Abstract
As the UK makes steps to reduce its greenhouse gas emissions to net zero, electric heat pumps are proposed as a solution for heat decarbonisation in the majority of homes. This paper used smart meter data to examine the gas and electricity consumption of 187 homes with air source heat pumps installed under the Green Homes Grant incentive scheme. It was found that 1 year after the installation, 24% of households were still connected to the gas distribution network and that 21% of households still used gas, either for cooking or for heating. We reflect on the implications of this ongoing gas connection and use in the light of the UK’s net zero plans, and discuss how households could be supported by government to fully transition away from using natural gas.
Introduction
The UK has a legally binding target of net zero carbon emissions by 2050, 1 and the combustion of unabated natural gas must therefore either become limited or be eliminated, depending on the ability to remove emissions from the atmosphere. 2 Yet, in 2023 natural gas still represented 35% of the UK’s primary energy consumption, a share that has remained approximately constant since 2010. 3 Because most UK households are heated with gas boilers, 34% of all UK natural gas is currently used in homes. As a result, the UK Heat and Buildings Strategy 4 has set out large scale plans for electrification of heat via heat pumps, as well as heat networks, and a potential role for hydrogen.
However, important uncertainties continue to affect the transition to clean heat. Firstly, the significant role of electrification in government strategy remains hampered by slow heat pump uptake in homes. The UK is still far behind similar European countries in per capita installations, 5 despite grant funding through the Boiler Upgrade Scheme. 6 Assuming the current growth in recorded heat pump installations is sustained, the stated 2028 objective of 600 thousand heat pumps per year would be reached 5 years late. 7 Secondly, the future role of hydrogen remains unclear, with some stating that it has no place in residential buildings 5 while others proposing that advanced applications (e.g. fuel cell technology in smart grid applications) could support the resilient provision of heat and power. 8 Thirdly, the future of the gas distribution network in the wider context of transitioning to net zero is undecided. Yet any eventual decommissioning needs to be clearly planned and well managed to limit the high costs and distributional impacts introduced by a reducing number of connected customers paying for the upkeep of the network. 9
Within this context, this paper focuses on gas supply outcomes in homes which transition to a heat pump. Approximately 55,000 heat pumps were installed in UK homes in 2024. 10 Following installation, households can choose to disconnect from their gas distribution network, in order to avoid paying a daily standing charge for gas, currently around £0.33 per day. 11 There are several reasons why households may not disconnect, for example, they may not know it is possible, they may be charged a fee by their energy supplier which could be perceived as too expensive, or they may have gas cookers or hobs which they do not wish to replace with electric alternatives.
The cost aspect of gas disconnection, as well as other potential barriers for households in doing so, are being investigated at the time of writing by the regulator Ofgem, who states that disconnection regulatory frameworks should “operate effectively” and “assist in achieving net zero”. 12 Yet, until recently, little data was available to investigate gas use after heat pump installation and thus understand whether households are disconnecting and, if not, whether there is ongoing demand for gas. The Centre for Net Zero recently published findings 13 that gas demand fell by 90% amongst a sample of 1,321 households who had installed heat pumps. The authors asserted that remaining gas use was due to cooking, although evidence for the latter was not provided.
In this paper we provide a detailed study of gas use following heat pump installation in the UK, finding a range of practices in terms of ongoing gas connection and gas use, placing the findings in the context of the UK’s future pathways to net zero, and raising questions of how to more effectively assist households in their own part of the transition.
Data
The dataset used in this study was collected as part of a research project evaluating the UK government’s Green Homes Grant (GHG) policy. 14 Under this policy a scheme was run with applications open between September 2020 and March 2021, providing subsidies for energy retrofit measures to existing homes, including air source heat pumps (ASHPs). 15 Half hourly electricity and gas smart meter data was collected by University College London (UCL) using the method described in 16 before and after the retrofit measures were installed in around 2500 homes, to investigate the impact of the retrofits using thermal performance estimates.
In this research, 187 homes were selected from the total 2500, which received only an ASHP under the GHG scheme (however other retrofit work outside of the scheme may have taken place). In most cases less than 12 months of pre-installation smart meter data was available 1 , whereas up to 2 years of data was available after heat pump installation. Importantly, this means that our analysis cannot provide robust annualised estimates of energy use before and after heat pump installation; instead we focus on gas (and electricity) use after households adopted heat pumps.
