ᐅ Hydraulic balancing for air-to-water heat pump + high-efficiency circulation pump
Created on: 3 Jan 2021 23:07
L
lesmue79
Warning: wall of text and lots of theorycrafting:
I am currently trying to optimize or fundamentally adjust the hydraulic and thermal balancing of my air-to-water heat pump system, including underfloor heating, but I am running into the following issues:
First, about the house: KfW-55 bungalow with controlled mechanical ventilation
Nearly 105 m2 (1130 ft²) of heated floor area
Air-to-water heat pump with underfloor heating throughout, 10cm (5 inches) pipe spacing, max 30°C (86°F) flow temperature. At -12°C (10°F) outside temperature, the calculated heating load is 3276 watts.
According to the datasheet, the heat pump delivers 3200 watts at -10°C (14°F) outside temperature with flow 35°C (95°F) and return 30°C (86°F).
All rooms are designed for 20°C (68°F), including the bathroom (to avoid an oversized heat pump by the general contractor). Additionally, for the bathroom, an electric radiator is planned to achieve a room temperature of 2°C (4°F) higher than the rest. However, in reality, the toilet, utility room, bedroom, and guest room should only be heated to 18°C (64°F) (it won’t be much lower in a new building). The bathroom is intended to be warmer, at around 21–22°C (70–72°F).
Currently, I have the following questions (though perhaps I am too focused on the self-regulation effect and avoiding actuator valves):
1. Circulation pump: Various guides, manuals, and forum posts recommend setting the circulation pump of the underfloor heating to a constant flow rate.
My conclusion: my circulation pump is a high-efficiency variable-speed pump, so I can set the flow rate on the manifold in L/min (based on the calculations from the general contractor / heating engineer) to whatever I want, but the flow always settles around 600–630 L/h (10–11 L/min). The only significant flow changes I get are when I activate the actuators and room thermostats, which then open or close the valves. The only adjustment parameter on the circulation pump is the minimum flow rate; no other settings are available. But I don’t fully understand how this function works.
2. Operating times of the heat pump / self-regulation: I usually read that the heat pump should run as long as possible, though some sources say short cycling a few times is normal.
My conclusion: if I run the system without actuators and room thermostats, the energy integral control does not work; the system basically runs almost 24/7 at low temperatures, with interruptions only for defrosting. As a result, with a flow temperature of 27°C (81°F), I only get about 19–20°C (66–68°F) room temperature, but I’d prefer around 21–23°C (70–73°F), especially in the bathroom. If I do it the other way, with energy integral control (EIC) and actuators and slightly higher curve so that 30°C (86°F) flow is demanded, the actuators close in the first rooms, which causes the flow to increase to the other rooms because the pump still distributes the volume flow among the remaining open valves. At the same time, the flow temperature rises for rooms where the actuators are still open until the desired temperature is reached and the actuators close. Then the energy integral kicks in and goes negative because actual flow temperature exceeds setpoint flow temperature, until the heat pump shuts off once the energy integral has been reduced.
So right now, I’m struggling with what is better: should the system just run steadily at a flow temperature of 27°C (81°F) (which I might still optimize a bit), with heating only interrupted for defrosting or when the compressor’s hysteresis is exceeded, causing the compressor to be locked out for a certain time? Or should I define time windows during which the system is allowed to operate?
Maybe I could manage this better by refining the balancing, but I guess I’ll have to throttle down so much for the energy integral to work that the flow rate will fall below the minimum required, and the bypass valve will open.
Or should I run the system at 30°C (86°F) flow with room thermostats and actuators, allowing the energy integral control to function properly and reach the desired room temperatures?
Another strange issue is: according to the heating load and underfloor heating calculations, the system requires about 840 L/h (14 L/min) at 4.4 K (7.9°F) delta T in the design case. If I set the flow according to this calculation or slightly lower, the pump only delivers 600–630 L/h (10–11 L/min) at a delta T of about 3–4 K (5.4–7.2°F).
According to the datasheet, the optimal flow rate for the heat pump is 540 L/h (9 L/min) at 5 K (9°F) delta T.
