Trystan Lea

Published: 3rd January, 2018

*Updated: July 2024*

Prior to replacement with the heat pump the original heating system was a multi-fuel stove with a back boiler connected to an 120 litre hot water tank (which also has an immersion heater) and 8x single panel radiators. The system did not really work, at least when we tried to fuel it with wood (it may well have been designed for coal).

Initially I wanted to work out whether it made any sense to connect the heat pump to the existing radiator system, would we get anything useful out of it? or did we need to replace the lot? In the end we did replace the whole system. We installed new larger radiators, a new hot water cylinder and mostly new piping.

The following details my calculations for the heat output we might have expected with the old radiator system.

The radiators were all single panel standard radiators. The following table lists the rated output of each radiator at deltaT 50K between the radiator mean water temperature and the room. These ratings can usually be found on the product listings e.g Screwfix K1 1000x600

Room |
Radiator |
Area |
Heat Output @ 50K |
Heat/m2 |
---|---|---|---|---|

Diningroom | 1200x500x45 | 0.6 m2 | 1051 W | 1752 W/m2 |

Livingroom | 1400x400x45 | 0.56 m2 | 1011 W | 1805 W/m2 |

Kitchen | 800x600x45 | 0.48 m2 | 821 W | 1710 W/m2 |

Hallway | 1200x600x45 | 0.72 m2 | 1231 W | 1710 W/m2 |

Bathroom | 800x600x45 | 0.48 m2 | 821 W | 1710 W/m2 |

Bed1 | 1000x400x45 | 0.4 m2 | 722 W | 1805 W/m2 |

Bed2 | 1000x400x45 | 0.4 m2 | 722 W | 1805 W/m2 |

Study | 1000x400x45 | 0.4 m2 | 722 W | 1805 W/m2 |

Total |
4.04 m2 |
7101W |
1758 W/m2 |

**Heat output at heat pump flow temperatures**

The following equations can be used to determine the heat output of the radiators at different water temperatures:

```
Delta_T = Mean water temperature (MWT) - Room temperature
Heat_output = Rated_Heat_Output x (Delta_T / Rated_Delta_T) ^ 1.3
```

If the radiator mean water temperature is 35°C and the room 20°C, Delta_T = 15K.

```
MWT 35°C: Heat_output = 7101 W x (15K / 50K) ^ 1.3 = 1484 W
MWT 40°C: Heat_output = 7101 W x (20K / 50K) ^ 1.3 = 2158 W
MWT 45°C: Heat_output = 7101 W x (25K / 50K) ^ 1.3 = 2884 W
MWT 50°C: Heat_output = 7101 W x (30K / 50K) ^ 1.3 = 3655 W
MWT 55°C: Heat_output = 7101 W x (35K / 50K) ^ 1.3 = 4466 W
```

My original 2017 heat loss calculation suggested a heat loss at design temperature of 6.1 kW and so I believed at the time that the existing radiator system would not work at all with the R410a Ecodan heat pump I planned on using, which could only reach a maximum temperature of 55°C at subzero temperatures.

In reality we have never needed more than 3.1 kW from the heat pump over a 24 hour period during sub-zero conditions and I have since revised my heat loss calculation based on better input assumptions down to 3.3 kW at design temperature. Subtracting ~0.4 kW of gains suggests the radiators would have needed to deliver 2.9 kW. The original radiator system should have been able to deliver this at a mean water temperature of 45°C and heat pump flow temperature of 47°C.

While the heat pump could certainly hit this flow temperature for short periods at sub zero temperatures and in-between defrosts. I wonder if it would have been able to sustain an average flow temperature of 47°C over a 24 hour period.. it would have been a good test!

I decided at the time to upgrade all the radiators and pipework, doubling the output capacity at a given water temperature to 15 kW @ DT50. This dropped the design flow temperature required from 47°C to 35°C.

Using the SCOP calculator tool, I estimate that dropping the design flow temperature from 47°C to 35°C increased space heating SCOP over the year from 3.4 to 4.2. My actual space heating SCOP does actually match the higher value! Though I do wonder if the SCOP tool is too optimistic in it's predicted SCOP for the higher design flow temperature.

Our space heating demand has been 6500 kWh/year on average. A SCOP of 4.2 and unit cost of electric of 20 p/kWh costs £310/year. A SCOP of 3.4 would have cost £382/year. This saving of only £72/year or £1440 over 20 years would take a long time to payback the additional cost of the radiator upgrades unfortunately. Using the Ecodan datasheet COP values instead of the carnot COP equation increases this saving slightly to £84/year or £1680 over 20 years. It's likely that a radiator system would last 40 years and so savings could be twice this amount over the long term.

It may arguably make sense if the existing radiator system is in good condition, well cleaned and sufficient in size to fall within the heat pump flow temperature range, to continue using the existing radiator system and then upgrade to a more efficient system at the next replacement point. An accurate heat loss is key to making this assessment! As my example above shows, my original calculation suggested that the original radiator system was significantly undersized, we have to deviate significantly from standard guidance to make that assessment.

```
Outside diameter: 490mm
Outside height: 915mm
Inside diameter: 440mm
Inside height: 875mm
Insulation: 25mm
```

Approximate volume taking into account dome top: 0.125 m3, density of water 1000 L/m3 suggests a tank capacity of ~125 Litres. Or checking common sizes on coppercylinder.co.uk a 36x18 inch tank holds approximately 120 Litres of water.

The thermal conductivity of polyurethane foam is around 0.022 W/m.K, at 0.25mm thickness it yields a U-value of 0.88 W/K.m2. At a cylinder surface area of 0.667m2, a water temperature of 45C and a bathroom temperature of 22C. Heat would be lost at a rate of 13.5 Watts. The hot water pipe exiting the top of the cylinder was uninsulated which would have likely contributed to additional heat loss.

We replaced the original cylinder with a special heat pump cylinder with a large surface area coil and thicker insulation.