New approaches for net zero data centres

This is the fourth article in our Data Centre Thought Leadership series in which Gavin Bollan explores a range of alternative primary and back-up power supply options for data centres, where reliability and resilience are key.

Data centre users demand reliability from operators. The lights (or more to the point, the servers) must stay on, a temporary interruption to electricity supply or even a fluctuation in voltage can cause an outage and with it countless adverse downstream effects on communications, transactions, and security.

To build resilience into their power supply options, operators are now considering alternative options; principally in the choice of back-up power - but where power grid conditions are poor and reinforcement requiring too long a wait, some are also considering alternatives to primary supply.

Firstly, primary supply. The preferred option for most data centre operators will be a stable grid supply for a plug and play environment for the data centre. Resilience and capacity building by the network operator may be required behind the scenes and come at a price, but the operator needs the ability to connect and effectively forget the power supply solution.

Secondly, back-up. Data centres have always integrated a high degree of redundancy into their backup. Even early purpose-built datacentres included a back-up battery system for microsecond corrections and to hold the fort until the main insurance against downtime kicked in – the emergency diesel generator (EDG).

There is a lot to be said for the EDG. They are cheap, scalable and there is a lot of choice on the market; their fuel is available in most of the world. If the primary power supply remains on, they will only ever be energised for testing, but this alone can lead to local pollution issues as diesel generators are usually at their dirtiest when first energised;(the author can attest to this first hand, having on more than one occasion had a face-full of soot whilst standing on a data centre roof when the EDGs started up). There is often a disconnect between the environmental performance achievable by EDGs and the increasingly stringent standards which apply to combustion plant. This is a particular concern in Europe where an aggregation of EDGs may lead to a data centre being technically considered a Large Combustion Plant with the pollution control expectations that follow from that designation under the Industrial Emissions Directive.

Drop-in solutions, such as biodiesel, will not improve the air emissions performance but will address the issue of fossil greenhouse gas emissions. Operators can expect to pay more for biofuels following initiatives such as the Sustainable Aviation Fuel (SAF) Mandate in the UK (and EU equivalent) which already require a percentage of bio- or e-fuel for aviation fuel and hence have significantly ramped up demand. This pressurises the supply chain and is prompting a flurry of activity around SAF production, with multiple technologies in development producing SAF from biomass, municipal solid waste, hydrogen and others. For the time being, the main established pathway is hydrotreatment of esters of fatty acids – and there isn’t going to be enough chip fat for everyone. Long term, when the supply side has caught up, drop-in fuels may start to look attractive again; clearly other options are also worth exploring and each operator’s strategy will look different.

Following on from our article on energy solutions for data centres, where we discuss the integration of renewable energy supply and large-scale battery storage systems, we will now look at more primary supply alternatives for primary and backup supply when circumstances dictate a less usual approach in the future.

Primary supply options

Primary supply alternative #1: Thermal plant

Sites in many prime locations have already been developed, where power, water and international data cables are close by or plentiful. Current developers may be forced to look at less optimal sites where one of the key requirements is constrained, or even absent. Should you just generate your own electricity?

One option we have seen considered in the UK is to generate on-site electricity from tried and tested thermal plant, where connection to the gas grid was a quicker proposition than a suitably upgraded electrical supply. Combined heat and power (CHP) using a combined-cycle gas turbine made a lot of sense to consumers of electricity and steam or hot water less than a decade ago when the greenhouse gas (GHG) emissions for a gas-fired thermal plant would be less than grid electricity. At the time, coal was still in use, and this raised the carbon intensity of grid electricity. These days, the proliferation of biomass, interconnection from clean neighbouring grids, solar, and especially wind means that a CHP is now always dirtier in GHG terms than just importing electricity from the grid. The price per kWh differential between gas (still cheap) and electricity (still not) gives a compelling reason for some industrial operators, particularly where electrification of heat is not practicable. For data centres, which are net rejectors and certainly not consumers of heat, a CHP will just add to this problem of heat to disperse; high grade heat is reusable in a combined-cycle system to generate more steam and electricity but lower grade heat in abundance must still be dealt with.

Unless a heat consumer is available (and can connect cheaply, which is highly unlikely), gas-fired generation starts to look very profligate and a magnet for planning objections on climate grounds. CHP is no longer the energy technology of the future in GHG terms and a dead end on the path to net zero for data centre operators.

Primary supply alternative#2: Nuclear power

For the last few years there has been increasing talk of the largest data centre operators investigating the viability of small modular nuclear reactors (SMR), with Rolls-Royce in the UK aiming to supply the first in the 2030’s. The licensing and consenting framework for SMR is nowhere near developed and whilst there is a working assumption that such devices may be available to commercial operators, the very concept of “regulatory hurdles” could have been invented for such an undertaking. The idea of placing such a potent and reliable energy source next to your data centre is appealing but for SMR, being a fast follower rather than first adopter may be a prudent strategy. And by “placing it next to” your data centre, this is not a shipping crate being plonked down on some hardstanding: an SMR requires a good couple of hectares.

Unless you are planning to buy two, there is still redundancy to consider. The Rolls Royce standard unit will run for between 18 and 24 months before an 18-day refuelling period is required.

