ASHP built-in pump control method

[siletto]

Hello again!

 

Sorry for posting a new thread, if it is inappropriate here, please move it to the right section.

It is a very technical question, and the more engineering-related heating forums still don't answer it.

 

I have a few weeks in the studying or planning of my own heat pump heating-cooling system, and as an engineer, I came across some questions that are basically unanswered throughout the net, or hardy reachable.

 

The question is that how heatpumps control their inbuilt circulators? I'm sure there are many solutions to this, but I seek reply from those that studied/installed more units.

 

AFAIK, most heat pumps use some kind of PWM controlled circulators now, at least the name-brand ones do.

Panasonic Aquarea uses fixed "power" percentage, that can be set from the menu, according to the manual.

LG ThermaV uses either fixed capacity,fixed flow, or fixed deltaT(5°C), or some "optimal" flow control method, it can be changed from the menu.

The cheaper chinese HPs that I came across does not have control change possibility, but they seem to use fixed-ish deltaT.

 

For those models that need external circulators, how would you set it?

 

 

What if you must use a buffer/LLH, because you have both fancoils, radiators and UFH, or separate wall/ceiling heat-cool and UFH, or multiple heat sources? The best available articles(like John Siegenthaler's) only describes systems, buffer connection schemes at their nominal performance. Even heating design engineers specify components based on max load. If sized well, at maximum capacity the primary flow is usually greater than secondary(given a 5°C dT HP and 5-7°C dT for UFH) the LLH works flawlessly.. But as a heat pump modulates, it usually reduces its pwm pump's speed, so primary flow becomes less(maintaining dT). The secondary pump, which is usually fixed speed, or constant height type,  pumps more water than the HP, so flow mixing occurs.

The same thing scould occur with antifreeze monobloc heat pumps with plate heat exchanger. It might be sized to have minusculeloss at nominal flowrates. The heat pump reduces flow, but the secondary pump is fixed, so dT on the secondary side should increase(??), because of the still high flow, but lower capacity. Or the forward temperature drops on the secondary side? Or it follows the primary side with a lag? Or it induces "mixing", just as well with a LLH?

 

IN my opinion, the  whole anti-buffer hunt is because of this flow modulation problem during part-load.

A fixed-deltaT type secondary pump would be the ideal, but only the most expensive Wilo and Grundfos pumps have that feature.

If you set a fixed capacity in the heat pump menu, does this means that the Design deltaT(5°C) increases to 15°C when the heat pump modulates to 1/3 capacity?

 

There is only one 2 minute long video about this topic on the HeatGeek YT channel, but no other source talks about this problem, but it would solve all the buffer-related problems nonetheless.

[JohnMo]

Multiple answers, because multiple heat pumps use different algorithms, so no set answer. Some heat pumps will try to target a fixed dT, modulate output and pump speed or just modulate output, while others do variable dT on a fixed speed or variable speed circulation pump and everything in-between.

 

Choose your heat pump - then ask specific questions?

 

If you want a buffer, just install as a two port, then different flow rates have less effect on system efficiency anyway. Heat pumps duty then becomes keep buffer temperature within limits, if house is drawing the same flow rate as heat pump delivery then the buffer is just by-passed anyway.

 

[JamesPa]

7 hours ago, siletto said:

I have a few weeks in the studying or planning of my own heat pump heating-cooling system, and as an engineer, I came across some questions that are basically unanswered throughout the net, or hardy reachable.

 

The question is that how heatpumps control their inbuilt circulators? I'm sure there are many solutions to this, but I seek reply from those that studied/installed more units.

As @johnmo says above its a good question, there are multiple answers (ie different heat pumps adopt differnt approaches), and flow rate isn't anyway that important (so long as its sufficient). 

 

In principle faster flow rate = lower deltaT across the emitters = higher average emitter temperature for any given leaving water temperature = higher COP.  However it also equals higher water pump power consumption, which is certainly significant in a very low loss house (say <4-6kW) but irrelevant in a high loss house.  Then there may be 'engineering' factors which we cant quantify, and there definitely is the very practical factor that the velocity of water through the pipes should be limited for reasons of noise and erosion (typically to somewhere between 1  and 2 m/s depending on material and location).

 

By the time you put all this into the mash you come to the conclusion that the rule of thumb to design for 5C DT across the emitters is a good starting point, but if you can do less without busting the water pump do so, and if you are forced to do a bit more (up to say 8C) don't sweat too much.  Unless of course you have a heat pump that insists on 'control'.

 

There are occasionally discussions on other forums about this, for example on openenergy monitor someone has done an analysis of water pump consumption VS COP and from that come out with an 'optimum' (which depends on various factors including house loss), and on the Renewable Energy Forum there is currently an individual that is claiming (without any theoretical justification and contrary to available empirical evidence) that heat pumps which maintain a fixed DT of 5 by modulating the water pump speed are somehow superior.  

