Radiators

Installed radiators

At 24 Dec 2023.

Flow temperature

Return temperature

According to BSRIA, BS EN 14511 (the standard used for testing heat pumps) specifies a flow temperature of 45°C and a return temperature of 40°C, i.e. a temperature drop across radiators of 5°C.

Octopus say they are sizing radiators using a "flow temperature" of 45°C. By studying the correction factors they are using (see beginning of the Radiators section below) they appear to be using a radiator temperature of 42°C. So they are presumably assuming a return temperature of 39°C, and a drop across the radiators of 6°C.

There seems to be a rule-of-thumb to use a drop through the radiators of 1/7 of the flow temperature, i.e. 7°C for 50°C flow and 8°C for 55°C. The government Report on compliance on page 63 uses 1/7 (quoting the original MCS Heat Emitter Guide), and Octopus's use of 6°C for 45°C flow and the general use of 10°C with 70°C flow are consistent with this. Hopefully this will not affect radiator sizes too much.

This rule-of-thumb doesn't make much sense to me. It would make more sense to use a fraction of the difference between the flow temperature and room temperature. The following table shows this fraction for various examples, assuming a room temperature of 20°C:

Hello Renewables aim for a 50/55°C flow temperature. This would suggest drop of 10/11°C across the radiators and an average radiator temperature of 45/49.5°C. So using an average radiator temperature of 47°C for sizing radiators would be consistent with this aim.

SCOP dependence on flow temperature

The MCS certified products directory provides SCOP values for heat pumps at different flow temperatures. These values are lower than those in the Daikin product catalogue, so they are presumably for different conditions.

The MCS certification for the Daikin EDLQ07CAV3 Monobloc (as used by Octopus) lists a SCOP of 3.57 at a flow temperature of 45°C.

The MCS certification for the Daikin ERGA06EVA heat pump provides the following table of SCOP values at different flow temperatures:

Using the Altherma M at a flow temperature of 45°C (as specified by Octopus) would produce a SCOP of 3.57.

Using the Altherma 3 R at a flow temperature of 50°C would produce a SCOP of 3.43 (0.14 less than Octopus)
Using the Altherma 3 R at a flow temperature of 55°C would produce a SCOP of 3.15 (0.42 less than Octopus).

Hello Renewables aim for a 50/55°C flow temperature. According to my flow temperature notes above, this would suggest drop of 10/11°C across the radiators and an average radiator temperature of 45/49.5°C. So using a flow temperature of 52°C and an average radiator temperature of 47°C for sizing radiators would be consistent with this aim.

If we assume a 6°C drop through the radiators, the average radiator temperature will be 47/52°C for a flow temperature of 50/55°C.

Derating

Updated Sep 2023 to use later Stelrad table (slightly smaller factors) and add extra entries.

According to Stelrad a radiator specified at ΔT 50°C has to be derated by the factors in this table:

Available radiators

Stelrad radiator types

Where two figures are given for distances from the wall these correspond to two possible orientations of the mounting brackets.

Water contents figures are approximate. See table in this Stelrad document for more accurate figures, broken down by radiator height.

Stelrad radiator weights

Weights in kg/m. I.e. multiply by radiator width to get radiator weight. Data from Stelrad Compact Horizontal and Compact K3 documents.

If the maximum radiator weight I can handle is 35kg, then the following table gives the maximum usable widths and corresponding powers:

Stelrad radiators

Kudox radiators

See Kudox Technical Data and Fitted Dimensions.

Heat losses

Updated Oct 2024:.

This table shows the heat losses at various outdoor temperatures according to my latest heat model (lossSL.yaml).

The ensuite and study doors are assumed to be open (assumed closed before Aug 2024).
All setpoints are assumed to be 21°C, except for the lounge entrance at 21.6 (MCS setpoints assumed before Aug 2024).

Total heat loss varies by 210W/°C for outside temperatures in range -5°C to 10°C, and can be approximated by 5053-210*T, where T is outside temperature.

