Excellent! CO2 heat pumps are going to become increasingly important. Understanding ways to reduce return temps to as close to room temp as possible will be essential to make them work well.
Fantastic video, but I would really have liked a proper section relating to underfloor heating. When you retrofit a heat pump to underfloor heating you are really stuck with the original design, and you need to manipulate within that framework, and with the possible limitations that gives. We have a 1999 well insulated house of 157m2 build with under floor heating. We have just had a Panasonic J series only for the underfloor heating. Underfloor heating was designed for an on/off gas boiler running through thermostats in each room (all disconnected now to run all open) SCOP is way too low at around 2.5 now at -5 degree outside temperature. Last evening it ran 39C into the house and 36-37 out. Pump runs 14+ l/min. How do we increase the DT of the system?
In a typical small modern house with builder-fitted undersized radiators (to leave some room for furniture), you have to run a flow temp of about 80 deg C just to keep the place warm. You obviously need a flow temp rather higher than your preferred hot water temperature if you have stored hot water.
You can experiment by limiting the radiator flow temperature and running more constantly, rather than the on-off cycling. For much of the year, the actual reqired radiator temperature can be consideranly lower, especially in reasonably well insulated houses. The response time may not be as fast as a boiler, so the way its used may take a bit of adjusting to.
Oh that thing you are showing at 6:12 is a plate heat exhanger, not a condenser. This is used for example in your average Combi boiler to heat up domestic water for your taps, showers etc. Hot water from the boiler travels on one side, water from the mains travels on the other side. boiler water gives up heat which is transfered on to the mains water as it passes through the exchanger and it arrives at your tap hot. Condersers work on the same principle, they are just copper pipes with aluminum fins, and the refrigerant exchanges heat with the environment (the air around it). And in the case of a heat pump the refrigerant enters the condenser super cooled, warms up, then travels to a heat exchanger similar to the one you are showing and transfers heat to the heating circuit water
The item I show is exactly like the condenser used in almost all water-heating heat pumps. (search images 'heat pump plate condenser'). The copper-aluminium heat-exchangers are for air, and typically Evaporators in ASHPs.
@@RossKempOnYourMum01 Exactly. Refrigeration evaporators and condensers are almost identical apart from the pressure rating and hard-solder connections, and bigger.
Very good video. Do you have some source where these equations, that are behind this simulation, can be seen. I am trying to make a model in simulink for my house heating system (condensing gas boiler).
Hi John. Sorry, but another question. In your demo diagram, you show the flow going into the top of the radiator, and return out from the bottom. The pump is also shown on the flow side. Given that heat rises, wouldn't the flow and return be the other way around, and the pump in the return? Reason I ask is because I had a job years ago where I questioned which way round they should be. There was a refurb job where they had re-piped the heating coils (Frost and Re-heat coils) on an air handling unit in a hospital, and soon after commissioning, the frost coil froze and burst. Boilers were at ground floor level, and AHU's were many floors above, so a pumped system, not gravity fed. I have been trying to find the answer to this question ever since, but no one seems to know the correct answer as to flow / return - top or bottom. If you or any of your viewers have an opinion, I'd love to know. Thanks in advance, too all participants!
The old-school way of plumbing rads is like I show, in at the top, and as you say hot water will stay above cold water, so the temperature and flow is even as the water flows (pumped) down the radiator. The return should always be at the bottom (the coldest part), but if you find a thermal image of a normal radiator (both connections at bottom, you will see the hot water rising to the top in just one of the vertical channels, then the hot flows along the top, and down the area of the radiator. The old method of top entry can be 'same side' or 'opposite' side. ususlly depending on if its height-length ratio. Re the pump position, it makes very little difference where this is in the circuit. The pressure gauge and expansion vessel would be in the 'suction' side of the pump.
Hi John. I really liked your video. I would like to have some conversation with you on heat pumps and circulators pumps in general. Two questions l have: 1. usually in the UK compare to Europe, space is key and when converting from a gas boiler to a heat pump there is no space for fitting a buffer tank or low loss header (in my opinion) so if the secondary side of the plate is connected directly to the radiators network, would the circulator pump flowrate ( and consequently the head) would be larger than when you had a gas boiler when no new radiators or pipework would be changed which again what would happen in most cases? 2. For the same KW requirment for a house or building, would you say that it is almost always true that a higher flow rate would be required for std refrigerant and C02 compare to a condensing gas boiler (again with a scenario where emitters system stays the same)?