The nuclear industry is heavily regulated and codified, not least to assure safe and stable operation of the assets, and the interface between providers and commercial consumers is sure to be complex; with complexity also comes expense. This is in addition to the estimated capex cost of £2bn, which puts this option beyond the means of most. Deep-seated concerns about safety in general, and the transportation of new and spent fuels in particular, mean that the social licence to operate a data centre with SMR power will be hard-won. Other territories such as Canada have an active SMR programme and enormous emphasis is expected on consultation and engagement with communities and first nations.

To round off by considering the other nuclear game in town: as the saying goes, nuclear fusion has been 20 years away for around the last 50 years.

Primary supply alternative #3: Geothermal energy

Geothermal systems can be designed for the loads and capacity factors which are attractive to data centre operators. Geothermal energy won’t come to you, however; you have to go to it. And “it” has a far from homogeneous global distribution. Geothermal sites which emit steam at an appropriate temperature and pressure, or can rapidly heat injected water to generate the same tend to be associated with areas of seismicity and vulcanicity which can mean several additional volumes to be added to your project risk register and resilience plan.

Steam is the energy vector for geothermal operation of electrically-powered assets and in the CHP section previously we have already discussed the pros and cons of steam turbine electrical generation. Heat is not the friend of the data centre.

Geothermal electrical generation is currently niche in most parts of the world, with gigawatt scales of generation only found in a handful of countries, where it can be assumed that the lowest-hanging fruit has been picked and exploited to generate grid baseload. As the technology becomes more efficient, more marginal sites may come into play but proximity to the source will still be the overriding consideration.

Back-up supply options

Long-term energy storage vectors (hydrogen, compressed air, redox) to consider as an alternative to diesel generation for back-up supply are in various states of maturity. Compressed air batteries, for example, have advantages in terms of stability and eventual scalability, and are pollution-free at point of use, but also bring heat rejection issues. A couple of emerging alternatives are discussed in more detail below.

Back-up supply alternative #1: Redox flow batteries

Redox flow batteries offer a very stable system that once charged, stays charged with little deterioration and so may increasingly be considered as a back-up supply which can be scaled in accordance with requirements. Systems to meet the load demand of a typical data centre are not yet commercialised and a large capacity system would mean the modular deployment of the smaller systems currently available. The overall energy density of such a system would be low and the footprint large, making this technology one to keep an eye on rather than immediately consider.

Back-up supply alternative #2: Hydrogen

The energy storage option that has had the most attention paid is hydrogen, not least here at SLR where for many years now we have supported and advised clients on such projects. Hydrogen can be generated from excess and otherwise unusable renewable energy, for instance during sunny periods when demand is low and windy periods where turbines spin without exporting because the wind farm grid connection is at capacity. As a primary fuel for a data centre, it makes little sense because why would you convert electricity to hydrogen and back if you run on electricity; electrons are also (usually) far easier to transport than molecules.

Generating your own hydrogen is possible but often the sort of idea that precipitates pained looks and the “let’s focus on being a data centre” chat from financial decision makers.

Hydrogen is reasonably transportable in the quantities it would be required in to act as a back-up fuel, and storage technology is mature. Storage of a lot of hydrogen on your site (5 tonnes or more) brings you under the glare of the Health and Safety Executive as a COMAH site, which many instinctively wish to avoid but is nonetheless perfectly manageable with the right will and advice. The storage of large quantities of explosive gas can make neighbours understandably nervous and the social licence issues mentioned previously would almost certainly feature.

Hydrogen is versatile in that it acts as a thermal fuel and a chemical feedstock for fuel cells. The emergence of hydrogen internal combustion engines in recent months means that hydrogen combustion generators are a distinct possibility. Fuel cells though remain the superior option for a data centre – no emissions (other than water), are otherwise clean and quiet and directly generate electricity.

Once the supply side market for hydrogen becomes properly established over the course of a few short years, the emergence of fuel cell backup power for data centres (and beyond) will become a viable option. Green hydrogen would of course be the fuel of choice in pursuit of a net zero strategy – generated from electricity produced by, or attributable to, renewable generation with no GHG emissions at the point of production or use. By an unfortunate quirk of regulation, your tankerload of hydrogen may currently NOT be delivered by a fuel cell truck and so diesel power clings on to the value chain.

How SLR supports data centre developers and operators

We support the sector by providing turnkey advisory solutions to the challenges of changing regulatory landscapes, carbon footprint reduction, energy supply and increasing energy efficient technologies, reducing water use, and the consumption of natural resources. ​

SLR is uniquely placed to provide solutions through the full life cycle of a data centre and our network of technical experts support clients in quickly bringing projects from concept to feasibility and planning to delivery.​ Our breadth of expertise and experience in this sector enables us to provide strategic advice and value which can unlock investment much more quickly and de-risk decision making.

• Site selection & feasibility studies
  • Evaluating geographical risks (such as; seismic, flooding, extreme weather)
  • Assessing grid reliability & energy supply
• Sustainability & ESG compliance
  • Implementing energy-efficient cooling & heat recovery solutions
  • Supporting carbon-neutral & renewable-powered data centre development
  • Conducting lifecycle emissions & environmental impact assessments (EIA)
• Risk, resilience & uptime optimisation
  • Designing for redundancy & disaster recovery planning
  • Optimising power & water usage efficiency (PUE & WUE)
• Planning, permitting & regulatory compliance
  • Successfully navigating planning & permitting regulations

The digital age is here to stay, and the need for resilient, energy-efficient, and sustainable data centres has never been greater. By taking proactive steps today, data centre operators can future proof their infrastructure while supporting a greener, more connected world.