Over-sized heat loss = over estimated design temperatures

Hardware Heatpump
1

I wanted to share some thoughts about where I think we are with heat pump system design in the UK as we start 2025.

There is a large performance gap between heat pumps that have been designed, installed and configured well (SPF H4 4.0 HeatpumpMonitor.org) and what most customers are likely to be getting (SPF H4 2.8 Electrification of Heat. The heat pump industry is diverse and while there are some amazing installers out there, the wider industry is struggling to deliver on this same quality and performance.

A significant factor in this difference is that the good installers know when to deviate from industry guidance such as CIBSE default air change rates and U-value assumptions in heat loss calculations. Whilst those who go by the book, inadvertently and arguably at no fault of their own, build in a lot of errors into system design that make higher performance results more difficult to achieve.

Overestimated heat loss = over estimated design temperatures and weather curve settings!

Over-estimated heat loss calculation is at the heart of the problem but this is not primarily about over-sizing the heat pump unit itself as many of us (myself included) have focused on.

Arguably the most important issue with over-estimating heat loss is that it over-estimates the design flow temperature that the system needs to run at.

If a system design is based on an over-estimated heat loss, it can appear that even with upgraded radiators that a 50C flow temperature at design temperature is required.

If this design temperature is then used to set the weather compensation curve, as one would expect, the system will run hotter than it needs to. It will run up to this high temperature, a room thermostat will then likely turn the system off for a period before the cycle repeats again. This looks to be a common issue with systems that were installed as part of the Electrification of Heat trial that achieved an average performance of SPF 2.8.

A simple case of lowering the weather comp curve would allow the heat pump to modulate back, run at lower flow temperatures and deliver better performance. Brands with degree minute algorithms such as Vaillant/Viessmann/Nibe will handle mild weather cycling pretty well even if the unit is over-sized.

In reality with accurate heat loss calculations itā€™s relatively easy to design systems to run at design flow temperature of 35-40C, even with radiators, and it is these design flow temperature and importantly associated configured weather compensation curves that deliver on the high performance results we see on HeatpumpMonitor.org.

This issue of over-estimated heat loss leading to over-estimated design temperatures and weather compensation settings can be solved by either 1. Improving heat loss calculation accuracy or 2. Heat pump manufacturers adopting automatic optimisation of weather comp curves. 3. Customer education on tuning heat pump settings, why low temp is better and how to avoid too much zoning etc. Or perhaps a mix of all threeā€¦

Right sizing the heat pump

Right sizing the heat pump is of course an associated factor, but is complicated to a greater degree by variations in how different makes and models of heat pump handle mild weather modulation and cycling and what their real world maximum capacity outputs are during defrost conditions. It seems that a 30-40% over-sizing factor on an accurate heat loss gives good results, but adding an extra factor on top of an over-sized heat loss is less than ideal.

Does the above seem like a reasonable summary of where we are? Any thoughts welcome!

Simple Heat Loss tool:

Explore how default CIBSE air change rates and other heat loss assumptions in pre-2000 properties can give the false perception that high design flow temperatures (~50C) are required even with radiator upgrades when 40C flow temperatures are not actually that hard to achieve in reality, see: Simple heat loss tool (40C or lower being the design flow temperature that really delivers on SPF H4 4.0+).

Related reading

15 Likes
2

Iā€™m not in the UK but your analysis is spot-on in my opinion.

This really hit home. The arguments I had about needing a smaller heat pump that were shut down by ā€œwe have to do it according to DIN xyz otherwise weā€™re liableā€ā€¦

7 Likes
3

One of the biggest things for me is the ability to see what is going on.

This is pretty much lacking in all heat pumps as standard, yet they all have the information you need.

I donā€™t see how anybody can operate a heat pump efficiently without feedback from changes you make.

A lot of problems can be fixed by making appropriate adjustments, itā€™s a lot easier when you can see what is wrong and what difference the changes make.

5 Likes
4

Yep. This is exactly whatā€™s happened here. Started with 50/-3 - 25/22 resulting in overheating and hitting the thermostat top end limit. The actual weather compensation points ended up being 42/-8 - 25/13, some 10c lower. Iā€™m in discussions with the installer now on things like not taking into account a large wall of 50mm IWI and the cold snap a couple of weeks ago ā€˜produced heatā€™ kwh figures revealing the actual heat loss to be quite a lot lower than calculated, of the order of 2x oversized.

Itā€™s almost like they took the EPC, divided space heating kWh by 2 and added 20% to come up with the figure which coincidentally is almost exactly the size of my heat pump.

3 Likes
5

I think youā€™re on the money with your analysis Tristan.

