Is on-site hydrogen production, storage and use as a fuel, a viable carbon-cutting energy option for data centres?

Using hydrogen as a fuel has mostly been discussed as a combustible solution for grid energy needs. Much press coverage has been devoted to its potential to provide zero-carbon energy where excess variable renewable power from wind or solar runs electrolysers to produce green hydrogen using clean energy that would otherwise go to waste. 

But could hydrogen be produced effectively at a data centre scale to achieve carbon savings? What is the potential for on-site production of low-carbon hydrogen (hydrogen produced without associated carbon emissions) for use as an energy carrier in data centres? Could fuel cells powered with clean on-site produced hydrogen replace batteries for back-up or even additional power?

Could the principles of developing such energy storage solutions assist Variable Renewable Energy (VRE) developments in proximity to new data centres and in turn drive greater consumption of low-carbon renewable energy within those facilities?

The fundamental characteristics of a medium to large data centre are high energy demand, energy storage, and power generation. This can be combined with an urgent need to cut Green House Gas (GHG) emissions and a need to use more VRE as primary power sources. 

The components of a hydrogen producing system, an energy generation source and an energy storage system already exist within conventional data centres. What are the possible synergies of converting these systems to hydrogen production to achieve carbon savings and drive-up use of VRE? 

Where to start? 

A new whitepaper from i3 Solutions Group [“The Case for On Premise Hydrogen Production in Data Centres for Greenhouse Gas Abatement Benefits”] considers an energy regime of the continuous operation of hydrogen-powered generation equipment during periods of high grid carbon intensity. In such a scenario the principal purpose of on-site hydrogen production and power delivery equipment is to operate as the prime source continuously for extended periods. This is a departure from the current situation where data centre power generation is operated as the prime source of electricity and the grid reverts to a standby source. 

Of the many ways to produce hydrogen, the electrolytic splitting of water is the focus of the paper.

Due to the low-temperature operation and the established and commercialized status of the technology, alkaline electrolysis is considered for the concept. The VRE energy storage regime studied in the paper requires a method for storage of generated hydrogen between periods of low grid carbon intensity when hydrogen would be produced on-site and periods of high grid carbon intensity when hydrogen would be consumed.

Having produced and stored the hydrogen through electrolysis the focus of the paper then turns to turning hydrogen into electricity. The paper says: “The use of fuel cells to achieve... energy extraction is more efficient and avoids due to the combination of any fuel with heat in the presence of air which includes both nitrogen and oxygen. The results show that carbon emissions saving can be achieved by employing on-site hydrogen production coupled with additional consumption of grid energy at times of low carbon intensity. However, the reductions achieved are very low.” 

Location, location, location

The concept examines the viability of utilizing additional grid electricity during periods of low grid carbon intensity to produce and store hydrogen. And using that stored on-site hydrogen to offset grid electricity consumption during periods of high carbon intensity. 

To explore the possible merits or deficiencies of such a concept a mathematical model has been created to consider the application of this concept to grid carbon variations in a series of different geographical locations that calculates the potential carbon reduction benefits. 

The model has used a notional data centre and a selection of applicable hydrogen technologies for hydrogen production, storage, and power generation.

Locations modelled include UK (Nationally), Scotland, Ireland, England (South East) which calculate the “Simulated Results for Model of 10MW DC with Matched Hydrogen Production & Power Generation Facilities – 2021 UK National Grid Carbon Intensity.” 

The report says: “The optimal solution to maximise electrolyser utilization for green hydrogen production is to site the installation in the geographical location of renewable power generation, with ample hydrogen storage capacity and ability to export surplus hydrogen. In this way, every available hour of surplus renewable energy production can be exploited using electrolyser plant operation.”

The paper states: “The merits of full-scale on-site hydrogen production range from low to medium depending on location.” 

The myriad commercial and technical considerations and calculations of the viability of hydrogen production, storage and its use to cut carbon in data centres are explored. Analysis of the simulation results examines the challenges and opportunities of current shortfalls and how they can be overcome. 

If not how, then when?

Will hydrogen production, storage and use be used to offset carbon in data centres? As more advantages and benefits need to be demonstrated today the answer is perhaps, but not yet. However, given the nature and scale of the challenges to cut data centre carbon intensity it doesn’t mean not ever, and could be sooner than one thinks.