We’ve published many posts and videos about our air source heat pump, and running costs have never been more important than they are today with energy costs hitting all-time highs. With this in mind, we’ve conducted a real-life experiment focused on the efficiency of our heat pump.
Our air source heat pump was a retrofit installation that replaced the previous homeowner’s oil boiler as our primary source of heating and we inherited all the associated pipework and its limitations which has contributed to some of our heating inefficiencies.
Our current heating circuits comprise of three underfloor heating zones, 12 radiators and two towel rails.
Due to the way our system was installed, our weather compensation doesn’t work as effectively as it should. It was also never enabled by the installer during the handover so we’ve always resorted to manually changing the flow temperatures based on the weather.
The installers slapped the main controller for the heat pump next to the hot water cylinder in our utility room cupboard. At any given point in time it’s over 27C in that cupboard when the heating is on, so it’s not ideal in giving the pump accurate indoor temperature readings.
In addition to the main controller, we have a smart Salus thermostat to switch on the radiator circuit and Heatmiser thermostats to switch on the UFH circuits.
Since we’ve been changing flow temperatures manually for years, we decided to do an experiment this autumn while temperatures aren’t too cold in the evenings to see if we could save money by dropping and playing around with flow temperatures.
As part of the cost cutting measures we’ve only left one underfloor heating zone on (in the kitchen/dining room) along with five radiators and the towel rails to only keep the rooms we’re using regularly nice and cosy. We also only ran the heating from 16:00 to 8:00 on set point flow temperatures. Yes, we know that’s not prescribed as best heat pump practice, but that’s part of the experiment.
We ran the heating overnight at flow temperatures of 45C, 40C and 35C, followed by full days (24 hours) running on weather compensation. On most evenings temperatures have dropped to between 6-9C while we conducted the experiment so it hasn’t been too cold, with daytime temperatures ranging from 10-16C.
As some added background, most heating engineers we’ve spoken to have categorically told us that it didn’t matter how many rads or UFH zones we had activated because our 18kW heat pump would always deliver the required heat to keep all rooms warm. While the heat pump software shows that at 45C flow temperature our return is 40C (that’s a delta T of 5) with every rad and UFH zone on, rads don’t all feel the same temperature and some rooms struggle to get to temperature. After investigations, we’ve put this down to dodgy plumbing.
However, with just one UFH zone on with five rads, within 30 minutes all rads are very warm at 45C flow temperature. Dropping the flow temperature to 40C didn’t affect things too much – the rads were a tad cooler, but rooms were still cosy and warm.
At 35C, the rads and UFH were notably cooler. Rooms took much longer to warm up (as expected) but once warm were fairly comfortable.
As soon as we activated more UFH heating zones and rads, we saw a drop off in the overall warmth of rooms at flow temperatures of 40C and 45C. This clearly shows that despite the heat pump producing the required heat, and with return temperatures of 35C and 40C respectively, something in the pipework is adversely affecting our heating system’s performance.
We kicked the experiment off at 45C flow temperature. The graph below shows our total electricity consumption via the SolarEdge app, and this accounts for all our consumption with notable spikes between 8:00 and 22:00 for things like running the dishwasher and oven, and for hot water. But the notable periods to look at from a heating perspective are from 22:00 to 8:00 when we’re not consuming power and temperatures outside are starting to drop.
On average, we’re running at 500W per hour to power all the various things we have in the house, which is higher than average, we know, but we are working on reducing this.
In this first graph, the early morning and evening spikes you see are the heat pump running at 45C, and we’re regularly breaking through the 3kWh mark (touching 4kWh) which is largely the heat pump doing its thing.
In the graph below, the heat pump is running at running at 40C.
In the next graph we have the heat pump at 40C in the morning, dropping to 35C for the evening session. The spikes in the evening are longer and flatter (to be expected with the lower flow temperature) but the electricity consumption spikes are generally higher, and we actually used more electricity that evening than on the night we ran at 40C. That was odd.
So 40C appears to use the least amount of power on a set point, and the house is lovely and warm (with one UFH zone and five rads running) and in this set up we can keep the heat pump off during daylight hours because that’s more cost effective than leaving the heat pump running during the day when there’s sunlight around (leading to solar gain), which is contrary to how a heat pump should run, but the graph below proves that. This isn’t something we’d get away with December though.
By comparison, we then ran the system in weather compensation all day (1.8 heat curve) and here’s what our daily consumption looked like. Much better. And that’s with a dodgy set up. The average flow temperature was around 35C-37C with energy consumption down and rooms feeling very comfortable. More on this in a little while.