540 L/h * 5 K * 1.163 = 3132 watts
620 L/h * 3.5 K * 1.163 = 2527 watts
840 L/h * 4.4 K * 1.163 = 4287 watts
Calculated heating load at -12°C (10°F) = 3176 watts (and this heating load is probably overestimated since controlled mechanical ventilation was not included in the calculation, and I want only 15–18°C (59–64°F) in four rooms instead of the calculated 20°C (68°F). Also, average outside temperatures for the heat pump in my area are closer to -10°C (14°F) rather than -12°C (10°F), so there is some margin).
Maybe I have now gotten too caught up in theoretical and calculated values and can’t see the forest for the trees?
I am currently trying to optimize or fundamentally adjust the hydraulic and thermal balancing of my air-to-water heat pump system, including underfloor heating, but I am running into the following issues:
First, about the house: KfW-55 bungalow with controlled mechanical ventilation
Nearly 105 m2 (1130 ft²) of heated floor area
Air-to-water heat pump with underfloor heating throughout, 10cm (5 inches) pipe spacing, max 30°C (86°F) flow temperature. At -12°C (10°F) outside temperature, the calculated heating load is 3276 watts.
According to the datasheet, the heat pump delivers 3200 watts at -10°C (14°F) outside temperature with flow 35°C (95°F) and return 30°C (86°F).
All rooms are designed for 20°C (68°F), including the bathroom (to avoid an oversized heat pump by the general contractor). Additionally, for the bathroom, an electric radiator is planned to achieve a room temperature of 2°C (4°F) higher than the rest. However, in reality, the toilet, utility room, bedroom, and guest room should only be heated to 18°C (64°F) (it won’t be much lower in a new building). The bathroom is intended to be warmer, at around 21–22°C (70–72°F).
Currently, I have the following questions (though perhaps I am too focused on the self-regulation effect and avoiding actuator valves):
1. Circulation pump: Various guides, manuals, and forum posts recommend setting the circulation pump of the underfloor heating to a constant flow rate.
My conclusion: my circulation pump is a high-efficiency variable-speed pump, so I can set the flow rate on the manifold in L/min (based on the calculations from the general contractor / heating engineer) to whatever I want, but the flow always settles around 600–630 L/h (10–11 L/min). The only significant flow changes I get are when I activate the actuators and room thermostats, which then open or close the valves. The only adjustment parameter on the circulation pump is the minimum flow rate; no other settings are available. But I don’t fully understand how this function works.
2. Operating times of the heat pump / self-regulation: I usually read that the heat pump should run as long as possible, though some sources say short cycling a few times is normal.
My conclusion: if I run the system without actuators and room thermostats, the energy integral control does not work; the system basically runs almost 24/7 at low temperatures, with interruptions only for defrosting. As a result, with a flow temperature of 27°C (81°F), I only get about 19–20°C (66–68°F) room temperature, but I’d prefer around 21–23°C (70–73°F), especially in the bathroom. If I do it the other way, with energy integral control (EIC) and actuators and slightly higher curve so that 30°C (86°F) flow is demanded, the actuators close in the first rooms, which causes the flow to increase to the other rooms because the pump still distributes the volume flow among the remaining open valves. At the same time, the flow temperature rises for rooms where the actuators are still open until the desired temperature is reached and the actuators close. Then the energy integral kicks in and goes negative because actual flow temperature exceeds setpoint flow temperature, until the heat pump shuts off once the energy integral has been reduced.
So right now, I’m struggling with what is better: should the system just run steadily at a flow temperature of 27°C (81°F) (which I might still optimize a bit), with heating only interrupted for defrosting or when the compressor’s hysteresis is exceeded, causing the compressor to be locked out for a certain time? Or should I define time windows during which the system is allowed to operate?
Maybe I could manage this better by refining the balancing, but I guess I’ll have to throttle down so much for the energy integral to work that the flow rate will fall below the minimum required, and the bypass valve will open.
Or should I run the system at 30°C (86°F) flow with room thermostats and actuators, allowing the energy integral control to function properly and reach the desired room temperatures?
Another strange issue is: according to the heating load and underfloor heating calculations, the system requires about 840 L/h (14 L/min) at 4.4 K (7.9°F) delta T in the design case. If I set the flow according to this calculation or slightly lower, the pump only delivers 600–630 L/h (10–11 L/min) at a delta T of about 3–4 K (5.4–7.2°F).
