Tricky heating control problem

[IMM]

Too crude and will not maintain the room temp accurately. Seamless temperature control can be has by using the correct controls. In this case a commercial controller is the only option.

[IMM]

If you are using modulated valves then Landis & Staefa have a wall mounted 2 stage controller with a built-in temp sensor. You can have an outside sensor which will raise and lower the setpoint. This unit is £130, so excelelnt value.

mixing valve is the 3/4" VMP43 at about £64 actuator is the SSB 61.00 0 -10v @ £108 two mixing valves, 2 actuators and one controller is: £474. Not bad, for what it will do.

[Aidan]

to modulate the radiator output with a boiler reset?

It's not practical, on it's own. You could use the boiler reset to supply the higher flow temperature required by the rads, and then have a mixing valve to supply the lower flow temperature required for the UFH. My background is in commercial systems, so I've never dealt with a boiler reset system. I'd be inclined to have 2 mixing valves to minimize the amount of control circuitry in the boiler.

The vessel's similar to a section of the primary header pipe, but with a big diameter and so a large water content. The primary circuit would contain 50 or 100 gallons, rather than 5 or so gallons. The buffer vessel's temperature will change gradually in response to heat inputs or outputs and the mixing valves won't be motoring up & down continually to compensate for a rapidly fluctuating flow temperature. A modulating burner will also help minimize the temperature fluctuations, but there will still be mismatches between the heat load and boiler output. Most 32 kW boilers will modulate down to about 9kW.

You need a primary pump for the boiler. If you've got a mixing valve, serving the UFH, then it has got to have it's own secondary pump. You could use the boiler pump to supply the rads with a boiler reset temperature control, or you could have another mixing valve & pump set. Decision time.

zone valve per circuit?

Just my preferred method. The use of 2-port zone valves will cause changes in the other zones' flow rates when they close.

about? For most of the year, a primary flow temperature of 55°C will be sufficient to heat the house.

Probably less efficient, but not much in it. A modulating condensing boiler with a 0-10V burner control input would be ideal. The flow temperature would be the minimum required, i.e., the maximum required by the rads, the UFH OR the DHWS.

seems a shame to heat the flow on/off at 82°C and then modulate the u.f.h. circuit down to 55°C. Wouldn't it be better to heat the boiler flow continuously at 55°C. I may be missing something here...

I had a non-condensing on/off type boiler in mind. The heating would operate at less than 55°C for much of the time. With boiler reset, you'd still have to boost the temperature when there was a DHWS demand. You'd then have to tolerate overheating of the rads, or stop them temporarily. If you had a mixing valve serving the rads, it would maintain a constant mixed flow temperature by just modulating down to compensate for the higher flow temperature. Again, my preferred method from commercial experience. You're not missing much.

temperatures than the thermostatic mixer setpoint, which seems desirable.

Unless the users had turned the thermostatic control down. I only mentioned the thermostatic mixer as a control device which works adequately, for most people, despite it's inherent lack of sophistication, i.e., electronics and motorized actuators.

rapidly but then drop off when the u.f.h. gets up to speed?

The value of the PID constants you've programmed into the control loop. For an explanation see

gets technical. The loop compares the sensor input with the setpoint and processes the output. The proportional output term is the error times the gain. So if it's 20°C and the setpoint is 21°C, the error is 1 °C. You could have a 10% output from one (proportional only)control loop to the UFH (gain = 10). At the same time, with the same sensor input, you could have a 50% output to the rads (gain = 50) from another loop. You programme in the proportional gain, the Integral & Derivative constants into each loop to get the response you want, depending on the system characteristics. My control theory isn't up to much, so you'll have to research it on Google if you want a more detailed explanation. I'm not going to even try to explain Integral and Derivative constants. You could program two loops to get one to react like a hare and the other like a tortoise.

You should tune the control loops (Ziegler and Nicholls method?) to get the optimum response. Most commercial control contractors take an educated guess at the initial PID values but most HVAC systems still work well with these.

do they offer over the standard boiler pump and zone valves arrangement?

Not commonly used in domestic installations. They give much better control but cost much more.

[IMM]

A boiler reset system as you call it is known as weather compensation, or just compensation. This is fully commercial, and if you are in commercial controls you would have come across this as an everyday thing.

I know of no domestic boiler tha will take an external control signal to modulate the burner.