 

So unless there is a good reason to the contrary, design for around 5C, if you end up doing better then do so.  Absolutely do NOT design in any system separation (3 or 4 port buffer, LLH or PHE) unless there is a very solid justification (rare in a domestic environment) and you have a well thought through control strategy to avoid mixing, because if you do you are almost certain to incur a COP penalty for ne good reason.  Basically KISS.

 

Edited March 20 by JamesPa

[siletto]

Thanks for the reply!

 

We plan to use the heat pump for both floor heating and cooling with floor and fancoils, so omitting the buffer completely might not be possible, that is why I seek an Optimum method using it. 

 

With a friend, on a test setup we came up with a pwm controlled 4 port buffered floor heating solution, but it uses a  condensing boiler. It should work with a HP as well, since it is only a pwm circulator, 2 temp sensors and an Arduino, but We've yet to try it. 

 

With a pwm secondary pump, One can preserve the buffering possibility in case of any unexpected heating anomaly, but eliminate the diluton of flow temperature.

 

I posted the question to hopefully better understand the possible pump concepts built into the machines. 

[JohnMo]

1 hour ago, siletto said:

With a friend, on a test setup we came up with a pwm controlled 4 port buffered floor heating solution, but it uses a  condensing boiler. It should work with a HP as well, since it is only a pwm circulator, 2 temp sensors and an Arduino, but We've yet to try it.

Why would even need to go that route with UFH. The floor is huge a buffer already. A half decent boiler has a PWM circulation pump built in, so will do it for you. Then just use a low hysterisis thermostat on a fixed flow temp, the boiler is happy with. If I set flow to 38 degs it will run for hours on end, with out stopping, direct to floor, no additional pumps, I just use a +0 -0.1 hysterisis thermostat to control, when average temp outside is below 5 degs for 6 hours. All other times the ASHP is available to run.

 

Heat pump does the same with UFH and fan coil using a +/-0.1 thermostat, that flow is set to 30 up to 33 degs.

[MikeSharp01]

All sounds like a candidate for a Model Based Systems Engineering (MBSE) with simulation (simulink or some such) approach. You have many knowns, you can build assumption ranges for the rest and play away to your hearts content.

[JamesPa]

11 hours ago, siletto said:

Thanks for the reply!

 

We plan to use the heat pump for both floor heating and cooling with floor and fancoils, so omitting the buffer completely might not be possible, that is why I seek an Optimum method using it. 

 

With a friend, on a test setup we came up with a pwm controlled 4 port buffered floor heating solution, but it uses a  condensing boiler. It should work with a HP as well, since it is only a pwm circulator, 2 temp sensors and an Arduino, but We've yet to try it. 

 

With a pwm secondary pump, One can preserve the buffering possibility in case of any unexpected heating anomaly, but eliminate the diluton of flow temperature.

 

I posted the question to hopefully better understand the possible pump concepts built into the machines. 

Why make it that complicated.  Connect the heat pump to the emitters, do a bit of balancing with manifold valves if necessary, get the WC curve as low as it can go whilst just heating your house, enjoy cheap low carbon heating.  If you buy a heat pump with decent controls no need for hacked control solutions either.

 

Seriously ashp systems dont have to be complicated and buffers etc (unless well set up which is rare) both cripple performance and make fault diagnosis difficult.  More to the point they are unnecessary in almost all cases.

 

What problem is the buffer purportedly solving in your case?

 

If you lack system volume for defrost add a 2 port volumiser.

Edited March 23 by JamesPa

[siletto]

12 hours ago, JamesPa said:

What problem is the buffer purportedly solving in your case?

 

If you lack system volume for defrost add a 2 port volumiser.

 

There is no problem with heating. The problem is that we don't want to cool the whole house all day, and using only a single fancoil/radiant panel at a time needs way less flow than any heatpump's minimum. 

[joth]

36 minutes ago, siletto said:

 

There is no problem with heating. The problem is that we don't want to cool the whole house all day, and using only a single fancoil/radiant panel at a time needs way less flow than any heatpump's minimum. 

This is what I found I needed too. Setup my system as two zones:

1 for ufh (no buffer) and

2 for a single FCU with a volumizer tank in series. 

The FCU services different rooms through the day (office or bedroom) switched via duct dampers. 

Edited March 23 by joth

[JamesPa]

1 hour ago, siletto said:

 

There is no problem with heating. The problem is that we don't want to cool the whole house all day, and using only a single fancoil/radiant panel at a time needs way less flow than any heatpump's minimum. 

Ok but does it really make sense to compromise the heating performance for occasional cooling?  

 

Is it flow that's the problem or cooling capacity.  If the latter it will presumably just cycle, but so what.  Some inefficiency but for a handful of days per year rather than throughout the heating season.  Fridges cycle, nobody gets concerned.

 

@joth seem to have a sensible suggestion which he has tried.

 

Adding a buffer (other than a 2 port volumiser) to your heating will almost certainly give you headaches and increase running costs, because you will almost certainly end up with mixing forcing you to increase the LWT and thus inevitably reducing COP.  Best avoided IMHO.