Sorry for the delay. Yes, you are right, often no room for buffers. A low loss header should take little room. Anyhow, i do like the simplicity of a direct connection as you describe. The flowrate for heat pumps should be higher, but heat pumps are usually considerably smaller than the boiler they replace, and run for far more hours (longer and lower). So the head pressure mau be similar. Persionally, I sometime look at existing radiators, and ask myself, what size heat pump would be happly with existing rads. Increasing house insulation, so as to reduce heat need, can e a good plan. I think I answered for 1) & 2) there. If you mean CO2 refrigerant, then the flow rate will probably be lower
Great John: can you say what the pressure drop that would be required to achieve the flow rate you indicate along a typical pipework set-up? My experience is that the pressure drop through typical long pipes with lots of 90 bends in them is far higher than the boiler internal pumps can provide. I use an ultrasonic flow tester to see the volume of water that is actually flowing. Too often I see a flow too low- not just for the radiators - but also for the heat exchanger in the boiler so it tends to shut down because of internal overheat. Typical cold water flow thru the mains supply is probably no more than 20 l/m at 10m of head; so 5 radiators with a 6 l/m requirement is going to need 30 l/m, that's sort of max for a typical central heating pump but it will only provide 4.5m of head. How much head does a typical system need for that 30l/m flow rate?
Its a little hard to be specific here, but generally, heat pumps are operated for longer hours, at a lower temperature, so the kW output is less, hence dt less. e.g. 2hr 'blast' of 20kW boiler at dt10 = 28lit/sec. Same heat as 4hrs of 10kW heat pump with dt5, also 28lit/min. So, often, the pipework can be good enough. Circ pumps are much lower energy now, and 8m head is not uncommon now. Not that you want to operate at the top end of the pump. I suggest you could play with my simulator to look at each radiator. I also have a pressure drop simulator, but you could also search 'online pressure drop' for better pressure drop tools. Ecoforest have quite high pressure pumps within them, and these automatically adjust themselves to achieve dt of 5. I think the maximum pressure is around 10m head, and the power consumption quite low. In the days of old circ pumps, the pump wattage was a problem... less so now! heatpumps.co.uk/technical/flow-rate-and-pressure-drop-simulator/
@@johncantor4056 hi John, thanks for the reply. It is not considered good practice to exceed 1.5m/s in waterpipe, maximum should be 2m/s. In 22mm pipe this equates to 24-36 litres per minute. You can see why just adding high pressure pumps soon exceed this number even if it can be achieved. It is better, I believe, to fit several low pressure pumps to the system so that each can overcome the back pressure of that part of the system. However, it does not overcome the problem of high flow rate required for low delta systems. Could be the customer needs bigger bore pipes
I have never found problems with velocities well over 2m/sec, apart from excess pumping power, but this is less of a problem now that motors are more energy-efficient. I haven't had noise problems in pipe, but have done in valves where the velocity could easily be a lot higher. Anyhow, to the pump pressure issue... I cannot see any difference between say one circ pump operating at 6m head, and two (at different points in the circuit) operating at 3m head each.
John, you briefly mentioned something about "a tendency for heat transfer to happen at the top". Could you say more about that and why? I think i understand but dont want to be guessing...
Hello Sam, I was trying to explain the situation where the flow-rate is low and hence the dt is large. The warm water entering the radiator rises to the top. The bottom of the radiator ends up being relatively cool. More heat is given off from the top warm bit of the radiator. If the flowrate were however very high, and the radiator quite even in surface temperature, then the radiant output is evenly spread, but now more heat is transferred to the air at the bottom of the radiator because the air is cooler here. Does that make sense?