For each home, available metadata included GHG scheme data and public energy performance certificate (EPC) data. The scheme data gave details of the type and dates of the measure installed, and the EPC data provided confirmation of the space and water heating systems present prior to ASHP install. This data was made available to the researchers by the UK Department for Energy Security and Net Zero (DESNZ). Data was pseudonymised and was required to be analysed within a secure research environment, UCL’s Data Safe Haven. This placed necessary constraints on what results can be shown 17 in order to prevent the identification of individual households; see the Methods section for more details on the constraints applied.
Homes were selected from the dataset using the condition that the only retrofit measure was an ASHP. This excludes hybrid heat pumps where an ASHP and a gas boiler were installed together under the scheme. However, it may not exclude all types of hybrid set-up: whilst a condition of the scheme was that households were not permitted to keep their existing gas boiler, they were permitted to buy a new one if they wished to have a boiler.
Finally, all heat pump installs were cross-checked against those lodged with the UK quality scheme TrustMark to verify that they had taken place, and the public EPC data was checked to ensure that all households included were previously gas heated.
Methods
This paper presents an exploratory data analysis of gas and electricity data in timeseries form, after heat pump installation, and beforehand where data exists. The data was interpreted by the authors using physical theory and literature to infer the possible energy service demands being met; there was no accompanying social data from occupants to confirm the reasons for different outcomes observed in the timeseries data. The interpretations made were discussed with civil servants associated with the Green Homes Grant scheme to check their viability and exclude alternative explanations.
The research was conducted as follows. Firstly, the number of households still connected to the gas network after heat pump installation was tracked over time by determining the last available valid gas smart meter reading for each home with ongoing valid electricity smart meter readings. It was then assumed that the lack of gas data thereafter was due to removal of the gas meter, rather than a fault in data collection. This assumption may not be correct in all cases, however given the continuing valid electricity data collection it is highly likely.
Secondly, analysis was carried out on households still connected 1 year after their heat pump was installed. Half hourly smart meter import data was aggregated to give daily energy data before and after heat pump installation, with the install date/period excluded. The pre and post heat pump install electricity and gas use were plotted for each of home with sufficient data, and then manually categorised into three groups: connected to gas meter but with zero consumption; continuing low level non-seasonal gas use; and continuing seasonally varying (i.e. winter) gas use.
To avoid statistical disclosure, any days with data from fewer than 3 households were not included in the outputs to avoid the risk of identifying individual homes. In addition, the average daily electricity and gas consumptions pre and post installation were both normalised (to maximum consumption) across all four datasets, transforming absolute scales into relative scales between 0 and 1. This means that gas and electricity consumption cannot be read from the plots in the results section. Instead, pre and post heat pump gas and electricity consumption are shown relative to each other, so that the proportional change in consumption following heat pump installation can be observed.
Results
Presence of a gas connection over time
Number of heat pump customers connected to the gas grid as a function of time after heat pump installation.
Table 1 shows that the number of connected customers decreased over time, with the disconnection rate peaking between 6 and 9 months after heat pump installation, and slowing thereafter. By the end of the data collection period – which is different for all customers but at least 2 years and 5 months – 20% of customers were still connected and the rate of disconnection was very low.
24% of customers were found to still be connected to the gas network 1 year after getting their heat pump. In the subsequent analysis, this set of customers were analysed in more detail to characterise their gas (and electricity) use and how it changed after heat pump installation, in both magnitude and seasonality. For example, if gas use before the heat pump was seasonally varying (winter dominated), and afterwards it remained seasonal but decreased in magnitude by two-thirds, this may indicate that gas continued to be used for space heating, but not as the main heat source.
The 45 homes with gas meter readings 1 year post-installation were divided into three groups on the basis of their energy demand, as explained below with accompanying plots of gas and electricity use (Figures 1–4). Due to the data privacy issues outlined earlier, data from individual households cannot be shown, thus the examples are aggregated over a minimum of 3 homes; as also discussed previously, energy use cannot be reported for these samples and is instead presented as percentages of the maximum daily energy use. Note that the data normally span one winter before the heat pump installations and up to three winters afterwards. Timeseries plot of daily gas and electricity use in households which retained their gas meter but used no gas after their heat pump was installed. Illustration of daily gas and electricity use in households whose gas use remains positive but low following heat pump installation. Illustration of daily gas and electricity use in households with seasonal gas and electricity use following heat pump intallation. Illustration of homes with low seasonal variation in electricity use following heat pump installation. Includes a subset which revert to seasonally varying use of gas 1 year later.