According to the datasheet, the optimal flow rate for the heat pump is 540 L/h (9 L/min) at 5 K (9°F) delta T.
540 L/h * 5 K * 1.163 = 3132 watts
620 L/h * 3.5 K * 1.163 = 2527 watts
840 L/h * 4.4 K * 1.163 = 4287 watts
Calculated heating load at -12°C (10°F) = 3176 watts (and this heating load is probably overestimated since controlled mechanical ventilation was not included in the calculation, and I want only 15–18°C (59–64°F) in four rooms instead of the calculated 20°C (68°F). Also, average outside temperatures for the heat pump in my area are closer to -10°C (14°F) rather than -12°C (10°F), so there is some margin).
Maybe I have now gotten too caught up in theoretical and calculated values and can’t see the forest for the trees?
H
Hausbau 5531 Dec 2021 17:55lesmue79 schrieb:
I will write something on this, but I need to use the laptop for it, because typing on the phone while on the couch in the evening makes the topic too extensive.Hello, have you forgotten about us? Please share your experiences with us... thank you...First of all, I am not a heating specialist, refrigeration technician, or anything like that, so please keep in mind that any adjustments you make to the control settings are at your own risk.
Happy New Year, and no, I haven’t forgotten about you. I just realized that the information you collect over two years can’t simply be condensed without taking some details out of context. I caught myself reaching 4-5 DIN A4 pages just gathering the design criteria from the general contractor and myself, and that’s only about heating loads, outdoor temperature, room temperature, etc.
So here is a brief summary: overall, I am satisfied with the system—it does what it is supposed to do, namely keep the house warm.
Here are some pros and cons:
I find Vaillant’s menu navigation rather confusing and not very clear.
For example, there are two technician levels: one can be found in the Multimatic controller, and the other in the VR71 (or whatever the name of the small display is exactly).
The Multimatic includes the basic technician functions that can still be managed by the end user, such as heating times, room temperatures, hot water temperature, heating curve, electric heating rod, bivalence point, and so on—basically the fundamental configuration settings.
The small display is intended for more skilled or professional users. You should be very careful when adjusting settings here, as it provides access to compressor hysteresis, compressor power limitation, silent mode, and similar advanced functions. So be cautious and don’t just press buttons randomly.
Generally, my 3.5 kW (about 12,000 BTU/h) system, which roughly matches the heating load, is oversized for the transitional period (roughly 0°C to 12°C (32°F to 54°F) outdoor temperature). Here’s how I approached it (now the first 4-6 pages would come into play):
I disabled the electric heating rod or set its activation point to -12°C (10°F) outside temperature, although in spring 2021, we had a low of -13.9°C (8.98°F).
I am still adjusting the heating curve as heating is somewhat of a hobby for me and relates to my professional background. Currently, I run a heating curve set at 0.15 or 0.20.
To prevent short cycling during the transitional period, I set the compressor hysteresis to its maximum of 15K (27°F). This means the actual measured flow temperature can exceed the calculated flow temperature (according to the heating curve) by 15K (27°F) before the compressor switches off, unless the control system turns the system off earlier.
Vaillant’s control works via the energy integral, meaning it compares the target flow temperature (set by the heating curve) to the actual flow temperature (the temperature currently delivered or ideally delivered). The system detects whether heating is needed (actual flow temperature equal to or below target) or if the house is sufficiently warm (actual flow temperature equal to or above target).
Behind the scenes, the so-called degree-minutes rise or fall; you can also adjust these in the controller’s settings. I have mine set from 60°C minutes up to a maximum of 100°C minutes.
The system starts or stops heating at 0°C degree-minutes, which means no heating demand because the actual flow temperature is above the target. If the opposite is true (actual flow temperature below target), the system counts in the negative range—for example, -100°C minutes—and activates the heating at that point, attempting to bring it back toward 0°C degree-minutes by raising the actual flow temperature above the target as heat is released. The goal is to keep the difference between target and actual flow temperature (the heating curve again) as small as possible to reduce cycling.
I now believe that a heating curve set too low can also contribute to short cycling. This is because the minimum power of the heat pump (in my case from 0°C to 0°C) is still high, so with a heating curve of 0.10 or 0.15, which calculates a flow temperature of about 23–24°C (73–75°F) for outdoor temperatures from 0°C to 10°C (32°F to 50°F), the actual flow temperature rises too quickly since the system itself delivers at least 24–25°C (75–77°F) flow temperature at its lowest output.