@SamDuke474 Sam, the graphs are at the bottom of the page here heatpumps.co.uk/cop-estimator/ Thanks for your initial question. You are the first to highlight this badly explained bit, so thanks for that. The point that I am trying to make is that with a condensing refrigerant, there is a constant temperature over the heat-exchanger’s refrigerant passageway (relating to the vapour pressure), but with a low water flow-rate, there is a large temperature variation along the water passageway. So, if the water flow-rate is low (hence dt [flow-return] is high), then most heat-transfer happens where the entering water is relatively cool. Little condensing happens at the top where the water temperature is close to the temperure of the refrigerant. (Left section of graphs ‘B’ & ‘C’). As a separate point, the vapour at the condenser inlet is superheated. It is many degrees higher, but this is a relatively small proportion of the total energy, and this vapour drops to ‘condensing temperature’ after passing down somewhere around 5 to 10% of the condenser. Anyhow, the point is that a fairly high flowrate (small dt) is generally desirable for best usage of the condensers surface. However, the CO2 Transcritical heat pump is different, an no condensing actually happens in the ‘condenser’. The CO2 refrigerant inlet experiences a gradient, and leaves the ‘condenser’ much cooler. So, with this refrigerant, a slow flow rate, and large rise in water temperure is desirable.
@@johncantor4056 makes total sense thanks. I wasn't sure it was as you show in your graphs or whether the refridgerant would also get supercooled and effectively mean the 'condensing' would happen over a small distance, limiting opportunity for heat transfer and forcing slower refridgerant movement. This also makes the CO2 transcritical behaviour make sense too! As you say, it is effectively a high glide refrigerant. Do you know if natural refrigerants like propane have a similar glide? I had heard some reference to it... Looking at the phase-change and specific heat capacity charts of CO2 it seems you can actually dynamically target different dTs by changing the pressure of the CO2, and therefore the 'width' of the transcritical region. Along that note, I wondered whether in heat pumps the condensing temperature is actively controlled by injecting more/less refrigerant into the system? Thanks so much for the detailed response and explanations - really appreciate it - I haven't seen anyone have a serious go at explaining this. In fact I've seen people speculating dt of 5 is just some myth that's permeated the industry... If you ever wanted to work together on some of the above questions I would love to help out :)
Great video, thanks. Can we consider that COP and efficiency drop if delta T is bigger? My ASPH, installed 7 years ago by the previous owners, has a delta T of 8-10, while the HP manual says it to be between 4-7 degrees. I will check today but I believe the water pump runs at the lowest speed as I can't increase anymore the flow from flow meters. I am thinking now about the reason the installers ran the pump at a low speed and got a higher delta T. All logic speaks to increase the flow to decrease the delta T in the specs of 4-7C. The heating works and it has been working for 7 years now, but apparently, delta T is higher... The only reason for me might be that zone actuators can clamp the zones in certain cases and the only open loop to the bathroom will run very fast and create noise. Also, another reason is to lower the pump consumption.
All I can think of is that you could try a compromise... lower the dt a little and see if you can record an improvement on running cost. There will usually be a region (maybe 4-7 dt) were the performance difference is not great. Circ pump power will increase if you speed the pump (unless you can open some flow restrictions to increse the flow) However, this exta may not be much with the latest pumps, and can often be outweighed by the compressor running cost saving.
@@johncantor4056 Thanks for your answer. Everything started when I decided to install auto-balancing actuators, and I noticed I could not reach delta T of 7 degrees even if I opened all zones fully. I estimated a flow of 6 L/min through the manifold; that's why I think the pump runs at minimum speed, and I can run it faster. The difference between 1 and 2 speed levels is 55W, which makes 40 kWh for a month - not much. Also, with this low flow, the HP cycles a lot when there is low demand and weather outside is hot. Maybe with higher flow this would be better. But still I also thinks that I have to measure the difference.
@@nkitanov I have never tried auto-balencing valves.. I have never looked into them.. Have you adjusted them relating to the radiator size? I'm wondering if they could be an unecessary restriction?? Are some of them around the max setting? I should get one to try it.
@@johncantor4056 it's UFH, and the problem is that I do not know the lengths and specs, so tuning them blindly. That's why thinking of using auto balance valves, which have temp sensors of the flow and return flow, and by opening and closing the valves, they try to keep delta T of 7 degrees. Anyway, I checked the pump, and to my surprise, it's running at maximum speed. So, I won't be able to achieve better delta T without adding an extra pump and will run it like that. It's a bit strange that the pump even on max speed cannot reach the needed circulation.
