how to control heated floors with high latency

[HAorso33]

I heat my home with hot water in the floors, I'm not sure how common this is outside Scandinavia, but there you go. The system works like normal radiators, but the water goes in plastic pipes in the floors. I've recently upgraded my dumb controller with a Z-wave based Danfoss HC 10 and I'm now starting to fine tune the heating.
I'm currently using the system with normal thermostats, opening and closing the valves as needed. However, my wooden floors react slowly to temp changes, it takes 2 hours to increase 1 degree Celsius. So when the set point is reached, the heat will still increase in the room and increase beyond the set point.

How can I translate this into a variable I can use in programming my system? I'm not a programmer so all input is appreciated

[Rvtravlr]

Just a thought, but you may want to use Fahrenheit as you will have finer control of the temperature (1* C = approximately 34* F). Then set the swing on your thermostat to 1/2* or 1*.

[tome10]

It sounds complicated. Is the water temp variable? Sounds like you would need to keep the floor a constant temp, and then vent air if the air temp gets too warm.

[HAorso33]

hmm, yeah, not sure it will do much.

[Priest]

Just a thought, but you may want to use Fahrenheit as you will have finer control of the temperature (1* C = approximately 34* F). Then set the swing on your thermostat to 1/2* or 1*.

[Uncle Michael]

I heat my home with hot water in the floors, I'm not sure how common this is outside Scandinavia, but there you go.

I have a similar setup, and you are correct, it is not common in the US.

I'm currently using the system with normal thermostats, opening and closing the valves as needed. However, my wooden floors react slowly to temp changes, it takes 2 hours to increase 1 degree Celsius. So when the set point is reached, the heat will still increase in the room and increase beyond the set point. How can I translate this into a variable I can use in programming my system?

What is the temperature of the water in your system? You may find it useful to reduce the temperature of the water you are circulating in your floor so that you do not deliver such a large excess of heat. I do not experience any overshoot on my set point. There is about a two hour lag if I increase the set point substantially, but that is not something I do frequently. Do you have an outside temperature reset feature on your hot water loop?

[HAorso33]

I do have outside sensors, but I have to create the logic to make use of it :-)

I keep the water temp at about 33C, and it's the same temp for all zones, the system adjusts room temp by opening/closing circulation.

Maybe I can create a virtual device and use that as a set point. If I have a set point @ 22C, then tell the system to close the valves when at 21.

[tome10]

I don't really know, but it seems having the water temp at 91 deg is overkill. Maybe 5 degrees over the setpoint would be sufficient? In order to try eliminating the overshoot of the target. I'm thinking slow and steady. It would be nice to have some kind of algorithm to take advantage of the 91 deg, but it seems it would always be prone to overshooting the target. Maybe if the water temp itself was variable, and the closer to setpoint it got, the lower the temp of the water. Of course, this would affect all zones, so it would have to be an average of multiple thermostats only clocking every hour so. everything would have to be slowed down as nothing would happen quickly. I just don't think it would work with high temp water with on/off control only. I bet that feels good to walk on though at full temp. Maybe my first comment is spot on. keep the floor warm, and vent air.

[HAorso33]

I don't really know, but it seems having the water temp at 91 deg is overkill. Maybe 5 degrees over the setpoint would be sufficient? In order to try eliminating the overshoot of the target. I'm thinking slow and steady. It would be nice to have some kind of algorithm to take advantage of the 91 deg, but it seems it would always be prone to overshooting the target. Maybe if the water temp itself was variable, and the closer to setpoint it got, the lower the temp of the water. Of course, this would affect all zones, so it would have to be an average of multiple thermostats only clocking every hour so. everything would have to be slowed down as nothing would happen quickly. I just don't think it would work with high temp water with on/off control only. I bet that feels good to walk on though at full temp. Maybe my first comment is spot on. keep the floor warm, and vent air.