Therefore, I am experimenting with the system’s base point (minimum heating circuit temperature in the Multimatic) and have set this to 25°C (77°F). This forces the system to start with at least 25°C flow temperature until the heating curve demands a higher temperature. This might be nonsense—it is still theoretical.
At the same time, the maximum temperature setting lets you influence the upper part of the heating curve. In my case, it is set to 30°C (86°F), because the underfloor heating is designed for this maximum temperature.
There is also a desired temperature setting, adjustable via a rotary knob in the Multimatic. However, this shifts the heating curve in parallel, which I do not want. For example, with a 0.10 curve ending at 30°C (86°F) flow temperature at -12°C (10°F) outside temperature, the maximum flow temperature also shifts upward depending on the set room temperature, not just the base point.
The system’s electricity consumption averages around 0.5–0.7 kWh depending on outdoor temperature, which currently ranges from a maximum of 10°C (50°F) to a minimum of -8°C (18°F)—it hasn’t been colder during this heating season. The annual heating efficiency (COP) displayed is somewhere between 4.5 and 5, but I am still experimenting a lot with the system.
Ultimately, I am satisfied if I see a total electricity consumption for household and heating of just under 4,400 kWh per year, while my photovoltaic system fed just under 7,000 kWh back into the grid. We don’t stress over this though; washing machine, dryer, dishwasher, infrared cabin, and so on run whenever needed, including at night. We do not organize our daily routine around the photovoltaic system.
What we could still optimize is our ventilation behavior, which would certainly allow for improvements in heating efficiency and temperature level. But the government firmly believes that window-tilt ventilation is not counterproductive. (Note the irony: "Because the air coming in through tilted windows is fresher than the air supplied by controlled mechanical ventilation.") I’ll leave it at that—everyone has their own opinion anyway.
If you have specific questions about individual menu items, please ask directly; otherwise, this would become too lengthy and my post even longer.
I deliberately have not gone into the parameters for hot water because we enjoy the luxury—or some might say the criticized energy loss—of a circulation line and because storage temperatures and time settings must be configured individually anyway.
Happy New Year, and no, I haven’t forgotten about you. I just realized that the information you collect over two years can’t simply be condensed without taking some details out of context. I caught myself reaching 4-5 DIN A4 pages just gathering the design criteria from the general contractor and myself, and that’s only about heating loads, outdoor temperature, room temperature, etc.
So here is a brief summary: overall, I am satisfied with the system—it does what it is supposed to do, namely keep the house warm.
Here are some pros and cons:
I find Vaillant’s menu navigation rather confusing and not very clear.
For example, there are two technician levels: one can be found in the Multimatic controller, and the other in the VR71 (or whatever the name of the small display is exactly).
The Multimatic includes the basic technician functions that can still be managed by the end user, such as heating times, room temperatures, hot water temperature, heating curve, electric heating rod, bivalence point, and so on—basically the fundamental configuration settings.
The small display is intended for more skilled or professional users. You should be very careful when adjusting settings here, as it provides access to compressor hysteresis, compressor power limitation, silent mode, and similar advanced functions. So be cautious and don’t just press buttons randomly.
Generally, my 3.5 kW (about 12,000 BTU/h) system, which roughly matches the heating load, is oversized for the transitional period (roughly 0°C to 12°C (32°F to 54°F) outdoor temperature). Here’s how I approached it (now the first 4-6 pages would come into play):
I disabled the electric heating rod or set its activation point to -12°C (10°F) outside temperature, although in spring 2021, we had a low of -13.9°C (8.98°F).
I am still adjusting the heating curve as heating is somewhat of a hobby for me and relates to my professional background. Currently, I run a heating curve set at 0.15 or 0.20.
To prevent short cycling during the transitional period, I set the compressor hysteresis to its maximum of 15K (27°F). This means the actual measured flow temperature can exceed the calculated flow temperature (according to the heating curve) by 15K (27°F) before the compressor switches off, unless the control system turns the system off earlier.
