Can data centres heat homes?

The UK Government is committed to achieving net zero carbon emissions by 2050 in line with the COP21 Paris Agreement. One key objective in this regard is to decarbonise the UK housing sector, which is responsible for about 14 percent of the country's carbon emissions mainly through using fossil fuels for heating. 

There have been discussions about phasing out gas-fired boilers in new constructions. In fact, there are proposals to ban the sale of domestic gas boilers altogether. Initially, this ban was supposed to be implemented by 2025, but it has since been extended to 2035 to give homeowners more time to switch to low-carbon alternatives. Therefore, finding affordable and sustainable alternatives to traditional gas-fired boilers has become a top priority. 

A potential solution is available in the form of data centres. According to the Environmental Investigation Agency (EIA), data centres consume around 1.0-1.5% of global electricity, primarily to power IT equipment. This results in a significant amount of heat generation, which cooling systems then need to remove. However, rather than treating this residual heat as a wasteful byproduct, it should be viewed as a valuable resource that can be captured and reused.

Waste, or valuable product? 

Using residual heat in district heating networks presents multiple advantages for various stakeholders, particularly in light of evolving sustainability targets and the need to transition away from traditional heating systems. 

For data centre operators, harnessing residual heat may generate attractive reductions in operational expenditure and improvements in Power Usage Effectiveness (PUE and Water Usage Effectiveness (WUE), by reducing the load on heat rejection plant. Additionally, monetising residual heat through partnerships with Energy Service Companies (ESCOs) offers an alternative income stream that would mitigate operating costs. The sustainability benefits are significant as partial shutdown of cooling infrastructure would reduce the data centre's energy consumption and carbon footprint. 

However, the true environmental benefit lies in lower external carbon emissions from district heating network users, who would otherwise need to rely upon burning fossil fuels to heat their homes. To maximise this impact, data centre operators should be encouraged to advocate the potential benefits to Municipalities and Planning Authorities. This would strengthen the facility's environmental credentials and potentially support the planning approval process. 

For the ESCOs, data centres emerge as a dependable and abundant heat source, catering to customer needs while reducing carbon emissions. Using low-grade residual heat from data centres as a primary source for heat pumps enables ESCOs to deliver hot water to their networks without the need for centralised boiler plant. With data centres increasingly transitioning to renewable energy sources, the residual heat becomes carbon neutral. 

Promoting adoption

Data centres and district heating systems offer numerous benefits, but their integration could be improved. The Nordics and the Netherlands have notable examples, but the UK needs to catch up. The constraints appear to be economic, rather than technical. 

Firstly, the main challenge in collecting and using residual heat from data centres is the need for established external infrastructure. Although planning authorities advocate for heat reclamation, the absence of a coherent plan for heat network development presents a significant obstacle.

While planning policy often requires developers to facilitate connections to future heating networks, this needs to align with relevant infrastructure development plans. The current disjointed approach means that compliance with regulatory obligations becomes a means to achieving planning permission, with little expectation that the connections will ever be used.

Secondly, from the data centre operator’s perspective, investing in expensive infrastructure to harvest residual heat only makes sense if this leads to worthwhile savings in operational expenditure, in the form of reduced power and water consumption. However, this may be undermined by misalignment of load profiles. Whilst demand from heating network peaks in winter and declines in summer, data centres operate inversely, using energy intensive mechanical cooling during periods of warm weather, but reverting to efficient 'free cooling' strategies during the winter. Hence, exporting residual heat during colder months may result in only marginal improvement in energy efficiency.  This undermines the business case for investment in heat reclaim infrastructure.

Thirdly, the energy supply companies that operate the heating networks also face challenges. Their business case for investment in heat pumps and pipework will assume that the infrastructure will be fully utilised. Uncertainties regarding operational factors in the data centre that are beyond their control, such as IT load growth and periodic maintenance shutdowns, may hinder their commitment. 

Even if these uncertainties are resolved, determining the energy cost poses another challenge. Data centre operators aim to sell heat at prices covering infrastructure costs, while ESCOs seek competitive prices to ensure profitability. Agreeing on a mutually acceptable cost remains pivotal for fostering collaborative partnerships. 

Exploring solutions

The UK Government needs to intervene and learn from successful models implemented in other countries to encourage construction of district heating network infrastructure to avoid falling behind other nations. 

Planning policy should be revisited to allow data centres to be located nearer to urban centres.  This would support Edge applications and bring them closer to where heating network users work and live, reducing infrastructure costs and pumping energy. Although this move might be opposed in some quarters, such attitudes could be overcome if data centres came to be seen as an essential part of an urban infrastructure, providing low-grade, low-carbon heat to support local communities. 

The solution to address the issue of load mismatch requires a large-scale approach. Developing extensive networks serving diverse commercial and domestic users would smooth out demand profiles. This approach allows ESCOs to capitalise on multiple heat sources based on season, unit costs, and heat availability. 

Offering grants and tax incentives to cover the new infrastructure costs could make planning conditions that mandate data centres to connect to district heating networks a reality. Government support would bolster the business case for data centre operators and ESCOs. Introducing tax breaks on energy costs would incentivise data centre operators to contribute residual heat, complemented by subsidies for ESCOs to harness and distribute this heat. 

The imperative of reducing operating temperatures for decarbonising heat networks cannot be overstated. Currently, many district heating networks operate within the medium temperature range, which necessitates the use of fossil fuels as a primary heat source.  Decarbonisation will require the adoption of alternative solutions such as ammonia heat pumps combined with low-grade heat sources. This will mean lower operating temperatures, and perhaps increased pumping power consumption, but the overall carbon reduction would be significant. 

Blueprint for success

Cundall has firsthand experience effectively utilising residual heat, as demonstrated in Odense, Denmark. Working closely with the client and local ESCO, Fjernvarme Fyn, we were responsible for the design of a hyperscale facility that will eventually export up to 45MW of residual heat into the local district heating network.

Fjernvarme Fyn's extensive heat network serves 65,000 metred connections, connecting various heat sources and user types to smooth out load demand profiles. The data centre is integrated into the district heating network through multi-stage ammonia heat pumps housed in a dedicated energy sector. 

The heat pumps capture low-grade residual heat from the data centre, increasing the district heating network's temperature by 70-75⁰C. Upon completion, it is expected to provide 165,000MWh of heat annually. This heat is entirely carbon emissions-free due to the data centre's reliance on renewable energy sources. 

To replicate this in the UK will require significant investment and political will over several years. But using residual heat from data centres to power large-scale district heating networks is more than just a theoretical concept, as the Odense scheme demonstrates. If replicated in the UK, such schemes could make a substantial contribution to achieving the nation's 2050 net zero carbon targets.