By Suchismita Sanyal, General Manager for Computational Science on Oct 1, 2020
Hydrogen could play a central role in tomorrow’s lower carbon energy system. Shell is exploring hydrogen’s potential for the heavy industry, transport and heating sectors, as well as in integrated solutions with other clean energy sources like offshore wind and solar. Shaping a sustainable hydrogen economy means creating a value chain for hydrogen that allows its production, storage and transport in safe, affordable and sustainable ways.
HPC allows Shell researchers to scale up and speed up the application of computational science to solve existing technical barriers to the hydrogen economy. Computational science refers to the use of physics-based models to predict the behaviour of materials and systems. By using HPC, Shell computational scientists help to enable the hydrogen economy in several ways. First, they model atomic and molecular scale structures to design new materials with specific and useful properties. Second, they create sophisticated computational and experimental methods to advance research and development using scientific images. Third, they simulate life cycles of whole assets or of critical components by using advanced computational methods and modelling of fluid flow, reactor engineering and erosion modelling.
Shell uses computational science to research more efficient electrolysis technologies, which in turn contributes to lowering the production costs for green hydrogen.
Hydrogen can be produced from various primary energy sources and by various technical processes. For example, an electrolyser splits water into hydrogen and oxygen. The production of hydrogen from electrolysis using renewable power produces almost no overall greenhouse gases emissions. The final product is known as green hydrogen. For now, electrolysis is more expensive than conventional hydrogen production methods, but - with anticipated process improvements - it is expected to offer significant cost-saving potential.
The chemical process of electrolysis poses some challenges that are better understood at a molecular level. Traditionally, material discovery, synthesis and selection for business applications has been driven by laboratory experimentations. Computational science has supported the work of experimentalists by providing predictions of the performance of materials based on the properties of their atomic structure. These predictions are based on foundational physics and chemistry computations. Modelling chemical reactions rather than testing them in laboratories saves time and resources and is a much faster way to test multiple hypotheses in parallel. Computational scientists in Shell have developed a digital model to understand the effect of the roughness of electrolyser surfaces in boosting hydrogen production.i
This research could be conducted in-house because Shell maintains a top-tier HPC facility. In fact, these detailed, molecular level computations, involve more than 100,000 interacting particles. Simulating this chemical reaction required a computing capacity of about 800,000 core hours. A core-hour is a measure of computation time, whereas a core refers to a processor that can independently perform or process computational tasks.
This means that 800,000 cores needed to run the algorithm together for one hour to solve this calculation. Without HPC, that would mean 100,000 domestic laptops working together for one hour, which would hardly be practical. It is not just the scale of HPC resources that was crucial for this simulation, but also the researchers’ expertise in code optimizations. They tailored the publicly available software GROMACS to the need of their simulation to make the code run faster, thus lowering the costs of the project.
Shell’s differentiated capability in physics-based artificial intelligence models is proving essential in optimizing chemical reactions to produce green hydrogen.
Shell is assessing the feasibility of large-scale electrolysis for its own assets as a starting point for deployment in industry. Identifying the best technical solutions early on with computational models at a systems level has the potential to unlock cost savings estimated in millions of dollars. This can contribute to making green hydrogen price competitive against other generation processes such as steam methane reforming.
Shell has been developing an integrated systems modelling capability for years to optimise the operations of its assets and processes, for example for Gas to Liquids operations. Our in-house ability to evaluate and analyse scenarios has increased manifold by integrating HPC infrastructures and code optimization, which has enabled research teams to run more scenarios in less time. This modelling capacity could now be used to run techno-economic integration studies to understand how to best integrate electrolysers within existing assets.
In an asset electrification study, Shell computational scientists modelled more than 1 million individual optimization options to assess in parallel multiple possible combinations and scenarios. This simulation utilised over 300,000 core hours. This system modelling capability powered with HPC is essential to understand how to best deploy electrolysers at Shell assets - for their own consumption - or at non-Shell assets. Scaling up the use of hydrogen as a fuel for industry is expected to be crucial to reduce the intensity of carbon emissions of industrial processes for which there are few other technical alternatives.
Another key aspect to the hydrogen economy is the safe storage and transport of liquid hydrogen fuel at large scale (around 20,000 m3). Unlike electricity, hydrogen can be stored in large amounts for extended periods of time. For that reason, liquid hydrogen produced on an industrial scale and transported via ships to the consumer could play an important part in the energy transition. In Japan, Shell is participating in a research project to develop a hydrogen energy supply chain, which includes the shipping of liquefied hydrogen.ii
But the safe storage and transportation of hydrogen - especially in large volume - presents several technical challenges to date. Shell’s researchers are addressing some of them, for example, by helping design new containers for hydrogen shipping. Using modern computational science methods, they developed an accurate thermodynamic model that can help improve the designs of the incumbent compressor technology.iii
The aim of the research is to increase the safety of operations and consequently enable a safer, faster and wider adoption of hydrogen as an energy carrier. The thermodynamic modelling capability in Shell is a deep expertise, developed through decades as a sub-discipline of reservoir simulation in the context of conventional oil and gas exploration. In relation to the storage of hydrogen, the same capability is being leveraged to predict – for safety purposes - the stability of a system where water and hydrogen meet. Pure hydrogen - which is required for storage due to safety reasons - burns almost invisibly if it is brought together with oxygen (a water molecule is made of one oxygen atom and two hydrogen atoms) and an ignition source. Modelling a fluid’s interaction between hydrogen and water is not a trivial modelling exercise, but the insights derived from it are essential to develop new container designs which can make the storage and transport of hydrogen safer.
Shell’s capability in high performance computing makes digitalisation an enabler of the energy transition. HPC helps lower the carbon intensity of today’s energy products and advance the clean technologies that will be central in tomorrow’s decarbonised energy system.
Suchismita Sanyal is the General Manager, Computational Science at Shell Technology Centre Bangalore. She leads a group of 60 researchers in the Computational Science group in Shell, delivering digital solutions across multiple Shell businesses and designed to position Shell on a stronger foothold in the energy transition.