The only way to adjust zone temp is by opening and closing the valves. I can't adjust water temp or open/close windows when I'm at work
I am collecting data now for future reference that may make it easier to create some logic. (indoor temp, outdoor temp, and the intervalls when the heat is ON/OFF)

[cjin]

The only way to adjust zone temp is by opening and closing the valves. I can't adjust water temp or open/close windows when I'm at work

I am collecting data now for future reference that may make it easier to create some logic. (indoor temp, outdoor temp, and the intervalls when the heat is ON/OFF)

[Uncle Michael]

I do have outside sensors, but I have to create the logic to make use of it.

It is very difficult to avoid overshooting the set point in a room if the heat content of the water in the floor overwhelms the rate of loss of heat by the room. The best way to avoid that is to adjust the temperature of the circulating water to match as closely as possible the rate heat is supplied to the rate it is lost. On cold days, the water needs to be warmer than on relatively warmer days. The specific function will depend on the rate of heat loss and the volume of water in the tubing under the room. I would guess that your building is extremely well insulated, given that you are circulating relatively cool water for this time of year in Norway. That probably makes the balance between heat gain and loss more delicate, since heat loss is slow.

This article describes the concept. A search for "outside temperature reset" will get additional references.
http://www.heat-timer.com/outdoor-re...ntrol-savings/

[billt]

Another approach is to control the floor temperature, which you can do with your zone valves without changing the main water temperature.

I installed wet UFH in 3 areas when I had to replace the floor and I just use a thermostat (implemented in HS) to control the floor temperature. (The intention was to use weather compensation (outdoor reset as the USAians call it for some inexplicable reason) to control the flow temperature, modulate the floor temperature and then add air temperature control on top. In the event just controlling the floor temperature gives good enough temperature control for our purposes, until I get round to finishing the control system.

The UFH pump runs 3 times a day and when the pump is running a thermostat script runs every 3 minutes (it doesn't need to be run often as the temperature changes are so slow). The thermostat script just switches the UFH zone valves on or off, with a 2C dead band. It would probably be better to run the pump continuously and just have a set back temperature during off periods.

As you have found, an air temperature thermostat isn't a good match to UFH. The commercial UFH thermostats seem to have both a floor temperature sensor as well as an air temperature sensor.

[Uncle Michael]

Another approach is to control the floor temperature, which you can do with your zone valves without changing the main water temperature.

The UFH pump runs 3 times a day and when the pump is running a thermostat script runs every 3 minutes (it doesn't need to be run often as the temperature changes are so slow). The thermostat script just switches the UFH zone valves on or off, with a 2C dead band. It would probably be better to run the pump continuously and just have a set back temperature during off periods.

Interesting approach. Do you know how uniform the floor and air temperature readings are (spacial and temporal)?

[tome10]

I agree that having the floor/radiator control would be beneficial, but in this situation only having on/off control of a relatively high heat always on radiator just doesn't seem possible. However, if floor temp sensor was in place then cycling the valve on for like 2 seconds every hour could be the way to create some stability. Maybe a good guestimate of only cycling the floor valve on for two seconds every 30mins to hour would be a good starting point even without floor temp sensors. Also, when I say vent, I mean the actuation of a fan somewhere via automation, I don't mean someone opening a window. I've seen in wall fan kits which flow air from one room to another. An automation event could be set up to utilize fans in this way to spread the overrun of heat to another room. If a majority of zones are overheated, there would have to be a whole house exhaust fan. But, moving of air should be last resort in case of overshooting of the target temp. I think the cycling of the valves for short durations every hour or so would be beneficial in this setup. If you're a calculous expert, all of this would be elementary, but for me, I'd just have to make changes and wait for results. This would where a floor temp sensor would be beneficial.

[Priest]

Another approach is to control the floor temperature, which you can do with your zone valves without changing the main water temperature.

This is basically what I was trying to suggest above. I'm glad that the theory has actually been put to practice somewhere already.