@@nkitanov So, 7 degrees is quite a lot for UH. If say your flow temperature was 35C, then the return would be only 28C. this would be felt as quite a gradient in the room. (hot and cooler areas). Can you reduce the dt to 4 or 5 max? It sounds like the auto adjusters might be 'throttling' the flow in an attemp to get a dt of 7. In general, in a normal situation, at least one adjustable flow restrictor on the manifold would be fully open. Often almost all of them are fully open. It would only be the shorter loops that need restrictiong, (as you are trying to achieve).
Sinmple answer.. dt of 3 to 5 ? Its actually hard to be specific, and may depend on the pattern the floor pipes are laid, and also on the temperature its designed to run at. So, if the dt is too much, then the flow to the floor loop is significantly warmer than the return back, so the floor area at the start of the loop is notably warmer than the end of the loop.... the floor feels uneven, and may, or man not be warm where you want it. If the flow is 45 and return 40 (dt5) then uneven-ness may not be too much, But if the flow were only say 30, and return 25 (dt5), then the unevenness could be significant. That said, uneveness may be more acceptable when its fairly mild out. The other thing to consider is this... you would emit about the same heat (kW) from a system with flow 45C, return 35C (dt10), as you would with a floor with flow 42C and return 38C (dt4). (mean water temp is the same) The COP should be a little better with the lower flow temperature, even though the return temp is higher. Finally, a spiral pipe design wont have the one-side-of-floor cold problem. Personally, i like the simplicity of having no mixer, but some people would advocate having a mixer and pump so that you can have the floor dt closer than the HP dt. I dont think the dt at the heat pump and for the floor are actually so fussy
Dear mr Cantor, I am curious to know what happens when the power output becomes very low. As far as I understand this may be down to 15% of full power. Does the deltaT remain the same? and thus it is the flow that is proportionally lowered? What is the reason for the "preferred" 5°C value of deltaT? is it a good compromise for the design of heat exchanger and circulation pumps, or there is another reason? thanks
Good question. IF the flowrate is the same (fixed-speed circ pump), then the flow-return dt reduces in proportion to heat output. Not seen evidence of any units happily dropping output more than around 30% in meduim conditions (lower on coldest day though... when you dont want low output!). IF the pump can modulate, then it could 'track' a dt of say 5, but at minimum compressor speed, the flowrate might be too low, so I think there would be a minimum speed 'capping'. I dont see much of a problem with low dt when at low speed, but possibly, some algorythms assume a certain dt, so very low dt could mess up the compressor speed control... possibly... I dont know. Another thing to consider is that any heat meter can loose accuracy when dt gets less than say 2 degrees... some even stop recording!, so it can be hard to proberly record heat output at minimum output. dt of 5 is I think a compromise. not wasting circ power, not too noisy, and the flowrate when dt is 5 is about right. If dt more, then average radiator temperature is significantly lower than the flow temperature... not ideal.
Hi mate. We’re installing hybrid systems, with a gas boiler and heat pump combined. Do we size the radiators at 50 or 30 degree delta T ? Thanks. Martin.
When the heat pump is running, it will always be better with dt30 design (radiators at 50c). A problem I have seen with a hybrid was the radiators were quite small, and if TRV close rads off, it can cycle inefficiently. I will be fitting bigger radiators in the main living room simply to keep the heat pump running more efficiently
When you talk about the temperature of the radiator, you're actually referring to the mean temperature of the water inside the radiator, right? I mean, measuring the radiator surface with an IR thermometer won't give you the correct value, right? That will give you a reading below the real value, right?
Yes, I am talking the average temperaure, so if radiator inlet were say 36c and outlet is 30c, then the mean is near enough 33c. The difference in temperure between water inside and the paint on the surface is very small. If you do a rough calc of heat transfer through the 2mm of steel (or aluminium), its small enough to ignore. No doubt a lot of laminar flow inside the radiator, but its still a small temperature difference, particularly at low heat pump temperatures.
Excellent dissemination of heat pump information, welldone John.
Excellent!
CO2 heat pumps are going to become increasingly important.
Understanding ways to reduce return temps to as close to room temp as possible will be essential to make them work well.
Fantastic video, but I would really have liked a proper section relating to underfloor heating.
When you retrofit a heat pump to underfloor heating you are really stuck with the original design, and you need to manipulate within that framework, and with the possible limitations that gives.
We have a 1999 well insulated house of 157m2 build with under floor heating.