We currently don't have a radiator in the ensuite (apart from the towel radiator), but we could install one. To handle the 200W heat loss at -2°C with a radiator temperature of 48°C and a room temperature of 21°C, we would need a radiator rated at 200/0.447 = 447W. A K1 450x600 radiator is rated at 454W.
Simon on 12 Sep 2024 calculated the ensuite loss at 300W, for which we would need a radiator rated at 300/0.447 = 671W. This would require a K1 600x700 @ 686W, or one of our spare P+ 600x600 @ 808W, or a K2 450x500 @ 685W.

To handle the 633W loss in the study at a radiator temperure of 48°C would require a radiator rated at 633/.447 = 1416W. The most powerful radiator we could install in the available width is a K3 600x400 rated at 956W. This would require a derating of at most 633/956 = 0.662, which corresponds to a ΔT of 36°C and hence a radiator temperature of 21+36 = 57°C.

Insulating both the ceiling and the garage wall of the study could reduce the loss to about 452W. With a K3 600x400 this would require a derating of at most 452/956 = 0.473, which corresponds to a ΔT of 28°C and hence a radiator temperature of 21+28 = 49°C.

Comparison of possible radiator changes

Written Sep 2024

This section compares the possible radiator changes investigated in the following sections. The possibilities are:

Plan Q adds a radiator in the back room to the existing set (minimum requirement for Octopus 2024).
Plan R is like plan Q but with a larger hall radiator.
Plan SL is like plan R but with an extra radiator in the lounge, and not such a big hall radiator.

Red cells indicate rooms where the temperature is more than 0.2°C below the target temperature (i.e. 20.7°C or less).

This table shows room temperatures with a radiator temperature of 48°C (flow temperature 50°C) and outside temperature of -2°C.

This table shows room temperatures with a radiator temperature of 52°C (flow temperature 55°C) and outside temperature of -5°C.

Performance with radiators installed for heat pump (plan SL)

Updated Feb 2025

This is the predicted performance of the radiators we have installed for use with the heat pump (installation completed 31 Jan 2025) based on my modelling (planSL.yaml). This includes additional back room and lounge radiators (see below) and a larger hall radiator,  This model splits the lounge into 3 areas: lounge entrance, lounge and lounge corner (to left of fireplace).

Setpoint is 21°C for all rooms, except for the lounge entrance at 21.6 and the bathroom at 22°C.
No heating in the ensuite.

Note that at -2°C outside, the Daikin heat pump is 13% more efficient at a flow temperature of 50°C (radiator temperature 48°C) than at 55°C (radiator temperature 52°C).

The Daikin heat pump was installed configured with a WD curve with a flow temperature of 50°C at -7°C outside, and a flow temperature of 25°C at 20°C outside. This produces a slope of (50-25)/(20- -7) = 0.93, and a flow temperature of 50 - (-2 - -7) * 0.93 = 45°C at -2°C outside.

The table below shpws minimum outside temperature for various flow temperatures (assuming radiator temperature is 2.5°C less than flow temperature), according to my model. For comparison, the corresponding outside temperatures using the Octopus WD configuration are also shown.

Based on the model results, a possible WD configuration is flow 30°C at outside 21°C, and flow 55°C at -6°C. This gives a slope of 0.93.

Performance with extra back room radiator and bigger hall radiator (plan R)

Written Sep 2024

Installing a bigger hall radiator should help with the study, where a K3 radiator would be needed even with extra insulation. It should also help with the lounge (and other rooms to a lesser extent).

This is the predicted performance of the radiators we have had installed since Dec 2023 plus the additional back room radiator (see below) and a larger hall radiator, based on my modelling (planR.yaml).

Setpoint is 21°C for all rooms.
No heating in the ensuite.

Radiator temperatures °C needed for various outside temperatures:

Performance with extra back room radiator (plan Q)

Written Sep 2024

Simon from Octopus calculated that we would need an extra K2 600x800 radiator in the back room.

This is the predicted performance of the radiators we have had installed since Dec 2023 plus the additional back room radiator, based on my modelling (planQ.yaml).

Setpoint is 21°C for all rooms.
No heating in the ensuite.

Installed radiator performance

Written Sep 2024

Predicted performance using my model

This is the predicted performance of the radiators we have had installed since Dec 2023, based on my modelling (planL.yaml).

