as follows:
- decrease water flowrate;
- water outlet temp rises;
- condenser saturation temp required increases.
- for the SAME heat-transfer rate (kW)

This would now seem to agree with what followed my original comment. Sorry about the confusion.

Following on the Tc,sat change should be an appropriate Pc,sat change.

For small flow-rate changes, the change in equilibrium Tc,sat is not too much, but at very low flows, the Tc,sat can rise substantially. I've actually seen this effect in my lab heat-pump, when the water flow is backed off hard, or stopped. The pump then trips on the high-pressure safety switch fairly promptly.

Practically, as Gary stated, the actual heat-transfer in the condenser most probably doesn't remain completely constant, but the analysis offers useful information nevertheless.

[Drew]

- for the SAME heat-transfer rate (kW)

I might be loosing the plot here, but just to make sure I understand: If the head pressure climbs the compressor delivers less energy, the heat exchanger will adjust to the new capacity and therefore exchange less kw?

[desA]

I might be loosing the plot here, but just to make sure I understand: If the head pressure climbs the compressor delivers less energy, the heat exchanger will adjust to the new capacity and therefore exchange less kw?

Will the compressor deliver less energy, or reset itself to a higher energy level, with consequent higher delivery to the condenser?

If this is correct, then surely the condenser output would rise & the heat-pump folks would have been correct all along?

[Drew]

If you look at a compressor capacity chart you will see that the higher the saturated condensing temp the less the capacity . The lower the suction the less the capacity. To obtain the best capacity from a compressor we keep the suction as high as possible and the head as low. These pressures are governed by ambients and design product temp. Within reason we try to stick within these parameters. Its also a matter of cost and rules of thumb.
The reason that the capacity is reduced with high head pressure and low suction is due to the reduced refrigerant flow rate with these pressures. With a reciprocating compressor the piston must never hit the valve plate so we have a gap (clearance volume) as a saftey. On the piston downstroke the refrigerant in the clearance, which is actually the same as the head pressure,expands before new refrigerant can be drawn in. If the pressure is high a lot of piston downstroke is wasted before drawing in new refrigerant. If the suction is low it is even worse. Not only is a lot of the downstroke wasted overcoming the high head pressure , but the suction valve will open even later. These issues reduce refrigerant flow through the system. The lower the flow rate the lower the capacity.
Ive read that on average 1kg/hr of R22 will absorb about 50w of heat at the evaporator. If we reduce this flow rate the capacity will decrease.

[desA]

If you look at a compressor capacity chart you will see that the higher the saturated condensing temp the less the capacity . The lower the suction the less the capacity. To obtain the best capacity from a compressor we keep the suction as high as possible and the head as low. These pressures are governed by ambients and design product temp. Within reason we try to stick within these parameters. Its also a matter of cost and rules of thumb.

For condenser heat output, at increased hp temp/pressure, you're correct in that the total heat output should reduce slightly. This is true. So, it should reduce then... fair-enough, on that logic.

Sidebar:
The compressor input is predicted to rise, though - not fall with rising Tsat,cond (at same Tsat,evap).

For a heat-pump, as the Tsat,cond rises, the percentage of the total condenser output, moves from being evaporator-dominated, to compressor-dominated. Take a Copeland chart & try this. Simulators predict the same thing.

Downside:
The problem with only looking at one aspect of the system like this is that we've forgotten about the other parts. Once the condenser stabilises, does the evaporator re-adjust itself to a new equilibrium?

Practical note:
One of the problems with adjusting water flow in such a way as to raise Tsat,cond, is that it becomes a problem to reach water temps of more than 60'C, since this can push compressors above their recommended safe operating range (65'C shown in Copeland US data-sheets, 75'C in European data sheets). For R-134a it becomes a real balancing act up at that temp range, I can tell you. I try, if possible, to keep as low an approach within the condenser, as possible, to compensate for this problem.

In a pool-heater this is generally not an issue, as you still have ample temp margin available - but, for an AWHP, this can become an issue.

[Gary]

Heat absorbed = Heat transfered + heat of compression = heat rejected

Using your formula:

If the heat rejected is less, then the heat absorbed must be less and the heat transferred must be less.

Which of these factors is the controlling factor? None of them and all of them. A change in any part of the system is a change in every part of the system.

[desA]

Well, this has indeed been a fascinating thread. I've certainly learned a great deal.

[Gary]

In order to properly evaluate a system, we need to view it as a heat transfer chain, with the goal being to identify and strengthen the weakest link.

