ASHP Design considerations

[JohnMo]

Subject comes up a lot, so thought I woul start a thread for other to add too.

 

CoP is the Coefficient of Performance, how much energy you put in, compared to how much energy you get out. An average system is going to get 3, a good simple open system gould be near 4.5 to 5 over the whole year (SCoP or seasonal coefficient of performance). 

 

Ideally size heat pump to match heat demand, if you size to complete all heating duties in 22 hours that will leave you 2 hours for domestic hot water heating (your cylinder). So as way of example, if design heat demand is 4kW, you need to deliver 96kWh of heat. Doing that in 22 hrs is 96 divided by 22 or 4.4kW.  The 4.4kW is the size of heat pump you would need at your outdoor design temp. So you would size at the nearest size of heat pump above that figure. If however you intend to batch charge a floor on cheap rate, calculate how many hours you cheap rate is, lets say 7 hours. 96kWh divided by 7 is 13.7 hrs - so a pretty huge heat pump. So you need to use common sense with sizing, will that to others to expand on.

 

Here you need to look at data sheets and figure out output at -3 (if that is your design temp), if its rates above 4.4kW its fine, below that its not fine.  Do not use the manufacturers model size as that information is generally meaningless.

 

System volume - ignore what the manual states as minimum volumes, as that is only to support defrost duties generally and is not suitable to manage cycling of the heat pump compressor. For this you need to go back to the manufactures datasheets and find the minimum output of the heat pump. If you then multiply this by 30-40, this will give you suitable volume of water in litres for the heat pump to work with.  If your system doesn't meet this volume, you need to add either a volumiser, ideally on the return side of the system, prior to DHW return tee, so it just heated in central heating mode. 

 

System design, you have two routes, open systems (so one zone), you can use the advice above on system volume and using a volumiser.  Or multiple zones with thermostats. Here your system volume is dictated by the smallest volume as a result of the various zones closing. But a volumiser is no longer an option as you also have to preserve a volumetric flow rate the heat pump demands, a volumiser is not suitable for this. This brings in a buffer, so you now size on the same basis 30 to 40 times your minimum output, but use your smallest zone to size against. Thes can be plumbed as 2, 3 or 4 port, each with its advantages and disadvantages.

 

Disadvantages of buffers are

1. Additional pumps are always need due to hydraulic separation between heat pump and heating system.

2. Distortion of the heat flows, due to mixing within buffer, so for a given heating flow temp your heat pump is asked to deliver a higher flow temp.

3. Additional costs for pump etc.

4. A drop in CoP, from distortion and the electrical power to run the additional pump or pumps.

 

Filters and strainers

You need something to catch debris in the system. A cheap easy option is a ball valve strainer, this gives an isolation valve and strainer in a single package - but with a substantial pressure drop and therefore a flow rate drop. So best advice is to avoid and use a good heating system filter.  If you have radiators you will need magnetic filtering as well as debris. Not all filters do both.

 

Generally size piping to give as low a flow pressure drop as practical, this allows the heat pump to flow as much or little as the controller needs. Low differential between flow and return temps, gives a higher mean flow temperature than a high dT. You get this higher mean temperature for same flow temperature, the flow temp governs the ultimately achievable CoP. Again you need to go in to the manufacturers data sets and get the circulation pump curve and work with this to design your system.

 

Flow temperatures, the lower the flow temp the higher the CoP, hence the statements that are banded about with reference to low and slow.

 

Cylinders, can be internal plate heat exchanger and pump or internal coil. Internal coil is least complex, but to get a good CoP you need a big a coil as you can get.  For me 3m2 is the smallest I would use, Heat Geek cylinder have a 6m2 coil. 

 

Your system can be very simple or very complex, but generally the more simple the more efficient it will be. A simple system is 

1. ASHP

2. 3 port diverter

3. A cylinder for DHW

4. Radiators or UFH

5. Maybe a volumiser

 

Other now contribute

[akjos]

Great write-up and very clear. Easy to follow for anyone designing a new system. 
One thing I’ll add, the heatloss calculation is such a crucial piece of information, so really make sure it’s correct. Whoever does it, needs to spend at least a few hours analysing all parts of the building.

[JamesPa]

Excellent.  Here is my contribution (I may add more)

 

Controller:  Generally use only the heat pump's own controller and under no circumstances use 'smart thermostats'.  Alternatively use an add in controller specifically designed for heat pumps, (currently only Homely, Havenwise or Adia). These interact with the heat pump bus and so only work with some heat pumps, but do so in. A way specifically tailored to heat pumps.  Anyone selling you a controller, or even just a thermostat, claiming to work with all heat pumps is (at least at time of writing) selling you one you shouldn't buy!