We have just had a Panasonic J series only for the underfloor heating. Underfloor heating was designed for an on/off gas boiler running through thermostats in each room (all disconnected now to run all open)
SCOP is way too low at around 2.5 now at -5 degree outside temperature. Last evening it ran 39C into the house and 36-37 out. Pump runs 14+ l/min.
How do we increase the DT of the system?
Many thanks Jock
In a typical small modern house with builder-fitted undersized radiators (to leave some room for furniture), you have to run a flow temp of about 80 deg C just to keep the place warm. You obviously need a flow temp rather higher than your preferred hot water temperature if you have stored hot water.
You can experiment by limiting the radiator flow temperature and running more constantly, rather than the on-off cycling. For much of the year, the actual reqired radiator temperature can be consideranly lower, especially in reasonably well insulated houses. The response time may not be as fast as a boiler, so the way its used may take a bit of adjusting to.
Many thanks 💗 really enjoyed your efforts.. subscribed
Thanks!
Oh that thing you are showing at 6:12 is a plate heat exhanger, not a condenser. This is used for example in your average Combi boiler to heat up domestic water for your taps, showers etc. Hot water from the boiler travels on one side, water from the mains travels on the other side. boiler water gives up heat which is transfered on to the mains water as it passes through the exchanger and it arrives at your tap hot. Condersers work on the same principle, they are just copper pipes with aluminum fins, and the refrigerant exchanges heat with the environment (the air around it). And in the case of a heat pump the refrigerant enters the condenser super cooled, warms up, then travels to a heat exchanger similar to the one you are showing and transfers heat to the heating circuit water
The item I show is exactly like the condenser used in almost all water-heating heat pumps. (search images 'heat pump plate condenser'). The copper-aluminium heat-exchangers are for air, and typically Evaporators in ASHPs.
@@johncantor4056 that is 100% a plate hex from a combi
@@RossKempOnYourMum01 Exactly. Refrigeration evaporators and condensers are almost identical apart from the pressure rating and hard-solder connections, and bigger.
Very good video. Do you have some source where these equations, that are behind this simulation, can be seen. I am trying to make a model in simulink for my house heating system (condensing gas boiler).
Hi John. Sorry, but another question. In your demo diagram, you show the flow going into the top of the radiator, and return out from the bottom. The pump is also shown on the flow side. Given that heat rises, wouldn't the flow and return be the other way around, and the pump in the return? Reason I ask is because I had a job years ago where I questioned which way round they should be. There was a refurb job where they had re-piped the heating coils (Frost and Re-heat coils) on an air handling unit in a hospital, and soon after commissioning, the frost coil froze and burst. Boilers were at ground floor level, and AHU's were many floors above, so a pumped system, not gravity fed. I have been trying to find the answer to this question ever since, but no one seems to know the correct answer as to flow / return - top or bottom. If you or any of your viewers have an opinion, I'd love to know. Thanks in advance, too all participants!
The old-school way of plumbing rads is like I show, in at the top, and as you say hot water will stay above cold water, so the temperature and flow is even as the water flows (pumped) down the radiator. The return should always be at the bottom (the coldest part), but if you find a thermal image of a normal radiator (both connections at bottom, you will see the hot water rising to the top in just one of the vertical channels, then the hot flows along the top, and down the area of the radiator. The old method of top entry can be 'same side' or 'opposite' side. ususlly depending on if its height-length ratio. Re the pump position, it makes very little difference where this is in the circuit. The pressure gauge and expansion vessel would be in the 'suction' side of the pump.
Hi John. I really liked your video. I would like to have some conversation with you on heat pumps and circulators pumps in general. Two questions l have:
1. usually in the UK compare to Europe, space is key and when converting from a gas boiler to a heat pump there is no space for fitting a buffer tank or low loss header (in my opinion) so if the secondary side of the plate is connected directly to the radiators network, would the circulator pump flowrate ( and consequently the head) would be larger than when you had a gas boiler when no new radiators or pipework would be changed which again what would happen in most cases?
2. For the same KW requirment for a house or building, would you say that it is almost always true that a higher flow rate would be required for std refrigerant and C02 compare to a condensing gas boiler (again with a scenario where emitters system stays the same)?