 

(Health warning - I'm not a heating engineer, just someone with a background in physics and engineering who has spent 2 years studying this as an amateur, in order to inform his own heat pump installation)

 

 

Edited March 23 by JamesPa

[JohnMo]

2 minutes ago, JamesPa said:

Ok but does it really make sense to compromise the heating performance for occasional cooling

Doesn't really make sense for cooling either. Heating and cooling are very similar, low and slow for heating gets best CoP, warm and slow gets best EER. Just use a flow temp about 16 degs for cooling, use floor as your buffer and let the fan coils tick away. Flow 7 dags at 30 outside gets an EER 3.7, while at 16 you are getting 5.5. Push the flow to 18 get an EER of 6.2. Getting close to half the running costs of 7 degs - zero condensation issues, heat pump ticks away, nice and happy.

 

Add some PV, it does it for free - or near enough.

[siletto]

12 hours ago, joth said:

This is what I found I needed too. Setup my system as two zones:

1 for ufh (no buffer) and

2 for a single FCU with a volumizer tank in series. 

The FCU services different rooms through the day (office or bedroom) switched via duct dampers. 

Thanks to all for the replies! 

 

The more I ask (and read) the more obvious it is for me that my ideas are not practically doable in a small-scale family home sized building, with a single heatpump. 

 

My initial plan was to use UFH for heating and use fancoils for heating at the same time, but the fancoils could be used for cooling(at 7 degs) alone, or combined with UFC(at 18deg). 

 

So that every system could be used for heating and cooling, separate or combined. I thought there are come clever hydraulics that solves this maintaining the advantages, but as You say, it cannot be done without normal buffer tanks, right? 

[JamesPa]

3 hours ago, siletto said:

Thanks to all for the replies! 

 

The more I ask (and read) the more obvious it is for me that my ideas are not practically doable in a small-scale family home sized building, with a single heatpump. 

 

My initial plan was to use UFH for heating and use fancoils for heating at the same time, but the fancoils could be used for cooling(at 7 degs) alone, or combined with UFC(at 18deg). 

 

So that every system could be used for heating and cooling, separate or combined. I thought there are come clever hydraulics that solves this maintaining the advantages, but as You say, it cannot be done without normal buffer tanks, right? 

Not sure what the problem is.  As I understand it you want to combine (for heating) fancoils and UFH.  Both will run at a low temperature so thats OK  No performance crippling mixing or buffer needed, just some valves so you can balance (not that you will need a lot of balancing if the fancoils have built in thermostats which modulate the fan - they will look after themselves!).  One water pump (the one in the heat pump) only unless house very large.

 

For cooling you want to run both fancoils and possibly UFH at at 7C - the problem is you will get condensation on all exposed pipework and the fancoils.  The fancoils probably have condensate drains which will need to be connected up, you may have to insulate and seal the exposed pipework.  Alternatively just run above the dewpoint for light cooling - some heat pump controllers handle this natively.    Still cant see why you need a buffer.  What function is it that you think is needed and that only a buffer can perform (I cant see one personally).

 

The only problem comes if you want to heat and cool at the same time.  You cant.

Edited March 24 by JamesPa

[JohnMo]

36 minutes ago, JamesPa said:

UFH at at 7C

Not something you want to do, or need to do, 16 to 18 is fine for UFH (cooling) If fan coils are sized for heating at low temperature they are huge capacity wise, so warmer temps for cooling is quite ok as well.

[siletto]

17 hours ago, JamesPa said:

Not sure what the problem is.  As I understand it you want to combine (for heating) fancoils and UFH.  Both will run at a low temperature so thats OK  No performance crippling mixing or buffer needed, just some valves so you can balance (not that you will need a lot of balancing if the fancoils have built in thermostats which modulate the fan - they will look after themselves!).  One water pump (the one in the heat pump) only unless house very large.

 

For cooling you want to run both fancoils and possibly UFH at at 7C - the problem is you will get condensation on all exposed pipework and the fancoils.  The fancoils probably have condensate drains which will need to be connected up, you may have to insulate and seal the exposed pipework.  Alternatively just run above the dewpoint for light cooling - some heat pump controllers handle this natively.    Still cant see why you need a buffer.  What function is it that you think is needed and that only a buffer can perform (I cant see one personally).

 

The only problem comes if you want to heat and cool at the same time.  You cant.

 

The fundamental approach was like supplying 7°C to the fancoils, and 16-18°C for the underfloor cooling. 

I get it what you say, omitting the 7°C side deletes the need for buffering. 

 

I am not sure if a fancoil gets any meaningful cooling power when fed with 18 water, but I get the idea! 

[JamesPa]

1 hour ago, siletto said:

The fundamental approach was like supplying 7°C to the fancoils, and 16-18°C for the underfloor cooling. 

Not sure you would want to do that, two different flow temperatures (heating or cooling) means you have to operate at the least efficient of the two (the colder for cooling, the warmer for heating) and mix down/up which sacrifices cop/eer.

 

I don't know how effective fancoils are at 16-18 TBH, hopefully someone will.  I don't get to find out until summer when I will try it on mine!  I can't see why they wouldn't be pretty good, after all you are forcing air over the fins pretty quickly.