 

If you are going to use only the heat pump's own controller and want to program different temperatures at different times of day, you need a heat pump that does this by shifting the weather compensation curve.  Some (for example, but not limited to Vaillants and the R32 Ideal models, maybe also the R290 Ideals) do this, others (eg Samsung) don't and rely, in effect, on a crude on/off thermostat function if you want programmable temperatures, which compromises performance. 

 

That said it's worth asking whether you really need to vary temperature at all unless you are doing it to take full advantage of the more complex time of use electricity tarrifs such as agile or cosy.  In this case you probably need Homely, Havenwise or Adia (or a battery) anyway unless you enjoy tinkering with home made control systems.

Edited Sunday at 17:55 by JamesPa

[JamesPa]

33 minutes ago, akjos said:

One thing I’ll add, the heatloss calculation is such a crucial piece of information, so really make sure it’s correct. Whoever does it, needs to spend at least a few hours analysing all parts of the building

I would add to this:

 

If it's a retrofit, particularly in a house which has had fabric upgrades at various times, try to 'sense check' the surveyed loss against historical consumption.  Surveyors (even those who spend a few hours) often ignore invisible fabric upgrades, overestimate air change rate and even double-count some losses, which can result in a calculated loss more than double the actual, which can make a huge difference.

 

As a very rough sense check, take annual consumption in kWh, divide by 2000.  Do the same again but dividing by 3000.  If the calculated loss in kW isn't between these two figures, look more carefully and try to reconcile.

 

If you have half hourly meter readings for a few weeks or a season, you may be able to use these to get a much more accurate figure.  The method and caveats are beyond the scope of this article, but it's not difficult and can, depending on the circumstances, be very informative or even decisive.

 

If it seems likely that assumed air change rate is the culprit, get a measurement done.  It costs a few hundred pounds only and will also tell you where you might be able to reduce draughts.

Edited Sunday at 17:59 by JamesPa

[Dillsue]

If you want an accurate sense check on heat loss calcs borrow enough electric fan heaters to match the calculated loss, spread them round the house and heat the house with them for 24/48 hours during a cold snap. If you log the external and internal temps hourly you'll be able to see if the heaters can keep the house up to temp.

 

Our calculated loss at -2 was 4.6kw and 4.1kw of fan heaters kept the house just below design temp when it was 0/-1 outside. Gave me a good degree of confidence that the 4.6kw calculated figure was good to work with.

[JohnMo]

43 minutes ago, JamesPa said:

you need a heat pump that does this by shifting the weather compensation curve. 

Most heat pumps will do this as a second set point, would normally allow you to run radiators and UFH at different temps and a mixing valve in between. You harness this part of the system via a zero volt contact. This can via a smart relay if you must or a simple time switch. So it works by adding temperature to normal weather compensation curve. In practice with UFH, the heat added during an off peak window, will make the heat pump shut down for many hours to carry you through a high tariff period.

[JamesPa]

9 minutes ago, JohnMo said:

Most heat pumps will do this as a second set point, would normally allow you to run radiators and UFH at different temps and a mixing valve in between. You harness this part of the system via a zero volt contact. This can via a smart relay if you must or a simple time switch. So it works by adding temperature to normal weather compensation curve. In practice with UFH, the heat added during an off peak window, will make the heat pump shut down for many hours to carry you through a high tariff period.

Fair enough, but this is 'tinkering with home made controls' which most ordinary people won't want to do. 

 

Most people expect to turn a dial calibrated in degrees C to change the temperature!  Quite why some heat pump manufacturers haven't realised that is a mystery!

Edited Sunday at 18:24 by JamesPa

[JamesPa]

45 minutes ago, Dillsue said:

If you want an accurate sense check on heat loss calcs borrow enough electric fan heaters to match the calculated loss, spread them round the house and heat the house with them for 24/48 hours during a cold snap. If you log the external and internal temps hourly you'll be able to see if the heaters can keep the house up to temp.

 

 

Seems sensible.  The only thing I would add is to do it on a day when the previous day, or better still two days, was the same temperature and you were heating the house.   Houses have a lag of sometimes 24hrs and you don't want to be caught out by that.  Also if it doesn't end up close to the same temperature as it started (say within a degree), you may need to apply a correction to account for the house heat capacity (my house, for example, appears to have a heat capacity of very roughly 20kWh/degree, so a 1 degree change in temperature over 1 day would result in roughly 1 kW error).

 

(Ps the reason I am so conscious about sense checking retrofit surveys is that I had 2 done, both taking 3 hrs and both coming up with 16kW.  Actual loss is 7kW.  Difference due to ach, ignored invisible fabric upgrades - even though I made a big point of telling them about them - and double counting.  I'm clearly not alone given what I read hear and on other forums, particularly rhh).

Edited Sunday at 18:58 by JamesPa