Sorry for the delay. Yes, you are right, often no room for buffers. A low loss header should take little room. Anyhow, i do like the simplicity of a direct connection as you describe. The flowrate for heat pumps should be higher, but heat pumps are usually considerably smaller than the boiler they replace, and run for far more hours (longer and lower). So the head pressure mau be similar. Persionally, I sometime look at existing radiators, and ask myself, what size heat pump would be happly with existing rads. Increasing house insulation, so as to reduce heat need, can e a good plan.
I think I answered for 1) & 2) there. If you mean CO2 refrigerant, then the flow rate will probably be lower
Great video
Great John: can you say what the pressure drop that would be required to achieve the flow rate you indicate along a typical pipework set-up?
My experience is that the pressure drop through typical long pipes with lots of 90 bends in them is far higher than the boiler internal pumps can provide.
I use an ultrasonic flow tester to see the volume of water that is actually flowing. Too often I see a flow too low- not just for the radiators - but also for the heat exchanger in the boiler so it tends to shut down because of internal overheat.
Typical cold water flow thru the mains supply is probably no more than 20 l/m at 10m of head; so 5 radiators with a 6 l/m requirement is going to need 30 l/m, that's sort of max for a typical central heating pump but it will only provide 4.5m of head. How much head does a typical system need for that 30l/m flow rate?
Its a little hard to be specific here, but generally, heat pumps are operated for longer hours, at a lower temperature, so the kW output is less, hence dt less. e.g. 2hr 'blast' of 20kW boiler at dt10 = 28lit/sec. Same heat as 4hrs of 10kW heat pump with dt5, also 28lit/min. So, often, the pipework can be good enough. Circ pumps are much lower energy now, and 8m head is not uncommon now. Not that you want to operate at the top end of the pump. I suggest you could play with my simulator to look at each radiator. I also have a pressure drop simulator, but you could also search 'online pressure drop' for better pressure drop tools. Ecoforest have quite high pressure pumps within them, and these automatically adjust themselves to achieve dt of 5. I think the maximum pressure is around 10m head, and the power consumption quite low. In the days of old circ pumps, the pump wattage was a problem... less so now! heatpumps.co.uk/technical/flow-rate-and-pressure-drop-simulator/
@@johncantor4056 hi John, thanks for the reply. It is not considered good practice to exceed 1.5m/s in waterpipe, maximum should be 2m/s.
In 22mm pipe this equates to 24-36 litres per minute.
You can see why just adding high pressure pumps soon exceed this number even if it can be achieved.
It is better, I believe, to fit several low pressure pumps to the system so that each can overcome the back pressure of that part of the system.
However, it does not overcome the problem of high flow rate required for low delta systems. Could be the customer needs bigger bore pipes
I have never found problems with velocities well over 2m/sec, apart from excess pumping power, but this is less of a problem now that motors are more energy-efficient. I haven't had noise problems in pipe, but have done in valves where the velocity could easily be a lot higher. Anyhow, to the pump pressure issue... I cannot see any difference between say one circ pump operating at 6m head, and two (at different points in the circuit) operating at 3m head each.
@@johncantor4056 Thanks for an excellent tool John, its very interesting.
John, you briefly mentioned something about "a tendency for heat transfer to happen at the top". Could you say more about that and why? I think i understand but dont want to be guessing...
Hello Sam, I was trying to explain the situation where the flow-rate is low and hence the dt is large. The warm water entering the radiator rises to the top. The bottom of the radiator ends up being relatively cool. More heat is given off from the top warm bit of the radiator. If the flowrate were however very high, and the radiator quite even in surface temperature, then the radiant output is evenly spread, but now more heat is transferred to the air at the bottom of the radiator because the air is cooler here. Does that make sense?
@@johncantor4056 oh sorry I meant when you were holding up the heat exchanger
@@SamDuke474 Ah.. OK. I did some graphs to help explain that... cannot find them!! I will try to seek them out. I need to tidy up my website!!
@SamDuke474 Sam, the graphs are at the bottom of the page here heatpumps.co.uk/cop-estimator/
Thanks for your initial question. You are the first to highlight this badly explained bit, so thanks for that.
The point that I am trying to make is that with a condensing refrigerant, there is a constant temperature over the heat-exchanger’s refrigerant passageway (relating to the vapour pressure), but with a low water flow-rate, there is a large temperature variation along the water passageway. So, if the water flow-rate is low (hence dt [flow-return] is high), then most heat-transfer happens where the entering water is relatively cool. Little condensing happens at the top where the water temperature is close to the temperure of the refrigerant. (Left section of graphs ‘B’ & ‘C’).