Alongside the loss in efficiency that obvious comes from a higher flow temperature due to an overstated heat loss there are two other effects:

  1. The system cycles on the room thermostat. This isnā€™t really a problem from an efficiency perspective but its a lost opportunity. Leaving aside any efficiency benefits, running your heating where the radiators produce a constant heat is just nicer to live with than the traditional blast of heat for an hour or two and then waiting for the room temperature to drop before doing it again. When talking to other people about my heat pump one of the things I try and get across is that it is simply better than a gas boiler.
  2. By overstating the heat loss you may force a home owner to go down a more disruptive and expensive retrofit than what could have been done. Reducing the heat requirement by 30-50% might mean that existing pipework and radiators are suitable, vastly simplifying a switch from boiler to heat pump. Yes, replacing everything will allow the lowest possible flow temp and most efficient install but we also need to be pragmatic about what is realistic for mass adoption of heat pumps into existing UK housing stock.

I think one of the key issues in heat loss surveys producing overstated results is that there is a cumulation of errors all in one direction. The U values for things like windows & insulation err on the side of caution, the surveyor errs on the side of caution when assessing the house, the MCS methodology errs on the side of caution by excluding solar gain and heat produced by living in the house, etc. It then errs on the side of caution with their selected desirable room temperatures, and again with air change rates. Then finally once you have the heat loss (which is already overstated) you then select the next available heat pump size up.

Iā€™m not sure what the answer is. Does the heat loss methodology get improved to produce a more accurate result, or do we just learn to apply a scaling factor to it?

I did my own heat loss calculation using the MCS spreadsheet when designing our system. The end result was very similar numbers to what the installer produced. They probably overstate the heat loss by 30-40% from what it seems to be at MCS specified room temperatures, and ends up being more like double what our actual heat loss is for our chosen room temperatures. Having said that the theoretical calculation correctly predicted the relative room temperatures. The rooms we wanted to warm up easily do, and the rooms which were a bit marginal on the heat loss spreadsheet are the ones we struggle to keep warm at very low flow temperatures.

The crux of Tristanā€™s question though is how do we make the best of an installed system, if the system is a bit oversized and a weather curve that is a bit steep. I donā€™t think more accurate calculations is the way forward, thereā€™s too much reasonable variation in peopleā€™s desires. I think the answer is moving towards a system which works out the best way to drive the heat pump for the house it is in. Some heat pumps have auto adaptions or WDC modulation based on what the inside temperature is doing which go some way to achieving this. There are other people on this forum, and companies like Homely, working on similar systems where there is a specific control system which then interfaces to the heat pump.

6 Likes
6

Is there ever a downside to larger rads?

7

(This was written before Iā€™d read @Jonathanā€™s excellent response, but not inconsistent I thinkā€¦)

It is also possible to exaggerate just how cold it gets in your local area, and for you (or your installer) to therefore base heat loss calculations on an over-conservative ambient temperature.

My installer did his calcs based on -2degC (this is central southern England, so on the face of it a reasonable choice) but out of interest I just checked the stats from my local weather station EGDM (using https://www.degreedays.net/) for 2024 ā€“ a fairly typical year Iā€™d guess).

Over the whole year, there were just 5 days that saw -2degC or lower, and only a single day where it occurred for more than 1 hour (1.6h to be precise). If I put in 0degC, EDGM saw 16 days that were at or below this, but only 3 where this was more than an hour (the worst day was 3h).

Given 1) your houseā€™s thermal inertia (which smoothes out temperature fluctuations quite effectively), and 2) the ease of putting on an extra layer of clothes for a few hours each year, it might be worth insisting that heat loss calcs be based more realistically. If this saves you a couple of kW on your heat pump sizing (which also reduces the likelihood of inefficient cycling), you might even be able to justify buying a fan heater for your living room (just in caseā€¦)

4 Likes
8

There can be I think.

I feel that my radiators are too big for my Daikin heat pump, they take it outside its operating range in terms of flow temperature. I want to go lower than the heat pump is happy at.

1 Like
9

Thanks all! reassuring confirmation :smile: I feel like Iā€™m at my wits end sometimes seeing what customers are getting, while knowing what they could be getting. Iā€™ve also just had an interesting call with an installer looking at a system that is apparently meant to be running on the manufactures auto optimisation settings but it doesnā€™t seem to be doing much optimisation, itā€™s just running the system at 55C (COP 2.5)!!

2 Likes
10

Basically, no, not for practical purposes. The larger the radiators the lower you can get your flow temperature and therefore the more efficiently your heat pump will run.

A downside would be where some radiators are much bigger in some rooms than others (relative to the room heat loss). In that case it becomes hard to balance the system, the rooms with the big radiators get too hot and those with smaller radiators struggle to get up to temperature. You can manage this by balancing the radiators using lock shields or TRVs but that undermines the benefits of the larger radiators. If I were advising somebody looking at new radiators, it would be to make sure that the rooms they are most concerned about being warm most of the time (living room, study, kitchen?) have the largest radiators relative to room size. This is what will determine the lowest flow temp that the system can be run at.