Homes with no gas use
6 homes out of 45 had used no gas since the heat pump was installed, and a proportion 2 of these were still connected when data collection ceased up to 3 years post-install. These customers retained their gas meters, and were therefore charged daily for their connection but were not using it. This is illustrated in Figure 1, which also clearly shows a new seasonal electricity demand attributed to the heat pump.
Homes with low, non-seasonal gas use
21 homes out of 45 were found to have ongoing non-zero, year-round gas use after their heat pump was installed (Figure 2). This low consistent gas use is most likely to be associated with cooking using gas hobs. Gas ovens are a less likely explanation as they have a lower prevalence than gas hobs. 18 An alternative explanation could be the continuing use of gas to heat water. This is less likely to be the case since research on hot water use 19 shows a small seasonal effect on heat demand.
Homes with seasonally varying gas use
18 of 45 homes displayed seasonally varying gas use after the heat pump installation: higher in winter than in summer. This group contains 12 homes which showed seasonal electricity use as well as seasonal gas use, and 6 homes without seasonal electricity use but with seasonal gas use.
Figure 3 shows the mean daily gas and electricity consumption of the subgroup of 12 homes which display both seasonally varying gas and electricity use after the heat pump. The electricity consumption is compatible with that of a heat pump: demand for space heating varies according to the indoor to outdoor temperature gradient and peaks around a third of the pre-heat pump peak gas consumption. The seasonality and magnitude of the winter post-heat pump gas consumption indicates it is also being used for space heating. The flat summer gas consumption also indicates gas cooking and/or water heating.
Two sources of metadata were employed to understand the continued presence of seasonal gas use for heating in homes where the intervention was labelled as an ASHP only. These consisted of EPC data from a maximum of 2 years prior to the heat pump installation, and GHG scheme records, including details of the install and proposed use. This metadata, whilst not explaining the post heat pump seasonal gas use in all homes, showed that some properties had gas fires (i.e. secondary, room heating), and others had gas boilers alongside the heat pump. Given that existing gas boilers were not permitted to be left in place under the GHG scheme, it is likely that the latter group purchased a new boiler at the same time as the ASHP.
Figure 4 shows the subgroup of 6 homes with very similar gas consumption after the heat pump installation to beforehand, and without any increase in electricity demand following ASHP installation. This suggests a retention of gas space and water heating, with no or limited use of the heat pump.
Furthermore, a small number of homes 3 were observed in which the electricity consumption increased and gas consumption decreased the first winter after the heat pump was installed, as expected - but then in the following winter low electricity consumption and high gas consumption were recorded. This suggests that the heat pump was used for one winter, but the household reverted to gas heating thereafter. Further investigation of metadata confirmed that the same occupants were present throughout: this change cannot be associated with a change of occupants, rather, it seems that the occupants did not continue to use their new heat pump.
For the remainder of the 6 homes with very similar gas consumption after the heat pump installation to that before its installation, the heat pump does not appear to have significantly altered their gas or electricity use, suggesting that it was not used to provide significant space or water heating. In order to investigate the cause of this, the validity of the installations was checked with Trustmark: the lodgement details confirmed the installation of a heat pump. Therefore, the cause of this finding is not clear: despite the installation of a heat pump, gas space and water heating appears to have been utilised. This could have been caused by a metadata error, in which a pseudo-anonymised house code was misaligned between datasets, however this is unlikely given the rigorous data collection process followed. 16 Alternatively, the household may have had a heat pump installed but decided to use a boiler instead.
Summary of results and interpretations
Summary of outcomes 1 year after heat pump installation.
Discussion
The analysis of smart meter data for a sample of homes retrofitted with a heat pump through the Green Homes Grant has found that 24% remained connected to the gas network 1 year after the installation date. Three main potential causes have emerged: ongoing payment for a gas connection (with no gas use), ongoing demand for gas cooking, and ongoing demand for gas heating. In this section these outcomes are discussed in the context of the UK’s transition to net zero.
Ongoing payment for a gas connection
3% of households in the sample remained connected to the gas network at least 1 year after their heat pump was installed but did not use any gas. These households would have paid an annual standing charge of around £105 4 . 20 Without further data collection it is not possible to determine why these households remained connected to the gas grid; potential reasons include: not knowing about the potential for or process to arrange gas disconnection; intention to disconnect in the future but delayed action for some reason; deciding not to disconnect for cost reasons; or retaining the connection in order to maintain the future possibility of gas heating or cooking without incurring disconnection and reconnection charges.
