In this article, we meet Akselos, an advanced engineering specialist technology company who specialise in predictive digital twin technology for structural integrity management, and have built the world’s largest digital twin.
Tell us about Akselos…
At Akselos we take incredibly complex structures and simulate them, building extremely detailed models using a technology we call Reduced Basis Finite Element Analysis (RB-FEA). The results are used to create a predictive digital twin or a ‘Digital Guardian’ as we prefer to call it using sensors, big data analytics and simulation. Our proposition is based on a simulation running in parallel with the in-service operation of an asset using a physics-based digital twin.
A prime example of this is a Floating Storage Production and Offloading (FSPO) which is essentially a converted oil tanker with a refinery, commonly used in oil production. Due to the complex nature of the design it’s very difficult to simulate using classical finite element methods, which doesn't have the capacity to simulate systems and structures of this magnitude. Using our RB-FEA technology, which is typically 1000x faster than conventional commercial methods, we can solve something like the example in less than 10 seconds.
How did the technology and subsequently Akselos form?
RB-FEA, was originally pioneered by the Massachusetts Institute of Technology (MIT) around 12 years ago as part of a research programme funded by the US Department of Defence. In 2012 Akselos was formed as a spin off by three technologists involved in the original research at MIT; David Knezevic, Thomas Leurent and Phong Huynh. The company has grown and we now have 3 main locations; Boston, Lausanne and Ho Chi Min City as well as representation in London, Aberdeen and Houston. The company has secured further investment funding in 2018 from Shell Ventures and Innogy Ventures, now employing around globally.
Tell us more about the RB-FEA technology?
RB-FEA is a completely new mathematical approach to solving the fundamentals of the Finite Element Method, it goes back to the equations of elasticity and how to solve them. The method is inherently parametric, meaning that we can define parameters like thickness, length or the position of a crack and update the solution very quickly. We take a structure and break it into structurally significant components, which are essentially small finite element models. We then run a pre-analysis of all these components in parallel, very quickly on the cloud, before connecting all the components together, solving the model in seconds - fast and accurate.
The concept of a digital twin isn’t new, however Askelos is different because it isn’t limited by scale. Our digital twin technology has the ability to model in service large-scale assets in the level of detail required to understand structural conditions and enable physics based predictive analytics. We believe the RB-FEA technology is completely unique. What this provides the customer with is a real time snapshot of the condition of its asset, enabling predictive and preventative maintenance, reducing downtime, cost and increasing operational efficiency.
What sectors do you work with?
Shell Ventures and Innogy Ventures are our two largest investors so we have a big focus on the energy sector, in particular oil, gas and wind. That said, the RB-FEA technology is generic and can be applied to any asset-centric sector. We have recently announced a new project sponsored by Eurostars involving the optimisation of wind turbine support structures with the objective of reducing materials costs by 25%.
What is Akselos’ focus going forward?
Growth is our primary objective. Up until now we have successfully positioned our technology towards the in-service segment, because that was why it was developed, and in the energy sector because that is where the big critical assets are. We plan to grow our position in upstream oil and gas and diversify in other industries, including marine, aerospace and defence and become established in the design process.
We hope to expand the use of simulation in digital twin technology. Currently simulation is often an offline activity due to the time it takes to run the models, and thus cannot be used for real time decision making. We want to bring this online.
What accomplishment is Akselos most proud of?
It would have to be our Digital Guardian concept. Imagine you have built a digital twin replica of a bridge and you can run real time simulations. Cars are using the bridge and the sensors are showing what is happening. Suddenly, the sensors begin to pickup unusual activity and you need to discover why. Our Digital Guardian can help monitor this and point you to the exact reason that sensor is reacting in that way. The whole idea is to have a simulation running in parallel with the in-service operation of an asset using a physics-based digital twin to protect the asset and the people using it.
We have also built a full digital twin model for a 300,000 ton FSPO which is currently in operation for Shell. We are led to believe this world’s largest predictive digital twin which we very are proud of.