On the subject of radiators, in the cold patch a couple of weeks ago (daytime high of -2C here) I was experimenting with very low flow temperatures <35C. The results of that doesnā€™t tally with my predictions of the heat that the radiators should be able to put into the rooms. My calculations, using Stelradā€™s correction factors, gives a theoretical total radiator output of <2kW at that sort of flow temperature. However the heatpump was steady state running at nearer 3kW heat production. Assuming the produced heat largely goes into the radiators, this must mean that the calculated heat outputs of the radiators are quite a way off.

Has anyone else found the same result? (or a different one?) I havenā€™t done any experiments at higher flower temps. If accepted formulae for radiator outputs for a given flow temperature are also underestimates then this would be a further factor in Tristanā€™s analysis that ā€œcalculatedā€ flow temperatures are far higher than they need to be.

2 Likes
11

Without some very strict oversight you will definitely see heat pumps set to static 55C with WC disabled in order to reduce callbacks.

1 Like
12

Ok I understand. Lucky most modern pumps modulate quite low really and I think on a national scale itā€™s less of an issue than the base COP.

13

To put some numbers against it. I used the MCS Heat Pump calculator spreadsheet to do a heat loss calculation and radiator sizing analysis, Iā€™m confident that the methodology was followed correctly and the correct data entered.

The spreadsheet produces a heat loss of 5.3kW @ -3.9C, this requires a flow temperature of 53C.

If I re-run the spreadsheet with the actual temperature we keep our rooms at its a heat loss of 4.7kW at -3.9C, requiring a flow temperature of 48C.

Experimental data is that heat input required is actually around 3kW at -2C and that heat can be produced at a flow temperature of 35C.

During the recent low temperatures we had -6C at night, rising to a little below 0C in the day, total produced heat in 24h was 75kWh for an input of 24kWh (includes defrosts). We let the internal temperature drop by 2C or so overnight.

So the actual heat loss is about 2/3 of the calculated heat loss, and the actual flow temperature is at least 10C lower than calculated. The difference is even larger if you take the original MCS estimate with the higher room temperature assumptions.

14

Michael de Podesta argues the case for Deliberately under-sized heat pumps. | Protons for Breakfast

3 Likes
15

Should just make them modular and stackable like batteries.

1 Like
16

Daikins are not happy below 30c, particularly when it is warmer outside.

I was hoping to run at sub 30c when it is warmer but the heat pump uses just as much electricity as it does running hotter, itā€™s less efficient and unstable.

I actually turn it up a bit when itā€™s warmer outside!

3 Likes
17

Great write up @TrystanLea
It aligns with my experience.

  1. Over estimated heat loss was at the heart of my problem. 9kW estimated vs the reality of 4.5kW resulting in an over-specced heat pump. Coupled with a design temperature of 50C it meant they didnā€™t touch the radiators, however the flip side of that was most of the rads are over-sized for lower heat loss. If only I had a heat pump that could modulate low enoughā€¦
  2. There was the quality of the installation (or lack of) and the attempt to install a fit and forget system with a 50C flow temperature and regular thermostat, hoping I would just accept hot radiators and not look at the elec bill.
  3. Then there was the choice of heat pump. The ā€œBeastā€/11kw Daikin just doesnā€™t have a good modulation range and canā€™t handle low flow temperatures with radiator circuit. It only modulates down to 1kw input before cycling starts and efficiency tumbles. Had a different brand of heat pump been fitted I suspect it could have handled mild weather cycling better. Viallant and Mitsuā€™s did do much better on the EoH trial (not that Iā€™m bitter :wink: ), although still not great. Lowering the weather compensation curve is not necessarily the answer when you have a significantly over-sized heat pump. And youā€™ll struggle to know how over-sized you are without proper monitoring in place.
3 Likes
18

I feel that as an academic exercise under sizing a heat-pump could make sense. But practically it is a bad idea.

People struggle to predict the correct heat-loss for a house, so correctly under-predicting the heat-loss sounds like it recipe for disaster. Getting it wrong by a few kW might be the difference between Ā£100 a year in backup heater costs and Ā£500.

4 Likes
19

Oh as a side note itā€™s just more evidence for actual testing e.g. EPCs become actual measurements rather than useless exercises.

1 Like
20

The rules for the BUS grant is also part of the problem. Using the fabric-based heat loss estimates, and marrying them with HDDs for my local area, I reckon one possible heat pump (Vaillant 7kw) would run out of puff for 2% of the annual heating plus hot water demand (I have made allowance for defrosts reducing actual output below the published levels in cold weather). A backup heater would easily and cheaply cover that 2%. But that 2% is enough to make the HP ineligible for the full BUS grant.

To get the grant I will have to increase the HP size- even though the heat losses are probably over-estimated. And 21deg is a higher temp than we will ever use.

Fortunately we are looking at high-end pumps with decent WC and low-output modulation, so we should be able to cop(e) (ahem). But the rules of the BUS are becoming a problem.

1 Like