It is therefore important to conduct research into why customers do not disconnect, to target effective policy and practices. For example, if such research highlighted no particular demand for gas, but instead barriers to gas disconnection such as lack of knowledge or disconnection cost, an obvious policy recommendation would be that the disconnection process is fully integrated into the heat pump installation process and is carried out by default, since this would reduce customers’ energy bills in the long term. On the other hand, research may indicate that households do still want to use gas, or at least they are committed to it by the appliances within their home. This is discussed next.
Ongoing demand for gas cooking
21% of homes were found to still use gas 1 year after heat pump installation. We estimated that 11% of households continued to use gas for non-seasonally varying energy demands, the most likely being cooking. If this interpretation is correct then these customers paid the standing charge for the sole purpose of gas cooking, meaning that for a typical annual gas cost of £20 for cooking, an extra £105 per year was incurred for remaining connected to the gas network.
Further investigation is required to determine why households continue to cook with gas despite the standing charge adding substantially to the operational cost of doing so. According to a recent survey of 2000 adults in the UK, 21 half cook with gas hobs, with familiarity being the main driver of preference for cooking technology, and perceived high capital and running costs being the main barrier of switching to electric cooking technologies. For the heat pump owners in our study, if the gas standing charge is fully attributed to cooking costs, the running cost of gas cooking exceeds that of electric cooking.
To address this, heat pump retrofit incentives could be introduced which include an efficient electric cooker/hob, in addition to the introduction of minimally disruptive solutions to reduce damage to kitchen fittings and fixtures and to ensure that the household is not long without a working hob. It is possible that the new cooker could be paid for in full or part using future savings from avoiding the gas standing charge.
It is clear from the above that with a little help for households it would be possible to bring them completely off gas once they adopt heat pumps. However, it is not currently clear from government strategy whether there is an aim to transition homes off gas, or to provide customers with a choice to retain it with a view to using it in future. Therefore we recommend clarification of heat policy aims in order to clearly plan the future maintenance or decommissioning of the gas network. 9
Ongoing demand for gas heating
10% of homes displayed seasonally dependent gas use. The majority of these also displayed seasonally dependent electricity use, which may indicate the use of the heat pump alongside gas heating provided by room gas fires or a boiler. Although there is no available qualitative data from these occupants to explain this finding, several possible reasons have been proposed in the literature. Some householders indicate uncertainty around whether the technology will meet their heating needs, 22 or a preference to have more control over the temperature in individual rooms. 23 Whilst it is often assumed that households install heat pumps in order to replace their entire heating demand, this is not always the case – research in Finland has found a high prevalence of households adding heat pumps to their existing heating system, essentially making their own hybrid systems. 24 Additional socio-technical research 25 would reveal the extent to which heat pumps are meeting the different heating needs of household members 26 which are set out in 27 and could develop mitigation strategies to further decarbonise these households.
A minority of the 10% of homes with persistent seasonally varying gas use showed no indication of heat pump use in their electricity data by the end of this study: either from the point of heat pump installation, or after the first winter with the heat pump. It is possible that the former may be caused by metadata errors, however the latter group clearly show short term use of the heat pump followed by reverting to gas heating. This may be due to dissatisfaction with or failure of the heat pump to deliver the desired comfort or cost. Reversion to gas heating does not meet the government’s policy objectives or support a transition to Net Zero, despite investment into low carbon technologies. Research to understand the causes of these practices would support the development of mitigation strategies, such as additional support for households with using their new heat pump, or guaranteed energy and cost performance.
Policy implications
Three main risks emerge from our findings on the persistence of gas use. Firstly, risk of missing national decarbonisation targets. Residential buildings in the UK emitted 52.8 MtCO2e in 2024, accounting for 14.2% of total UK emissions. 28 Yet, most future pathways to reach our net zero goals expect this sector to transition to low carbon buildings, fully decarbonising by 2050. The government’s Net Zero Strategy and the balanced pathway futures from the Climate Change Committee’s carbon budgets 6 & 7 assume emissions from building will drop between 98-100% by 2050.5,29 While other decarbonisation options exist, heat pumps are expected to provide the bulk of this effort, heating up to 80% of buildings by 2050. 5
The results in this study suggest that up to 10% of homes that installed a heat pump continued to use natural gas for heating in some way, and that 12% continued to use gas for cooking. Although it was not possible to quantify annual gas use changes following heat pump installation due to data limitations, previous work 13 estimated a 90% reduction. If the persisting 10% or 985 kWh per household were scaled up across 80% of 27 million existing households adopting heat pumps in future (not counting new homes or households), this would lead to continued sector emissions of 4.3 MtCO2/year by 2050. This would in turn increase the reliance the CCC expects the UK to have on carbon dioxide removal technologies by 12%. However, the novel technology options that provide the bulk of this removal are still not available at scale. 30 In summary, the failure to fully decarbonise homes risks meeting our net zero targets more difficult.