Vaillant’s control works via the energy integral, meaning it compares the target flow temperature (set by the heating curve) to the actual flow temperature (the temperature currently delivered or ideally delivered). The system detects whether heating is needed (actual flow temperature equal to or below target) or if the house is sufficiently warm (actual flow temperature equal to or above target).
Behind the scenes, the so-called degree-minutes rise or fall; you can also adjust these in the controller’s settings. I have mine set from 60°C minutes up to a maximum of 100°C minutes.
The system starts or stops heating at 0°C degree-minutes, which means no heating demand because the actual flow temperature is above the target. If the opposite is true (actual flow temperature below target), the system counts in the negative range—for example, -100°C minutes—and activates the heating at that point, attempting to bring it back toward 0°C degree-minutes by raising the actual flow temperature above the target as heat is released. The goal is to keep the difference between target and actual flow temperature (the heating curve again) as small as possible to reduce cycling.
I now believe that a heating curve set too low can also contribute to short cycling. This is because the minimum power of the heat pump (in my case from 0°C to 0°C) is still high, so with a heating curve of 0.10 or 0.15, which calculates a flow temperature of about 23–24°C (73–75°F) for outdoor temperatures from 0°C to 10°C (32°F to 50°F), the actual flow temperature rises too quickly since the system itself delivers at least 24–25°C (75–77°F) flow temperature at its lowest output.
Therefore, I am experimenting with the system’s base point (minimum heating circuit temperature in the Multimatic) and have set this to 25°C (77°F). This forces the system to start with at least 25°C flow temperature until the heating curve demands a higher temperature. This might be nonsense—it is still theoretical.
At the same time, the maximum temperature setting lets you influence the upper part of the heating curve. In my case, it is set to 30°C (86°F), because the underfloor heating is designed for this maximum temperature.
There is also a desired temperature setting, adjustable via a rotary knob in the Multimatic. However, this shifts the heating curve in parallel, which I do not want. For example, with a 0.10 curve ending at 30°C (86°F) flow temperature at -12°C (10°F) outside temperature, the maximum flow temperature also shifts upward depending on the set room temperature, not just the base point.
The system’s electricity consumption averages around 0.5–0.7 kWh depending on outdoor temperature, which currently ranges from a maximum of 10°C (50°F) to a minimum of -8°C (18°F)—it hasn’t been colder during this heating season. The annual heating efficiency (COP) displayed is somewhere between 4.5 and 5, but I am still experimenting a lot with the system.
Ultimately, I am satisfied if I see a total electricity consumption for household and heating of just under 4,400 kWh per year, while my photovoltaic system fed just under 7,000 kWh back into the grid. We don’t stress over this though; washing machine, dryer, dishwasher, infrared cabin, and so on run whenever needed, including at night. We do not organize our daily routine around the photovoltaic system.
What we could still optimize is our ventilation behavior, which would certainly allow for improvements in heating efficiency and temperature level. But the government firmly believes that window-tilt ventilation is not counterproductive. (Note the irony: "Because the air coming in through tilted windows is fresher than the air supplied by controlled mechanical ventilation.") I’ll leave it at that—everyone has their own opinion anyway.
If you have specific questions about individual menu items, please ask directly; otherwise, this would become too lengthy and my post even longer.
I deliberately have not gone into the parameters for hot water because we enjoy the luxury—or some might say the criticized energy loss—of a circulation line and because storage temperatures and time settings must be configured individually anyway.
A few more details for the statisticians, taken from the Vaillant controller:
The total values should be divided by 2 for the years 2020 and 2021.
Coefficient of Performance (COP):
Monthly COP Heating 6.4 December 2021
Overall COP Heating 5.6 2020–2021
(I’m not entirely sure; I think the Vaillant controller calculates this a bit more accurately.)
Energy Costs:
Total energy consumption 4847 kWh
So on average per year:
2400 kWh x €0.30 = €720/year = €60/month
Note that from the 2400 kWh, the photovoltaic share still needs to be subtracted, meaning basically the total hot water production in summer 2020/2021.
Starts/Operating Hours:
Compressor hours 7385 h
Compressor starts 4678
Runtime-to-start ratio 1.5 h per start
Heating Element:
Electric heating element consumption 1405 kWh
(Heating up screed from December 2019 to mid-January 2020)
The total values should be divided by 2 for the years 2020 and 2021.
