There has been a lot of interest recently looking at whether heat pumps should be left on all the time – and how this works with time of use (ToU) tariffs.

Over the last year I have had the good fortune to spend a lot of time learning about the real-world impact of mixing technologies where heat pumps are involved. Just before Christmas, my colleague William Jamieson and I started analysing the open-source data from the Electrification of Heat trial. This dataset has around 350 homes with high quality data.

I wanted to get to the bottom the best control strategy – not just for efficiency but also for the cheapest heating costs when using smart tariffs, solar and batteries.

The first task was to identify which properties in the EoH trial were controlled constantly and which were intermittent. We used ‘R’ to group the properties into either variable or constant heating patterns, using the heat pump flow temperature.

The tariffs used for comparison

For the comparisons we used the standard capped rate, Economy 7, Agile, Cosy, Intelligent Go and Intelligent Flux. The Agile cost data was sourced from the last 12 months. If you want a description of how these tariffs work, try this for Agile, and this for Intelligent Flux, and this for Intelligent Go. We included Economy 7, as there are many providers who offer these tariffs. It should be noted that ‘Go’ is an EV tariff – so won’t be accessible for everyone.

This is not supposed to be a recommendation of any one company’s tariffs, but Octopus publish a very useful API which made it useful for our work – and they should be commended for this open approach.

Background electricity use

We then created some background electricity profiles using a Loughborough University tool. The tool randomizes the electricity use for each day of the year. We then assigned these profiles to properties based on the total floor area for each property, so homes get assigned standardised low, medium or high background electricity consumption on top of the heat pump energy use based on the size of the house.

Solar and battery

We then modelled solar energy generation for 3 cases: Zero production, 2000 kWh (a solar system size of about 2.25kW) and 4000 kWh (a solar system size of about 4.5kW) and mapped it across each day using a solar distribution pattern for Sheffield.

Batteries were more complex to consider, because there are different strategies for charging. Our analysis replicated a very simple smart charging logic, where the system does not know the amount of solar generation forecasted for the day but chooses to charge the battery during periods where energy is cheaper. We used 3 different battery sizes (plus zero) for each property.

Analysis

We ran every single possible heat pump dataset against each of the tariff options through ‘R’ to give us a series of output charts where every combination is identified, from zero solar and zero battery, with low background electricity use / and no heat pump consumption, through to the opposite extremes. Heat pump electricity consumption was modelled from 0 kWh to 8,000 kWh.

To give a simplified illustration of how this works, I created this chart below, which looks at electricity use for one day. It gives a cost for each half hour chunk for each tariff: the second chart adds it all up. This particular chart is for a large house with no solar or battery, using a heat pump on a cold day with sub-zero temperatures outside, electric cooking, and some EV charging.

If you were then to repeat this process for 365 days a year, for houses with different size solar systems and different size batteries included in the model, we get to the chart below which plots each of the 350 houses in each use case.

Results

Across the top we consider 3 different size batteries (4, 9 and 13.5kWh), down the right-hand side are the 2 solar systems generation amounts (2000, 4000 kWh). Down the left-hand side are the bill savings for each possibility. The coloured dots represent the outputs for each of the 350 homes we modelled, comparing a heat pump using a flat tariff against a heat pump using Cosy and Go to keep the chart relatively easy to read…

This led to some useful insights.

It became obvious that you need to be careful moving to the “Cosy” tariff unless you are really willing to modify your heating time of use – which is exactly what the tariff is trying to encourage – or have a battery.

Two of the best performing tariffs were Go and Cosy. Agile loses out to these two tariffs in some scenarios, but we are confident that a smarter battery charging algorithm would increase the benefits even further, to the point where it would be better value than either.

Additionally, the smarter battery management platforms can take into account the predicted solar energy for the day to modify the charging pattern for maximum gains.

Real world examples

We came up with four example scenarios that demonstrate some interesting findings. Where we use “Agile” for a comparison, the cost data comes from the recorded values from 2023.

In all cases, we have selected one real world heat pump dataset for each case – with an annual SPF between 2.9 and 3.1.

Modelling assumptions in the charts below:

We used the typical domestic consumption values used by OFGEM, for both heating requirement and electricity consumption. The flat electricity rate is assumed at 27p/kWh. Solar export is assumed to be at 12p/kWh in all cases. Gas costs use a rate of 7p/kWh, with a boiler efficiency of 85%. A heat pump using a “constant” heating pattern is used.

In all cases the bill amounts on the Y-axis refer to bills for the entire house except the second case.