Setpoint is 21°C for all rooms.
150W towel in ensuite is assumed to be permanently on.

This is the predicted performance of the radiators we have had installed since Dec 2023, with a 378W radiator added in the ensuite, based on my modelling (planL.yaml).

Setpoint is 21°C for all rooms.

Required flow temperature for Octopus 2022 heat losses

Assumptions:
Outside temperature -2°C.
MCS required temperatures.
Flow temperature is 3°C greater than the radiator temperature.
150W towel in ensuite is on.

According to Octopus's heat loss figures, the back room requires significant extra heating.
At a 50°C flow temperature and hence a radiator temperature of 48°C, ΔT is 27°C and the derating factor is 0.447.
Under these conditions the existing radiator would output 1558*0.447 = 696W. The heat loss is 1134, so an extra 1134-696 = 438W is needed.
This would require an extra radiator with a rating of 438/0.447 = 980W. A K2 600x600 is rated at 1039W. Simon on 12 Sep 2024 reckoned that a K2 600x800 @ 1385W would be needed for the heat loss he had calculated.

Required flow temperature for my heat losses

Updated Oct 2024

Assumptions:
Outside temperature -2°C.
All rooms at 21°C.
Flow temperature is 3°C greater than the radiator temperature.
150W towel in ensuite is on.

Effect of hall setpoint on hall, study and lounge losses (signed figures are relative to 21°C):

Radiators for 55°C flow temperature

Written Sep 2023.

If we assume a 6°C drop through the radiators, the average radiator temperature will be 52°C for a flow temperature of 55°C.

For a room temperature of 21°C, ΔT will be 31°C and correction factor will be 0.534.
For a room temperature of 18°C, ΔT will be 34°C and correction factor will be 0.606.

Prices are from Stelrad website, including VAT. They could probably be bought cheaper elsewhere, and an installer wouldn't need to pay VAT.

Modelling was done using homeW52.yaml specification.

Lounge radiator

The output of the specified vertical radiator isn't sufficient, but there isn't a proper alternative unless we are willing to install an extra radiator somewhere in the lounge. However, there are some things we can do to compensate for the low output.

Firstly, we can size the hall radiator for a temperature of 21°C (instead of the standard 18°C) to assist the lounge radiator (and the study radiator, which is also underpowered). We can also allow the hall temperature to rise above 21°C to allow extra heat to flow into the lounge and study. Anyway, in practice it would not be possible to maintain a 3°C difference between the hall and the other two rooms with open doorways. (I've tried sizing the hall radiator even larger, but it doesn't help much.)

Secondly, in cold weather we could raise the heat pump flow temperature to 60°C (say) to provide extra assistance.

Bathroom radiator

The available width between the door (open) and the bath is 600mm. The fittings each side of the radiator (TRV and lockshield) will need about 160mm, so the maximum radiator width is 400mm.

Radiators for 60°C flow temperature

Written Sep 2023.

If we assume a 6°C drop through the radiators, the average radiator temperature will be 57°C for a flow temperature of 60°C.

For a room temperature of 21°C, ΔT will be 36°C and correction factor will be 0.651.
For a room temperature of 18°C, ΔT will be 39°C and correction factor will be 0.723.

Prices are from Stelrad website, including VAT. They could probably be bought cheaper elsewhere, and an installer wouldn't need to pay VAT.

Modelling was done using homeW57.yaml specification.

With these radiators, a radiator temperature of 48°C (flow temperature of 51°C) would be needed for an outside temperature of 7°C.

5 new radiators are needed, at a cost of £747, compared to 6 new radiators at a cost of £912 if we used a 55°C flow temperature.

A compromise might be to use the existing hall radiator and change the other 5 radiators to the 55°C flow specification. This would cost £806, i.e. £59 more than the 60°C flow specification, but £106 less than changing all 6 radiators. The hall radiator could be upgraded later if required.

The performance of this compromise would be:

At 55°C flow, the lounge would be 0.2°C cooler and the study would be 0.1°C cooler than if the hall radiator was replaced. These differences would probably not be noticeable. The hall would be 0.9°C cooler.