For example: If we reduce the airflow through an evaporator, the slower moving air has more time to cool as it passes through the coil, thus the delta-T increases. At the same time, the SST has decreased, verifying that less heat is being transferred due to the reduced air volume.

Assuming a fixed speed compressor, this tells us that an increase in delta-T must indicate reduced airflow and that reduced airflow = reduced heat transfer. In other words, high delta-T is an indicator of airflow problems.

There are similar indicators for each link in our heat transfer chain.

[Drew]

Sidebar:
The compressor input is predicted to rise, though - not fall with rising Tsat,cond (at same Tsat,evap).

the rise in input doesnt out do the loss in refrigeration effect and the result is a loss.

[Drew]

In order to properly evaluate a system, we need to view it as a heat transfer chain, with the goal being to identify and strengthen the weakest link.

Well put.

It just takes one undersized component to drag the system down to a less effeicient capacity.

[Gary]

Sidebar:
The compressor input is predicted to rise, though - not fall with rising Tsat,cond (at same Tsat,evap).

In fact, rising Tsat,cond is generally accompanied by rising Tsat,evap... yet we know there is a drop in capacity. The loss must therefore be in refrigerant mass flow.

[desA]

Ran up a test on the air-to-water heat-pump in my lab today.

I let the system stabilise at a full-flow condition on the water side, for around 30 minutes & recorded system readings. Then reduced the water flowrate substantially, let system stabilise to same tank water temp (allowed 45 minutes) & took new readings.

The trends observed were at follows:
1. Tsat,con rose from 48'C to 53.8'C on gauge (+5.8'C);
2. Psat,con rose from 1150 kPa to 1350 kPa on gauge (+200kPa);
3. Tsat,evap rose 0.5'C (within experimental error)(on gauge);
4. Compressor amperage rose 12.2% (actual recorded, marginally different from Copeland data sheet);
5. dT,water rose from 1'C to 5.7'C (poor sensor placement - indicator only)
6. First test : Tw,in=38.0'C ; Tw,out=39.0'C ; Tc,sup=60.2'C; Tc,exit=43.6'C ; Tw,tank=37.1'C (sensor position & mixing); Tevap,sup=23.4'C;
7. Second test : Tw,in=38.8'C ; Tw,out=44.5'C ; Tc,sup=65.8'C ; Tc,exit=48.7'C ; Tw,tank=38.2'C (sensor position & mixing); Tevap,sup=22.6'C;

When calculations & curve-fits are performed against the Copeland Europe predictions in 'Select 7.1', the following information emerges:
a. Condenser output load reduces (-3.06%);
b. Compressor input power increases (+14.522%);
c. COP,hp reduces substantially (-15.355%)

Air on/off evaporator coil remained similar in both tests (performed in same test run). The slight difference in recorded compressor power increase & Copeland predicted values is attributed to a system refrigerant charge still slightly under the optimum conditions. Optimum charge determination is still in progress for this heat-pump.

[Gary]

The approach temp seems to have remained relatively stable (Tsat and water out temp rising in unison), as I would expect, but not knowing the actual Tsat and leaving water temp I am wondering what your approach temp is.

[desA]

The trends observed were at follows:
1. Tsat,con rose from 48'C to 53.8'C on gauge (+5.8'C);
2. Psat,con rose from 1150 kPa to 1350 kPa on gauge (+200kPa);
3. Tsat,evap rose 0.5'C (within experimental error)(on gauge);
4. Compressor amperage rose 12.2% (actual recorded, marginally different from Copeland data sheet);
5. dT,water rose from 1'C to 5.7'C (poor sensor placement - indicator only)
6. First test : Tw,in=38.0'C ; Tw,out=39.0'C ; Tc,sup=60.2'C; Tc,exit=43.6'C ; Tw,tank=37.1'C (sensor position & mixing); Tevap,sup=23.4'C;
7. Second test : Tw,in=38.8'C ; Tw,out=44.5'C ; Tc,sup=65.8'C ; Tc,exit=48.7'C ; Tw,tank=38.2'C (sensor position & mixing); Tevap,sup=22.6'C;

There we are, Gary - I've added in the missing data you need.

I'll add in a few more items, if you'd like to crack it open - I'd love to hear your thoughts.

[Gary]

The approach was 9K in the first run and 9.3K in the second run. It should be no more than 11K, so you are getting good heat transfer. Contrast this to the 17K approach in Drew's pool heater and you see how very sick the pool heater is.