·       Gas and background electricity is included in the column marked “gas”.

·       Total electricity (i.e. the heat pump and the background electricity use), is counted in the other columns where a heat pump is used.

The electricity standing charge has not been included in the bills, and we have indicated the gas standing charge for the gas heated house.

Case 1: Switching to a heat pump in a typical house and using a smart tariff

The first case lets us consider an average household, switching to use a heat pump but not being able to install solar or a battery. One example here would be a family making a “distress” purchase when a boiler breaks down. The heat pump is controlled to keep the house warm constantly, rather than “twice a day” like a boiler. This avoids all the heating energy being used at the “peaks” in the tariff profile.

We can see in the chart that whilst the Agile tariff wins, Go (an EV tariff), Flux or a flat tariff are the next cheapest. The Cosy tariff is slightly better than sticking with a gas boiler. Using an E7 tariff would be bad idea.

Case 2: A small property adding a 4kWh battery and moving to a smart tariff (with no solar or heat pump included)

Our theoretical case here is a pensioner living in sheltered accommodation, or a family living in a tower block. I use this example because tower blocks are considered difficult to heat with heat pumps and generally can’t access solar.

In the chart we can see that on a standard flat tariff there are nearly 20% savings (approx £85 a year) to be made using a modestly sized battery. Moving to Agile would increase the saving to around £160. This is not really into the territory of making good payback periods – but those verging into fuel poverty would find these very useful savings on their energy bills if there was grant funding available for the installation under SHDF, ECO or HUG.

Case 3: A typical house with a 9kWh battery (£7,000), heat pump, solar and smart tariff

Here we have chosen the “typical” house – often considered to be a 3-bed semi, with average electricity use and average heating requirement for the UK, installing 4kW of solar, a 9kWh battery and a heat pump.

The savings here are pretty dramatic. “Go” wins here, with the total house running cost reducing by a little over 50% (£1,100). I’m not that surprised by this result because it’s not dissimilar to my personal situation (though I have been using Agile). Again just noting here that Go is an EV tariff, so if you have an EV there will be additional savings from charging at night as well, but the tariff is not open to everyone.

I’m not going to go into payback calculations in detail here, but obviously in combination with the £7,500 BUS grant, this makes for a pretty sensible payback period – around 7 years.

Case 4: A typical house with a 9kWh battery (£7,000), heat pump, solar and smart tariff, and the levies removed moved from the variable rate of the electricity bill

Finally, we take Case 3, and throw one more element into the mix: removing the environmental tariffs from the variable electricity rate.

This is an idea that has been put forward for at least 2 or 3 years in policy circles. It has been mentioned by Government ministers, as well as making its way into policy consultations, though there has been little sign of progress recently. Still, it was an easy thing to add into our model, so we did it. We used an electricity rate discounted by 15% to model this.

This is not a perfect reflection of how this would work in the real world when implemented – you’ll have to view the results with that in mind.

Conclusions

Coming out of this work there are some good lessons learned.

Running a heat pump continuously can reduce running costs when combined with smart tariffs, but the benefit comes from reduced heating demand in expensive tariff periods, not necessarily from good control of the heat pump…

I can think of several reasons why this might be the case which I will discuss in the next article. The EoH dataset offers many opportunities for more detailed analysis of the factors driving performance.

We have ascertained that batteries might not only be for people with solar panels. This has significant implications for helping reduce strain on the grid, as well as offering savings to people from many different demographics.

Where many social tenants have seen the benefit of solar being added to their homes, others without access to a roof have missed out. This could lead to social landlords considering installing batteries for people living in HMOs, and blocks of flats. As previously mentioned, funding under ECO / HUG or SHDF also seem like good opportunities for helping people in those demographics.

Most of the value of a battery comes from the ability to use a time of use tariff. It’s vital that consumers are educated about the importance of automatic battery integrations to get the best value from batteries. I have heard many people being promised generous savings with batteries that fail to materialise because the batteries are not set up properly, with the customer left to figure out the solutions themselves.

It needs to be fool-proof for them to do this without any technical knowledge – but the savings can be high. It would be good for battery manufacturers engage with this issue and allow people to monitor and check that any tariff integrations work properly and continue to save customers money versus flat tariffs.

We have not discussed payback periods anywhere in this article, because of course all this equipment comes at a cost. We have also not – as yet – got to the bottom of the best heat pump control strategy. The next two articles will consider this in more detail.

A big thank you to my long suffering colleague William Jamieson for doing all the difficult bits!