As a separate point, the vapour at the condenser inlet is superheated. It is many degrees higher, but this is a relatively small proportion of the total energy, and this vapour drops to ‘condensing temperature’ after passing down somewhere around 5 to 10% of the condenser.
Anyhow, the point is that a fairly high flowrate (small dt) is generally desirable for best usage of the condensers surface.
However, the CO2 Transcritical heat pump is different, an no condensing actually happens in the ‘condenser’. The CO2 refrigerant inlet experiences a gradient, and leaves the ‘condenser’ much cooler. So, with this refrigerant, a slow flow rate, and large rise in water temperure is desirable.
@@johncantor4056 makes total sense thanks. I wasn't sure it was as you show in your graphs or whether the refridgerant would also get supercooled and effectively mean the 'condensing' would happen over a small distance, limiting opportunity for heat transfer and forcing slower refridgerant movement. This also makes the CO2 transcritical behaviour make sense too! As you say, it is effectively a high glide refrigerant. Do you know if natural refrigerants like propane have a similar glide? I had heard some reference to it...
Looking at the phase-change and specific heat capacity charts of CO2 it seems you can actually dynamically target different dTs by changing the pressure of the CO2, and therefore the 'width' of the transcritical region.
Along that note, I wondered whether in heat pumps the condensing temperature is actively controlled by injecting more/less refrigerant into the system?
Thanks so much for the detailed response and explanations - really appreciate it - I haven't seen anyone have a serious go at explaining this. In fact I've seen people speculating dt of 5 is just some myth that's permeated the industry... If you ever wanted to work together on some of the above questions I would love to help out :)
Great video, thanks. Can we consider that COP and efficiency drop if delta T is bigger? My ASPH, installed 7 years ago by the previous owners, has a delta T of 8-10, while the HP manual says it to be between 4-7 degrees. I will check today but I believe the water pump runs at the lowest speed as I can't increase anymore the flow from flow meters. I am thinking now about the reason the installers ran the pump at a low speed and got a higher delta T. All logic speaks to increase the flow to decrease the delta T in the specs of 4-7C. The heating works and it has been working for 7 years now, but apparently, delta T is higher... The only reason for me might be that zone actuators can clamp the zones in certain cases and the only open loop to the bathroom will run very fast and create noise. Also, another reason is to lower the pump consumption.
All I can think of is that you could try a compromise... lower the dt a little and see if you can record an improvement on running cost. There will usually be a region (maybe 4-7 dt) were the performance difference is not great. Circ pump power will increase if you speed the pump (unless you can open some flow restrictions to increse the flow) However, this exta may not be much with the latest pumps, and can often be outweighed by the compressor running cost saving.
@@johncantor4056 Thanks for your answer. Everything started when I decided to install auto-balancing actuators, and I noticed I could not reach delta T of 7 degrees even if I opened all zones fully. I estimated a flow of 6 L/min through the manifold; that's why I think the pump runs at minimum speed, and I can run it faster. The difference between 1 and 2 speed levels is 55W, which makes 40 kWh for a month - not much. Also, with this low flow, the HP cycles a lot when there is low demand and weather outside is hot. Maybe with higher flow this would be better. But still I also thinks that I have to measure the difference.
@@nkitanov I have never tried auto-balencing valves.. I have never looked into them.. Have you adjusted them relating to the radiator size? I'm wondering if they could be an unecessary restriction?? Are some of them around the max setting? I should get one to try it.
@@johncantor4056 it's UFH, and the problem is that I do not know the lengths and specs, so tuning them blindly. That's why thinking of using auto balance valves, which have temp sensors of the flow and return flow, and by opening and closing the valves, they try to keep delta T of 7 degrees. Anyway, I checked the pump, and to my surprise, it's running at maximum speed. So, I won't be able to achieve better delta T without adding an extra pump and will run it like that. It's a bit strange that the pump even on max speed cannot reach the needed circulation.