The tangible difference between 45C, 40C and weather compensation
Weather compensation, even on our poorly designed system, is the most efficient means of running your heat pump. There’s no denying that, but I can see why some people are out off by it and don’t give it a chance.
Since upgrading the rads in our coldest rooms to Stelrad K3s (which are amazing and massively recommended) when running our heat pump or HVO boiler at anything over 40C makes those rooms hot. You’re probably thinking, “well, duh”.
But I find it’s uncomfortably warm at times, so we’ve had to set the temperature on the Eve smart TRVs to regulate the rooms so that they don’t become saunas. There’s absolutely nothing efficient about that.
With weather compensation, the flow temperatures in late October have ranged between 32-38C. When you touch the radiators or towel rails, they feel warm. At 45C, they are hot. At 36 or 37C they feel extremely comfortable.
Each and every room (radiator or UFH) has comfortably maintained 21C (our target temperature). We’re not heating our hallway, so it’s around 19.5-20C, and when you enter the TV room it’s just a perfect temperature, because heat is gently warming the room and maintaining it all the time.
It’s not chucking out heat and you’re not getting blasted by the rads. So it involves a mental switch and approach, and giving it chance. It’s very tempting to try it for an afternoon and just think it’s rubbish, and thinking it’s not going to keep you warm because you can’t sense ‘real’ heat coming from the rads.
Gentle heat leads to gentle bills. And as the weather changes and it gets colder, the weather compensation will ramp the heat pump up and chuck out more heating when needed, but with mild autumn temperatures, when you just need to get the chill out of the air, you don’t need to ramp up the heating and waste energy.
What is weather compensation?
This is one of the best explanations of weather compensation courtesy of Derek that he posted on the Renewable Heating Hub forums:
I suspect that most properties in the UK would fall in the 0.8 to 1.2 weather compensation range, in that for each 1C change in outdoor air temperature, the WFT needs to be increased or decreased by somewhere in the region of 0.8C to 1.2C for your particular home. A good starting point would probably be to set the curve to a WFT of 25C at 20C ambient and 50C WFT at -5C ambient. That gives a slope of 1, 1C change in WFT for each 1C change in outdoor temperature.
Auto adaptation or modulation control has now been incorporated in some systems, to add to, or subtract from, the weather compensation curve, if the indoor temperature deviates from the desired setpoint. But these only work correctly if the system is correctly installed, configured and optimised.
Weather compensation by trial and error
As mentioned, due to a lazy retrofit installation of our air source heat pump, our weather compensation is unreliable. But if the outdoor temperatures are constant, as they often are in autumn, you can save electricity.
To get it right requires a fair amount of trial and error. For us a curve of 1.6-1.8 works quite well in the autumn with 7-15C ambient temperatures. It takes a lot of time and patience to give each curve enough time to see if it’s working well. At 1.6-1.8, with outdoor temperatures at 7-15C, our flow temperature was at around 37C, and left all rooms with rads and UFH very comfortable due to the gentle heat. It was also significantly cheaper to run.
In our case, with curves around 1-1.2, the flow temperature dropped to 30C and below. It wasn’t too bad in the areas heated by UFH, but rooms were too cool that were served by rads.
So it’s a matter of tweaking things and patiently waiting for results until you hit your sweet spot. We generally let each curve have at least 24 hours to do its thing, while taking outside temperatures into consideration.
Air source heat pump recommendations
If you’re looking to install a heat pump, or are getting one put in, be adamant that your installer takes the necessary steps to ensure you can get weather compensation running. Insist that they show you how it works so that you can potentially tweak it during the handover.
Sadly, many installers don’t enable weather compensation because rads don’t feel warm and they don’t want the hassle of coming back to your property to explain or tweak things, so they just a set point.
The MCS (micro certification scheme) requires weather compensation to be available on all space heating systems but frustratingly does not stipulate how this must be set up and does not mandate installers to enable it.
But take it from us. Weather compensation works. Not only does it work, it’ll save you money because it’s extremely efficient and your house feels warm with out the sensation that heat is being blasted in through the radiators. It’s the most comfortable heat we’ve experienced.
In closing, if you have questions or need advice, please visit the Renewable Heating Hub forums where you can ask questions and get advice and feedback from homeowners across the UK that are running a variety of different heat pumps to warm their homes.
I would also strongly recommend that you down the MCS’s best practice guide for domestic heat pumps to familiarise yourself with the basics.