Those Phase 2 trials were April 2009 - April 2010 and heat pump technology has advanced a lot in the intervening years. I regard my GSHP, installed in 2015 as ‘old school’ because it doesn’t have an inverter drive for the compressor and is only ever On or Off - but that was standard for GSHPs in 2015. The use of older technology in those pre-2009 systems will account for at least some of the differences in performance.
Thanks @MyForest ! there is so much data available for that study! including a 1GB+ download!
This detailed analysis report by UCL that was part of the study is really interesting, I’ve only skim read the head line results so far: https://doc.ukdataservice.ac.uk/doc/8151/mrdoc/pdf/8151_decc_rhpp_detailed_analysis_report.pdf
The summary of results are given on table 1:
and a comparison with the EST trial in table 2:
That’s a pretty decent sample size and such poor results even for the H2 boundary! 
A quick look at the mean for the 23 heat pumps on our heatpumpmonitor.org page, suggests at least a SPF H2 boundary mean of 3.56 (with a range of 2.6 to 4.7). Median of 3.6.
We dont have a full year of data yet for most heat pumps and Id expect a reduction in COP over the summer given higher standby and higher proportion of DHW, so it will be interesting to see where the figures come out at.
2 Likes
Yes, Im not totally convinced yet that it’s better technology though happy to be proven otherwise. The RHPP trial a couple of years later seems to have worse H4 boundary performance for the ASHPs, though better H5 boundary…
I wonder if most of the difference between what we are seeing and these trials, is that we are a selection of highly engaged users? Taking care to understand and drive the heat pumps in such a way as to get higher performance?
I wonder if there are any more recent trials? say 2017-2019?
1 Like
Looking at the RHPP trial ASHP H2 Histogram, there are a few heatpumps getting SPF H2 4+ and a few more around 3.4-3.5
1 Like
Perhaps something to be added to the dashboard - a “How much do you tune your system” 1-5 scale perhaps.
This seems to be borne out by the dashboard - at first sight it looks like those engaged in tuning get better SCOP (which is not unreasonable).
Perhaps also an installation rating “How would you rate your installation” - subjective but will probably give an idea if the issue is the system itself or how it has been installed/configured.
2 Likes
Yes that’s a good idea.
Found another trial carried out by Mitsubishi themselves, published 2013 and audited by the EST that achieved an average SPF H4 of 2.9 across 23 sites, though I cant find much more detail on it apart from this press release: https://www.aspenrenewables.co.uk/file_upload/ecodan-trials.pdf
This report for/from RECC published in 2021 seems to give a good overview of the different trials: https://www.recc.org.uk/pdf/performance-data-research-focused.pdf There seems to be another more recent trial by Ofgem, based on RHI / MMSP data, I think the date is ~ 2018/2019 :
The Ofgem dataset results do not all conform to a specific boundary. Most are likely H2 but some may well extend further e.g to include the circulation pump and indoor controls, Im not sure about the immersion heater. Looks like an average of ~2.7 for ASHP’s.
1 Like
I think for me, my main motivation for heatpumpmonitor.org is to try and work out what would it take to shift the mean into the 3.5-4.0 range at least for the H2 boundary and perhaps the 3.3-3.8 for the H4 boundary? I calculated using my 2020 data that if I used my immersion heater a more normal amount e.g once per week for legionella protection that would have dropped my SPF H4 that year from 3.91 to 3.86.
I guess the other side of this is as others have pointed out, there’s also a risk of getting too fixated at trying to get the highest SPF’s/COP’s at the expense of other goals such as using less electric or using electric at cheaper times of the day, or making use of PV divert directly etc…
1 Like
Surely that goes hand in hand - higher COP means less primary consumption no matter where it comes from?
I did a theoretical calculation for our house that suggests that if i was to zone upstairs and downstairs separately that could save electric overall even though there’s a small reduction in COP due to needing higher flow temperatures to compensate… I haven’t tried it in practice to confirm but only talking small % points…
2 Likes
The other scenario is that it might make sense if you have underfloor to charge it up at a higher flow temperature overnight at a time that’s better for the grid. That might even use less primary energy at a lower COP due potential for higher % from wind vs gas generation at peak times…
2 Likes
I’d certainly be highly sceptical that technology advances account for all the differences between that 2009-10 trial and what we’re seeing today, but inverter drive for the compressor motor was surely a game changer (since it lowers the average flow temperature, which we know improves CoP). Not sure if we have any ASHPs that are non-inverter on HeatpumpMonitor currently (do we even record that?)