@@nkitanov So, 7 degrees is quite a lot for UH. If say your flow temperature was 35C, then the return would be only 28C. this would be felt as quite a gradient in the room. (hot and cooler areas). Can you reduce the dt to 4 or 5 max? It sounds like the auto adjusters might be 'throttling' the flow in an attemp to get a dt of 7. In general, in a normal situation, at least one adjustable flow restrictor on the manifold would be fully open. Often almost all of them are fully open. It would only be the shorter loops that need restrictiong, (as you are trying to achieve).
Hi John great video, as a rule of thumb what delta T would you be looking for on a UFH loop with a heat pump?
Sinmple answer.. dt of 3 to 5 ? Its actually hard to be specific, and may depend on the pattern the floor pipes are laid, and also on the temperature its designed to run at. So, if the dt is too much, then the flow to the floor loop is significantly warmer than the return back, so the floor area at the start of the loop is notably warmer than the end of the loop.... the floor feels uneven, and may, or man not be warm where you want it. If the flow is 45 and return 40 (dt5) then uneven-ness may not be too much, But if the flow were only say 30, and return 25 (dt5), then the unevenness could be significant. That said, uneveness may be more acceptable when its fairly mild out. The other thing to consider is this... you would emit about the same heat (kW) from a system with flow 45C, return 35C (dt10), as you would with a floor with flow 42C and return 38C (dt4). (mean water temp is the same) The COP should be a little better with the lower flow temperature, even though the return temp is higher. Finally, a spiral pipe design wont have the one-side-of-floor cold problem.
Personally, i like the simplicity of having no mixer, but some people would advocate having a mixer and pump so that you can have the floor dt closer than the HP dt. I dont think the dt at the heat pump and for the floor are actually so fussy
Great video! Is that simulator available anywhere to download?
Nevermind, just found it on google!
Sorry, the simulator is only on my website
Dear mr Cantor, I am curious to know what happens when the power output becomes very low. As far as I understand this may be down to 15% of full power. Does the deltaT remain the same? and thus it is the flow that is proportionally lowered? What is the reason for the "preferred" 5°C value of deltaT? is it a good compromise for the design of heat exchanger and circulation pumps, or there is another reason? thanks
Good question. IF the flowrate is the same (fixed-speed circ pump), then the flow-return dt reduces in proportion to heat output. Not seen evidence of any units happily dropping output more than around 30% in meduim conditions (lower on coldest day though... when you dont want low output!). IF the pump can modulate, then it could 'track' a dt of say 5, but at minimum compressor speed, the flowrate might be too low, so I think there would be a minimum speed 'capping'. I dont see much of a problem with low dt when at low speed, but possibly, some algorythms assume a certain dt, so very low dt could mess up the compressor speed control... possibly... I dont know. Another thing to consider is that any heat meter can loose accuracy when dt gets less than say 2 degrees... some even stop recording!, so it can be hard to proberly record heat output at minimum output. dt of 5 is I think a compromise. not wasting circ power, not too noisy, and the flowrate when dt is 5 is about right. If dt more, then average radiator temperature is significantly lower than the flow temperature... not ideal.
It also for the glide of the refriger gas they use like R410, propane or R32, different gas different glide of condensation different delta T
@@Bushtuckerman71 can you please elaborate? possibly linking where to find the theory. thanks
Hi mate. We’re installing hybrid systems, with a gas boiler and heat pump combined. Do we size the radiators at 50 or 30 degree delta T ? Thanks. Martin.
When the heat pump is running, it will always be better with dt30 design (radiators at 50c). A problem I have seen with a hybrid was the radiators were quite small, and if TRV close rads off, it can cycle inefficiently. I will be fitting bigger radiators in the main living room simply to keep the heat pump running more efficiently
@@johncantor4056 thanks John 👍 have you an email I can contact you direct mate ?
When you talk about the temperature of the radiator, you're actually referring to the mean temperature of the water inside the radiator, right?
I mean, measuring the radiator surface with an IR thermometer won't give you the correct value, right?
That will give you a reading below the real value, right?
Yes, I am talking the average temperaure, so if radiator inlet were say 36c and outlet is 30c, then the mean is near enough 33c. The difference in temperure between water inside and the paint on the surface is very small. If you do a rough calc of heat transfer through the 2mm of steel (or aluminium), its small enough to ignore. No doubt a lot of laminar flow inside the radiator, but its still a small temperature difference, particularly at low heat pump temperatures.