At 60°C flow, the temperatues would be unchanged by not replacing the hall radiator.

With these radiators, a radiator temperature of 53°C (flow temperature of 56°C) would be needed for an outside temperature of 7°C.

Radiators for 65°C flow temperature

Written Sep 2023.

I am assuming a flow temperaure of 65°C and return temperature of 59°C, hence a radiator temperature of 62°C. This could be for the existing gas boiler (or a replacement boiler), or possibly for the Vaillant Arotherm plus heat pump.

For a room temperature of 21°C, ΔT will be 41°C and correction factor will be 0.771.
For a room temperature of 18°C, ΔT will be 44°C and correction factor will be 0.846.

Prices are from Stelrad website, including VAT. They could probably be bought cheaper elsewhere, and an installer wouldn't need to pay VAT.

Modelling was done using home W65.yaml specification.

With these radiators, the table below shows the heating powers and radiator temperatures needed for the specified outside temperatures. The radiator temperature is based on the lounge radiator, which is the main limiting radiator. Flow temperatures for a heat pump would typically be about 3°C higher than the radiator temperature.

Radiators for 70°C radiator temperature

Written Sep 2023.

I am assuming a flow temperaure of 75°C and return temperature of 65°C, hence a radiator temperature of 70°C. This could be for the existing gas boiler (or a replacement boiler), or possibly for the Vaillant Arotherm plus heat pump.

For a room temperature of 21°C, ΔT will be 49°C and correction factor will be 0.974.
For a room temperature of 18°C, ΔT will be 52°C and correction factor will be 1.052.

Prices are from Stelrad website, including VAT. They could probably be bought cheaper elsewhere, and an installer wouldn't need to pay VAT.

Modelling was done using home W70.yaml specification.

With these radiators, a radiator temperature of 58°C (flow temperature of 61°C) would be needed for an outside temperature of 7°C.

Octopus specifications

Octopus are sizing radiators assuming a flow temperature of 45°C.

The actual correction factors used by Octopus are 0.324 for a room temperature of 21°C and 0.386 for a room temperature of 18°C. These correspond to ΔT of 21.0°C and 24.0°C respectively, and hence to a radiator water temperature of 42.0°C in both cases.

Octopus specified radiators:

MCS Heat Emitter Guide

This is my attempt to size radiators and choose the flow temperature using the MCS Heat Emitter Guide (section 4 of MGD-007 Reference Data and Tools).

The guide uses an "Oversize Factor" (radiator rated output divided room heat loss) to look up the required flow temperature (or equivalently "star rating") using a graph. The graph corresponds to the first two columns in the following table:

The third column is simply the reciprocal of the first column, and corresponds to the "Operating factor" in the Stelrad radiator temperature table. The fourth column is the difference between radiator temperature and room temperature, obtained by interpolating in the Stelrad table. The fifth column is the difference between the second and fourth columns.

Note that flow temperature is not the same as radiator temperature, since there is a temperature drop across the radiator. My guess is that MCS are assuming a room temperature of 20°C and a drop across the radiators of 8°C (although it could equally well be 21°C and 6°C respectively).

The first two columns can be approximated by:

FT = 27.8 + 52.3 / OF

Analysis of existing radiators:

The alternative output of 616W for the bathroom towel radiator is if we painted it white.

Analysis using 4.5 Guidance Table:

The guidance in the MCS table is always "Caution" for flow temperatures of 51°C and above. The guidance for 46-50°C flow changes from "Go ahead" to "Caution" when the specific heat loss is 100 or above.
"Caution" means: "System can perform at these design conditions with extra consideration on the emitter and heat pump design sought from the specialist designer/manufacturer."

An "oversize factor" of 2.4 corresponds to a radiator correction factor 1/2.4 = 0.42.
An "oversize factor" of 1.9 corresponds to a radiator correction factor 1/1.9 = 0.53.
In my radiator design above, I am using correction factors in the range 0.43 to 0.49 for a flow temperature of 50°C, and 0.53 to 0.61 for a flow temperature of 55°C. The ranges are to cover varying target room temperatures.