Note that adjusting the water flow has very little effect upon the approach temp. It is an entirely different issue.

[Gary]

There we are, Gary - I've added in the missing data you need.

I'll add in a few more items, if you'd like to crack it open - I'd love to hear your thoughts.

First we would need a better description of your system. Is it a cap tube or TXV system? Is it designed for comfort cooling and/or heating or for water heating?

Any further information you can provide about the design/purpose might prove to be helpful.

[Gary]

There we are, Gary - I've added in the missing data you need.

I'll add in a few more items, if you'd like to crack it open - I'd love to hear your thoughts.

Also, if we are to evaluate your system, you should start a new thread for that purpose.

[Drew]

As Gary said ealier a raised head pressure would normally cause an increase in evap pressure due to the reduced refrigerant flow. If they did increase proportionally the capacity loss according to compressor capacity charts wouldnt be much. But as we see with DesA 's experiment under reduced flow rate which forced the hp up the suction didnt climb as much. Therefore kW output would be reduced?
With the sick pool heater that I was working on it eventually had a Hp of 52C and Suction of -5C (due to low ambients)

[Drew]

If you knew the flow rates of the water we could see the two different Kw outputs of your system?
This would explain a lot.

kw = l/s x 4.19 x temp diff.

If we could work out the refrigerant flow rate we could see this transfer taking place through the chain.

We could also follow the heat tranfer chain to the evap where the heat gained in the refrigerant will equal the heat transfered through the heat excahanger and will equal the heat lost from the air.

[Gary]

As Gary said ealier a raised head pressure would normally cause an increase in evap pressure due to the reduced refrigerant flow.

Hmmm... on second thought this would be a result of warmer liquid entering the coil and the refrigerant mass flow should actually increase. Part of the heat is simply recirculating.

[desA]

Excellent feedback.

The exit approach temp is a very interesting indicator. Thanks for that, Gary.

I'll sort out the water flow-meter in the next few days, so that more accurate water measurements are available (beats bucket & timer).

As Gary suggested, I'll then start a new thread so that we can dissect the system in more detail.

[desA]

With the sick pool heater that I was working on it eventually had a Hp of 52C and Suction of -5C (due to low ambients)

Ahah... -5'C suction temp.

What refrigerant was it using? R-22, or R-134a?

Added later:

Currentley the HP is 47 degreesC , LP is -2 degreesC, 3 degreesC across heat exchanger. Ambient is 10 DegreesC.

I missed that in the earlier post. With an ambient like that, you are on a hiding to nothing, with an air-to-water pool heat-pump. If the ambient goes any lower, the defrost mechanism will probably begin to kick in - if it's equipped with one (some aren't in Australia).

[Magoo]

Drew . Sell the owner a spar pool and throw all the heat into it, you are on a mission to no were with the swimming pool. Given the reduced heat pick up from lower ambient, and losses from pool overnight.
Do the sums, volume of water in a ground pool, loses to ground, add losses to atmosphere even with a cover, thermal separation temps. Add reduced h/p capacity with low ambient.
everyone will theorise till the cows come home, but nothing will change.
magoo

[Drew]

Hey Magoo

Im over the pool heater. It has been sorted out. Not the way I would have liked , but anyway. This thread has gone beyond that. Im enjoying the theory.

[Drew]

Its running on R22. The low suction is a reflection of the ambient at the moment. When the suction lifts in summer our head will also climb. Hopefully not to high. Probably have to lift the water flow rate then to lower the hp.

[Drew]

Ahah... -5'C suction temp.

What refrigerant was it using? R-22, or R-134a?

Added later:


I missed that in the earlier post. With an ambient like that, you are on a hiding to nothing, with an air-to-water pool heat-pump. If the ambient goes any lower, the defrost mechanism will probably begin to kick in - if it's equipped with one (some aren't in Australia).

It was going into defrost often. In tasmania we will battle to maintain the suction above zero. That has to be taken into account in the heat load. Their is no doubt that their is an issue with the pool heater. My whole issue has been with the fact that the suppliers throttled down the water flow thinking that they increased the capacity. That is why i look forward to seeing your water flow figures so i can appply the formula.

[desA]

It was going into defrost often. In tasmania we will battle to maintain the suction above zero. That has to be taken into account in the heat load. Their is no doubt that their is an issue with the pool heater.