3 Likes
I built a deeper slab for just this theory. I don’t think the theory holds up in practice. I find it still loses heat too rapidly and the better solution is to manage the room temp within small differences in temperature.
4 Likes
https://usmart.io/org/esc/discovery/discovery-view-detail/5325ef18-9cd1-493c-beae-e278d8998400
Nearly 750 heat pumps were installed as part of the BEIS Electrification of Heat project. You can see from the above dataset that there is a variety of different systems installed, each of these will have MMSP monitoring. Fingers crossed that this data becomes available for analysis.
2 Likes
Yes looking forward to the results from that one!
I’ve created a bar chart of the SPF’s of the system’s on heatpumpmonitor.org so far, listing them as H2 boundary for now until we can be sure otherwise, I know most will include the central heating pump consumption as well which usually comes under H4 - but then not the immersion heater.
2 Likes
I notice that some of the feeds occasionally have missing data in 1 or more feeds. Is there any way to capture periods of missing data and excluding those from the averages, or if this is too difficult have a % data integrity variable somewhere?
1 Like
Hi Trystan
I agree re being fixated on SCoP (hence my comment elsewhere about comparing actual historic gas use to heatpump electricity use). We mostly keep our houses warm for longer with our heat pumps - somewhere there is an optimium balance between turning off/setting back and achieving the highest SCoP in terms of cost, and likely carbon emissions.
There is another BEIS (no more!) Funded project collecting heatpump operational data. If i spot it published I’ll put a link here.
.
4 Likes
Thinking about the different System Boundaries following the Eco Home Lab session earlier. I’m reporting H3 currently (well, mostly; actually including E(bt_pump) too - but not including E(b_pump) which is downstream of my Low Loss Header and not currently included in any metering).
Would we plan to standardise on reporting against one boundary, maybe H3 or H4, or would it be better to log the different consumption readings separately so we could report on any one of H1 through H4 as required?
I do have a separate sub-meter on the immersion heater in the DHW tank (which is almost exclusively used for excess solar PV diversion) and NIBE Uplink reports E(Aux) so I could subtract that from the metered heat pump supply and report it to a separate Feed.
1 Like
On my system, I’m reporting H2 plus E.B_pump, or H4 minus E.Immersion, because the electric meter feeds the heat pump and all the pumps but the immersion is on a separate meter (and barely used).
I could probably estimate the consumption of the pumps (120W) and use that to compute consumption at the H2 boundary. This is made tricky by the primary pump being in H2, and the CH & DHW pumps in H4 (standard config for Mitsubishu pre-plumbed cylinders). Probably easier to add immersion and call it H4 (at the expense of COP).
2 Likes
On my system the boundary also moves!
I have my original solar cylinder in place. In theory the coil is nowhere near big enough, but given that I have solar thermal and a PV divert i couldn’t see any point in spending around £5000+ on installing a new tank.
The way I have it set up is that the ASHP runs on DHW in off-peak hours and early evening (not good from a ADMD perspective, but, as it turns out, quite useful from a Demand Flexibility payment perspective!). Despite the small coil, the ASHP will get the top of the tank to just about 50C, though I have the set point at 46C because the ASHP temp sensor is below the hottest point and it does sometimes get stuck trying to get to 50C!
At night, off-peak, after the ASHP has done its best, I by-pass my solar diverter and finish it up to 50C with the immersion. I don’t bother for the afternoon run as i dont need the volume of water then. Once a week I take it to 60C.
The immersion puts about 1kWh in, but given that I mostly don’t need it at all in summer, and the ASHP CoP wont be very good for that last bit of heating, I imagine the additional cost per year is not more than a few pounds. This is where a bit of flexible thinking about systems could save vast sums in installation costs! My whole install was under £7k, it would have been a more traditional £12k with a new tank, which all the installers said i needed.
I do monitor the grid immersion use, so could add this in, but it wont really add up becaise of the solar divert/thermal, hence comparing to my historic gas use rather than the EPC number.
R
2 Likes