The second risk is that, although only a small percentage of our sample were shown to revert to gas heating in full, leaving customers connected to the gas network opens the door for more doing so in future. Government reticence to banning either gas boiler sales or the connection of new homes to the gas grid, along with a continued narrative of repurposing the grid to deliver hydrogen for heating, gives a signal that the gas network will remain relevant in future and that reverting to gas heating remains an option.
The third risk is the potentially significant costs for households who stay connected to the gas grid whilst others disconnect. Gas grid infrastructure is long-lived and capital intensive, and current cost schedules expect up to £3bn to be recovered after 2050. Further, current estimates suggest that decommissioning the grid could cost over £50bn. 31 These costs would normally be recovered through consumer bills, but the necessary transition away from natural gas makes this less likely with every step the system takes forward. Recognising this risk, Ofgem’s RIIO-3 price control has allowed for higher investment cost recovery on consumer bills for the period 2026-31, 32 and makes provisions to consider a decommissioning fund from customer levies. Notwithstanding, as the number of connections dwindles, cost pressures will increase for those who either choose to stay connected, or cannot afford to leave. 9
To address these risks, we propose a clear steer from government in two areas. Firstly, clarity on the future of the gas network. Ensuring that it is widely understood the gas network will be decommissioned, and making it clear on when it will occur and how this will be managed fairly, could help households make decisions on heat pump adoption and gas disconnection. Secondly, targeted support for households regarding the disconnection process. Providing guidance and finance to support households through the process of gas disconnection, including helping replace cookers, will help maximise the decarbonisation and cost saving potential of heat pump installation.
Conclusion
This research studied smart meter data collected from homes before and after the installation of an air source heat pump through the Green Homes Grant in the UK. Most households who installed a heat pump disconnected from the gas network within 1 year; however, a 24% did not, and 21% continued to use gas. We inferred from the seasonal patterns of gas usage several possible different ways it may be used following heat pump installations. The only explanation of this persisting gas usage is the continued use of gas appliances within these homes for cooking (11%) and heating (10%).
The sample analysed in this study contained under 300 heat pumps, where smart meter data enabled the change in gas and electricity use to be determined upon installation of a heat pump on a known date. Future research should add breadth and depth to this study. Broader analysis of a representative sample would provide clearer insights into the scale of the issue and the consequent missed opportunity in decarbonisation. Deeper analysis could uncover whether heat pumps are fully meeting households’ heating needs, and to what extent there is a demand for gas cooking or whether targeted interventions and subsidies could help households move to electric cooking and therefore away from gas use completely.
In the UK, heat pump adoption is situated within an unclear policy context. It is not clear to households what other steps to take when they get a heat pump installed, regarding use of other heating systems, other appliances, and gas disconnection. At the same time, it is not clear how the wider energy system will be decarbonised, and how decisions made by individual households may interact with, affect, and be affected by this wider context. For example, if households assume that hydrogen heating is forthcoming, they may decide to not invest in a heat pump and/or to remain on the gas network. Equally, if government assumes that there will be an ongoing demand for gas from households, there may not be decisive plans made to decommission the gas network.
Therefore, two important recommendations from this study consist of a clear strategy for the decommissioning of the gas grid, alongside a clear policy steer to support and incentivise households with heat pumps towards complete electrification of their homes. These two aspects will reinforce one another, as households understand the implications of decarbonisation of the wider system for their homes, and the gas grid’s decommissioning can be managed without uncertainty over whether there will be ongoing demand for gas in households and other sectors.
Footnotes
Acknowledgments
The authors would like to thank Dr Daniel Scamman for his helpful feedback, and DESNZ for useful discussion and interpretation checking.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by UK Research and Innovation, under grants: EDOL (grant number EP/X00967X/1), SERL (EP/P032761/1), Energy Demand Research Centre (grant number EP/Y010078/1).
Data Availability Statement
The data from this study has not been made available according to the terms of use