No. Best to raise and lower the boiler temp to suit the UFH demands. This can be set via an outside weather compensator. The unit you need is the Seimans BLC1.A:

Boiler Load Compensator is a heating controller that can be used as either a:

1. Weather Compensator
2. Load Compensator.

The BLC1.A can be set up to control conventional or underfloor heating systems and tuned to condensing boilers.

on unit. from Newway:

01905 794242. About £200.

This is ideal to have on the UFH circuit. Have one condensing combi boiler do the UFH and this unit controls its temperature output. Have another combi do the rads and this need not be under outside waether contro,a s it is just a boost. The two combi's combine to give high flow hot water.

You will still need a two stage controller to seamlessly combine the UFH and booster rads and maintain the internal temperatures.

What are your DHW needs? How many baths, showers etc?

[John Aston]

1 bath and 3 showers for a family of five.

Thanks for your thoughts. I'll take a look at those controllers you suggested.

[IMM]

There are a number of routes you can take. You obviously are installing a new system(s) and boiler. Are you keeping the same cylinder, or is it a full new system? How is your cold water mains. Can it cope with all of the house?

[John Aston]

Thanks, Aidan, for taking the time to compose that informative reply. I appreciate the help.

[John Aston]

I had a thought further thought about controlling UFH as a primary heat source and using radiators as boost heaters only.

The UFH is controlled by room thermostats. My problem is that when the weather is cold, the UFH output is not sufficiently high to heat he rooms. At this point the radiators come on. Rather than put a room temperature TRV on each radiator, though, I could use a *water temperature* valve (Danfoss FJVR). That means that the maximum kW output of the radiator will be fixed at a user-defined level whatever the room temperature. From my thermal calculations, I know the kW shortfall for each room when the UFH is at its maximum output and the outside temperature is at -2°C. I can adjust each radiator output to exactly match this shortfall on a room-by-room basis.

As the outside temperature increases, weather compensation will reduce the radiator output as the boiler flow temperature falls.

This seems a better solution than using TRVs. With TRVs, I would have two parallel heating systems fighting to reach the same setpoint.

[Aidan]

reduce the

Summat along these lines?

article uses BTU/hr, degF, etc., but I'm sure you'll make sense of it.

I'm not sure that you're taking account of transient heat gains, e.g.,

3 or 4 people in one room with the lights and TV on will provide a substantial heat gain. Similarly, east-facing rooms may get big a.m. solar gains through the windows, west facing may get big p.m. solar gains and they'll all have fairly constant heat losses overnight.

The radiator TRV's, on a 'normal' system would compensate by reducing the flow rate to the rads as the room temperature went up. With your system, the response time of the UFH would be too slow to compensate for this, so you'd probably need some control on the rads, which I'm not sure you've got. Not picking holes, I haven't read the proposals in detail so you may already have got this covered. I'm not familair with the workings of the Danfoss valve, I'll have to read up on it.

[Andrew Gabriel]

magic. I'm in the middle of designing a control system to do just this. I'll have to read the article in detail to see if there is anything I've not taken into account.

My father installed what was originally coal-fired central heating in late 1950's, and he used this scheme for the temperature control. It was being pushed by the Coal Board at the time. It was called BMT -- no idea what it stands for. It used an outdoor temperature phial on a very long capilliary tube which operated a bellows in a rather complicated valve which set the radiator water temperature by mixing the boiler water with the return water to create an automatically adjusting flow temperature. The unit was powered by the water pressure across the pump, which had to run continuously. The BMT unit continued to be used with a later gas fired boiler up until 2000, when the whole lot was replaced with a small wall mounting low volume boiler. In that

40 year period, I think the only thing that ever went wrong with it was an O-ring went hard and started leaking.

[John Aston]

Yes, exactly along those lines. Thanks for the link.

Agree with you, Aidan.

I suppose that any house with underfloor heating (whether or not there are radiators installed) will have this response problem. There don't seem to be many unhappy UFH users though so I guess that it's all relative. My house is mid-terrace with a south-facing wall partially shaded by trees and a north facing wall which is fairly well sheltered from the wind. Hopefully, the external influences with be small.

My UFH will be installed on a suspended timber floor, which I'm told has a faster response that a screeded floor. Don't know how much faster, though!

The problem with TRVs is that the radiators could take over from the UFH as the primary source of heating, which would lead to a less efficient heating system.

Basically, if the UFH can respond fast enough (or I can live with the fluctuations), I'm sorted.