We partnered with Havenwise to analyse a full year of real-world UK heat pump performance data.

The analysis compares how Mitsubishi Electric, Samsung and Vaillant systems behave once they’re installed in real British homes.

Looking at the data side by side raises some genuinely interesting questions about performance, measurement and what “good” actually looks like in practice.

Just an evidence-led look at what the numbers do (and don’t) tell us.

Full video here:

@hcas, I've received a few emails and message in the past week asking when Daikin will be available on Havenwise as it's been listed as "coming soon" for quite some time. Any updates would be helpful.

@editor Oh I am glad we are Vaillant and Mitsubishi Experts using Havenwise, great video 

Interesting video, with some intriguing findings, particularly with regard to the Mitsubishi pumps potentially over-reporting consumption.

One slight anomaly though is with regard to the discussion on flow temperatures, and having a flow temperature that jumps around between the 30s and 50s being described as inefficient.  Havenwise control, as far as I've analysed it, actively fluctuates the flow temperatures across this range when providing it's heating in 'short, sharp blasts'.  It probably is inefficient, in terms of COP performance, relative to stable, lower flow temperatures, but Havenwise algorithms (at least as far as my sample of one heat pump goes) chooses to run the pumps in this way, often with the compressor frequency resembling that for DHW heating for short periods of time.

I've discussed it previously in a post on my 'diary' blog https://renewableheatinghub.co.uk/forums/postid/53002/

It's a long post, but included within it is a snapshot of the target flow temperature being set by Havenwise over a typical day, which I've copied below for easy reference.  As you can see, the flow temperature varies across the day to a peak of circa 55°C, for heating.  The overall impact is a 'net' flow temperature in line with the averages that you report in the video, but the method of achieving this isn't via a linear, low and slow operation, as you might expect.

This isn't a criticism of Havenwise, by the way, as this operation is aligned with longish periods of the heat pump sitting idle so it doesn't run like this constantly throughout a day.  It's intended to provide a bit of insight from a user perspective.

So.. I’m currently on holiday. The 1st few days I lowered the target temperature to 18c.

The daily cost of this was under £1 a day.

I decided to change the setting of the heat pump to holiday mode. As the graph shows it was clearly a wrong decision as the daily cost shot up by over 300%.

have to say I’m rather surprised by this. Is there anything that I could be missing here?

@westkent, I don’t use Havenwise, so this is just a theory on why Holiday Mode pushed your daily costs up so much… when you manually set the target to 18C, the heat pump only ran as needed to hold that level, keeping runtime low and costs mostly under £1. That worked brilliantly for the first few days as you’ve described.

Havenwise Holiday Mode (I think) drops the target to 16C for frost protection and only activates heating if the house falls below it. On paper, a deeper setback should save even more energy than 18C, but your graph shows the opposite happened.

What I suspect is going on is some extra automated behaviour kicking in under Holiday Mode, perhaps slightly more frequent circulation to keep things from freezing, a periodic DHW anti-legionella cycle (even though I think this one is unlikely) or the app’s optimisation applying a different flow temperature curve or pump logic that’s less gentle than your manual setting.

Another factor could be timing: looking at the dates (Jan 16-22), the big spikes line up with what might have been a colder snap mid-period. Maintaining even 16C during that would naturally cost more than the milder earlier days at 18C. Just speculating here.

Either way, the manual approach clearly came out on top for a short holiday like this. For longer trips, Holiday Mode is still handy for proper protection, but next time it might be worth sticking with a manual setback or just monitoring the app closely for the first day after switching modes.

Curious to hear how it goes if you experiment again, and wonder if other Havenwise have experienced similar things in Holiday Mode.

Posted by: @editor

↑

We partnered with Havenwise to analyse a full year of real-world UK heat pump performance data.

The analysis compares how Mitsubishi Electric, Samsung and Vaillant systems behave once they’re installed in real British homes.

Looking at the data side by side raises some genuinely interesting questions about performance, measurement and what “good” actually looks like in practice.

Just an evidence-led look at what the numbers do (and don’t) tell us.

That is quite concerning as I have a Samsung Gen 6 in our current house and have just had a Gen7 R290 installed in what will be soon be our new home.  I had the choice of Vaillant or Samsung from the plumbers with the Vaillant being quote over £1000 more. I choose the Samsung on cost and after reading quite a lot on the Open Energy Monitor forum from people who have them and have taken the time to set them up well and I expected it to be similar to run to any other well set up HP.  I have a Homely installed on the Gen 6 and intend to use the same in the new house. I am hoping that Homely will run it as efficiently as possible but from your video that isn't necessarily the case, or at least not with Havenwise.

@editor Good question! Should be very soon, as in the coming weeks.

If people download the Havenwise app, create an account, and select Daikin as their manufacturer, they are automatically added to our wait list. And they'll receive a message when they are invited to complete the onboarding. 

@westkent I'd suggest sending a message to our customer support team via the button in the app. The explanations from Mars are definitely plausible, and the team can confirm by looking at the data from your account.