A few thoughts - why not preheat the air stream into the evaporator, so that it never sees a very cold ambient & does not need to go into defrost mode?

You could pull off some alternative warm air-stream for this.

In Australia, some folks actually put air-heater elements in front of the evaporator. for winter use. Sounds odd, but seems to be used.

-----------

Heat-transfer computation (local method):
For the water temp values, I reserve judgment on their accuracy. Current probe placement is not trustworthy at this point, in my view, & will probably throw you a few curved balls when you try & compute the instant condenser output. I NEVER trust local water probes (type K has accuracy around 0.4'C at very best, & closer to 1-2'C with poor placement), for heat-transfer calculations. Instead, I measure the end user of the energy e.g. water storage tank (250L), or swimming pool, over time & then compute the average heat absorbed by the water.

[Drew]

Nice logic with regard to applying that formula. How practically do you do it?

You would have to take the heat loss of the water container (pool) into account?

Whereas the other method of measuring the inlet and outlet of the heat exchanger would give you an instant reading of what the unit delivers without having to factor in other variables.? I do understand about the accuracy though.

[desA]

I always measure what the end user is actually doing.

An estimate can be worked out for the convective & evaporative cooling from the pool if uncovered, or conductive/convective from the pool cover.

This can then be worked backwards to determine what the heat-pump actually delivers.

This method turns out to be more accurate, in most cases than using your equation which is actually:

Q'w = m'w*Cpw*dTw = rw*V'w*Cpw*dTw

where:
rw = water density [kg/m3] ~ 1000 kg/m3
V'w = water volume flowrate [m3/s]
Cpw = water specific heat [J/kg.K] ~ 4186.8 J/kg.K
dTw = Tw,out - Tw,in ['C]

Now, what happens on a small dTw, with probe uncertainty of +- 1'C & Tw,in=35'C; Tw,out=36'C?

dTw = (36+1)-(35+1) = 37-36=1'C
dTw = (36+1)-(35-1) = 37-34=3'C
dTw = (36-1)-(35+1) = 35-36=-1'C
dTw = (36-1)-(35-1) = 35-34=1'C

You'll have a range of interesting heat-transfer values.

The problem gets less significant for larger values of dTw, but still represents a significant experimental uncertainty.

[Gary]

That is why i look forward to seeing your water flow figures so i can appply the formula.

I had the impression that you were no longer involved with or had access to the pool heater.

If you still have access then go over there and measure the subcooling (SCT minus liquid line temp). If the subcooling is more than 15F/8.5K, then remove refrigerant until the subcooling is down to 15F/8.5K and let us know the results.

[desA]

Check system for noncondensables...

How do we check for incondensibles being present in the system?

What is standard practice to remove air, other than through evacuation? Is a high-point bleeder point used e.g. service valve located at highest point in system?

Are you able to check this, Drew?

[Drew]

Hey Gary.
I dont have access to the pool unless i do it in my own time. I was going to use the flow rates on the figues that where obtained during DesA 's experiment to prove my point. maybe in due course i could obtain the subcooling specs and this would explain a lot.

DesA I hear you about heater banks on the evap surface to maintain a higher suction. I cant see this as being as effective as fitting electric elements in the water stream?
Being a tech and not an engineer i pick up formulas as i go and from what i read, so forgive me if I dont know the correct terms.

Condesables can be proved to be in a system by turning it off for a while to cool to ambient temp. We assume that the refrigerant should be saturated in the system, which means that we have vapour and liquid present. We then can read off our guages to see whether the ambient temp lines up with our guage pressure/ temp. If there is air in the system we what have the sum of the refrigerant and the air pressure which would be higher.
We also use this test to confirm what refrigerant is in a cylinder. This has become a bit difficult with the new refrigerants, but it gives us an idea. The only way to remove air in a system is to reclaim and evacuate.

[desA]

DesA I hear you about heater banks on the evap surface to maintain a higher suction. I cant see this as being as effective as fitting electric elements in the water stream?

I understand that parts (?all?) of Australia were looking at legislation which make it illegal to use direct heating of the water stream - for energy-efficiency purposes. This forces the use of indirect heating mechanisms.

Under this scenario, the heating of the air-stream into the heat-pump could be considered as indirect heating.

Direct electrical heaters in water, for a swimming pool, do scare me somewhat - from a health & safety point-of-view.

[Drew]

Valid point. The ones that have been installed came as a kit from a pool shop. Not quite sure of the law in Tasmania with regard to that.