Coefficient of Performance (COP):
Monthly COP Heating 6.4 December 2021
Overall COP Heating 5.6 2020–2021
(I’m not entirely sure; I think the Vaillant controller calculates this a bit more accurately.)
Energy Costs:
Total energy consumption 4847 kWh
So on average per year:
2400 kWh x €0.30 = €720/year = €60/month
Note that from the 2400 kWh, the photovoltaic share still needs to be subtracted, meaning basically the total hot water production in summer 2020/2021.
Starts/Operating Hours:
Compressor hours 7385 h
Compressor starts 4678
Runtime-to-start ratio 1.5 h per start
Heating Element:
Electric heating element consumption 1405 kWh
(Heating up screed from December 2019 to mid-January 2020)
First of all, thank you very much! Really well summarized! 🙂
I don’t understand this relationship with the compressor. At these transitional temperatures, I have cycles of about 30 minutes, and I would like to optimize that. If the set supply temperature is around 25°C (77°F), does the compressor really go up to 40°C (104°F) when it is allowed to exceed the target by 15K (27°F)?
lesmue79 schrieb:
To prevent short cycling during the transitional period, I set the compressor hysteresis to the maximum of 15K (27°F). This means that the actual/measured supply temperature is allowed to exceed the calculated supply temperature (for example, according to the heating curve) by 15K (27°F) before the compressor is switched off, unless the system control switches off the unit earlier.
...
By now, I believe that a heating curve set too low can also contribute to short cycling because the minimum output of the heat pump (in my case from 0°C to 0°C (32°F to 32°F)) is already so high that—for example, at a heating curve of 0.10 or 0.15, which calculates a supply temperature of about 23–24°C (73–75°F) at outdoor temperatures between 0°C and 10°C (32°F and 50°F)—the actual supply temperature rises too quickly. This happens because the system delivers a minimum supply temperature of at least 24–25°C (75–77°F) even at the lowest limit.
I don’t understand this relationship with the compressor. At these transitional temperatures, I have cycles of about 30 minutes, and I would like to optimize that. If the set supply temperature is around 25°C (77°F), does the compressor really go up to 40°C (104°F) when it is allowed to exceed the target by 15K (27°F)?
OWLer schrieb:
First of all, thank you very much! Really well explained! 🙂
I don’t quite understand the connection with the compressor. At these transition temperatures, I have about 30-minute cycles and would like to optimize that. If the target flow temperature is around 25°C (77°F), does the compressor then go up to 40°C (104°F) if it’s allowed to exceed by 15K? Yes, initially about the 40°C (104°F); under normal conditions, the system should have mostly worked off the energy integral. The degree-minutes are calculated based on the temperature difference.
If you have an energy integral set, for example, at 100, and the actual flow temperature exceeds the setpoint by 5K—say 30°C (86°F) flow instead of the required 25°C (77°F)—it would take 20 minutes to reduce the 100 degree-minutes before the heating cycle ends. So the higher the hysteresis, the faster the energy integral is reduced.
So, the compressor either switches off because:
A: The energy integral has been worked off; for example, from -180°C degree-minutes (compressor on) to 0°C degree-minutes (compressor off),
or
B: The compressor hysteresis limit is reached and forced shutoff occurs (actual flow temperature + compressor hysteresis = compressor off).
Regarding B, with unfavorable settings, it can happen that the compressor shuts off due to the hysteresis faster than through the energy integral.
In my factory settings, for example, the hysteresis was set to 3K. If there is a too high heating curve combined with individual room control, it can happen that about 95% of the actuators are controlling the heating circuits and only 1-2 circuits remain open. These get more flow automatically, heat up faster, and the actual flow temperature rises within a few minutes by 3K (the set hysteresis) above the flow temperature setpoint. Then the compressor is forcibly shut off and, I believe, locked out for 3 minutes before the cycle starts over, hoping that eventually the energy integral is worked off.
I find forced cut-offs always worse than a smooth reduction by working through the energy integral. Whether it needs to be 15K as in my case is up to each individual.
lesmue79 schrieb:
Degree minutesI was thinking more about geometry and its angular minutes. Interesting.I have now set it to 100 and will monitor it. The compressor hysteresis was set to 7°C (45°F) for me.
Similar topics