What is a real time digital twin?

August 2024

On the 6th August Dr Andrea Novoa presented her research work ‘a real-time digital twin of a hydrogen-fuelled annular combustor’ as part of the AI for Net Zero webinar series.  

Dr Novoa’s webinar gave us an insightful view into what is meant by a real-time digital twin by starting off her presentation by giving us a succinct definition:  

“a real-time digital twin is a set of virtual information constructs that mimic the structure context and behaviour of a real system. It is dynamically updated with data from the physical twin. It is the bidirectional interaction between the virtual and physical that is central to the digital twin.” 

Continuing with the webinar, Dr Novoa explains to us about the complexities and challenges of building such systems.

Dr Novoa paints the picture of the current situation that focuses on low fidelity models at low computational costs, that put data and physical models together but use many assumptions and thus add bias and/or error to the output.  

This is where Dr Novoa is stepping in with her bias-aware real-time data assimilation tool that puts everything together in a real example of a real-time digital twin of a hydrogen-fuelled annular combustor. 

This is a very complex system consisting of multi physics, and a combination between acoustics flame dynamics, aerodynamics, and the turbulent environment of the combustion chamber.  

In this webinar Dr Novoa explains how she collected the raw data she needed to create the latest digital twin of the annular combustor contributing to the overall success of the AI for NetZero Energy & Transport project. 

This webinar, along with the entire series, can be seen on the AI for NetZero YouTube channel.


 

Dr Georgios Rigas is joining the AI for Decarbonisation’s Virtual Centre of Excellence’s Expert Working Group

AI for Decarbonisation’s Virtual Centre of Excellence (ADViCE) is aimed at the development of innovative artificial intelligence technologies for decarbonisation applications to support the transition to net zero.  

The Centre's ambition is to foster a cross-sector collaboration engaging the high-emitting sectors and technical experts that can inform stakeholders how AI solutions are essential in the move to Net Zero.  

The sector-agnostic working group, that Dr Rigas is forming part of, aims to deliver recommendations to drive the broader Department for Energy Security and Net Zero’s  (DESNZ) AI for decarbonisation interventions aligned to 3 critical themes: 

  • AI for decarbonisation policy and regulation alignment 
  • Unlocking and enabling investment and innovation 
  • Data accessibility and capability 

To find out more about ADViCE and how you can get involved visit the programme website: AI for Decarbonisation’s Virtual Centre of Excellence (ADViCE) - Digital Catapult | Digital Catapult (digicatapult.org.uk) 

 

Dr Georgios Rigas presents in AI for Net Zero webinar series

July 2024 

As part of the AI for Net Zero webinar series Dr Georgios Rigas presented a summary of his research theme focusing on real-time digital optimization and decision-making for fluid-based energy and transport systems. 

Dr Rigas is an expert in the fields of aerodynamics and artificial intelligence (AI). Having studied both topics separately he found that AI was exactly the tool he needed to answer problems he found in his aerodynamic studies. This research project truly bridges the gap between fluid-based energy and transport systems and real-time digital optimization (RL), also known as decision-making tools.   

The webinar goes into the details of how algorithm learnings can increase the aerodynamic efficiencies of heavy goods vehicles, how it can increase energy production of wind farms, and the role it’s playing in effective hydrogen combustion.   

The research is steaming ahead, and Dr Rigas reported that the next most important steps are to reinforce algorithm learnings, especially for windfarms, by setting up pilot test sites to see how the theoretical situations perform in real environments. 

Dr Rigas wrapped up his presentation by inviting us to get in touch with advice about standardising guidelines for scientific machine learning, a piece of work that he is currently putting together. He also tells us to watch out for a specialised training workshop on machine learning and RL for physical systems, that should be happening later this year. 

The audience was rather active, asking questions about real-life fuel savings, developments in physics, and improvements in parameterization. 

Dr Rigas’ webinar can be rewatched the AI for NetZero YouTube channel along with the entire series.     


 

AI for Net Zero work presented to the UK Department for Energy Security & Net Zero

Dr Georgios Rigas and Prof. Sylvain Laizet presented the AI for Net Zero project to the UK Department for Energy Security & Net Zero (DESNZ).  

On 20 May 2024, Dr Georgios Rigas and Prof. Sylvain Laizet attended a meeting held by Deputy Director of DESNZ, Anna Stephenson, and in conjunction with the Imperial Policy Forum, to present some of the initial outcomes of the AI for Net Zero project.  

Dr Rigas and Prof. Laizet started the presentation with an overview of the project, and its objective of merging the digital world with the physical world to increase efficiencies in existing road transport and wind energy technologies.   

Another very interesting topic presented during this session was about the growing need for Information and Communication technologies (ICT) and its enormous use of electricity and generation of e-waste. Dr Rigas and Prof. Laizet explored the use of AI in delivering carbon neutral digital research infrastructure (DRI).  

The presentation received an amazing turnout of more than 30 government representatives and is thought to be just the beginning of further knowledge sharing sessions with policy makers of this kind. 

If you would like to find out more the presentation can be downloaded here: Rigas_Laizet_AI-NetZero.pdf 


 

AI for Net Zero Webinar Series

June 2024  

The AI for Net Zero community is pleased to announce the launch of the AI for NetZero webinar series

This webinar series brings together experts, innovators, and enthusiasts from both artificial intelligence and sustainability domains. AI’s role in the progression towards a NetZero society is essential, so please do join our series to find out more about groundbreaking research and developments in this field.  

About  

The AI for Net Zero webinar series is covering all the different areas of research that the UKRI-funded AI for NetZero projects are exploring, including energy, transport, environment, agriculture, and food systems. 

During this series we will hear from senior academics and PDRAs delving into all the diverse applications of artificial intelligence and beyond. 

When: Every Tuesday at 2pm (UK time). 

Where: Future and past webinars can be viewed on the AI for NetZero YouTube channel 

Find out more about the webinar series and who’s presenting next on the AI for NetZero Webinar Series website. 


 

Second AI for Net Zero Project Advisory Board Meeting

April 2024

The "AI for Net Zero" team successfully concluded its second Advisory Board meeting on April 23, 2024, hosted at Oxford University. The session brought together board members to review the project's recent advancements and strategize on future initiatives.

Principal Investigators delivered insightful presentations on the progress of their respective work packages. These presentations sparked a dynamic roundtable discussion involving external board members from both industry and government sectors, including representatives from Nvidia, Atkins, Catesby Projects, EngSys, and the Department of Energy Security and Net Zero.

The next Advisory Board meeting is scheduled for September 2024 where further developments and strategic plans will be discussed.


 

Dr Andrew Mole wins national presentation prize for turbulence research

April 2024

Andrew Mole, Research Associate, won first prize at the UK Turbulence Consortium annual meeting for his presentation on wind turbine adjustments for increased power production. Using cutting-edge AI and computational methods, Andrew presented an innovative approach aimed at maximising power generation from wind farms, a critical step in transitioning away from fossil fuels.

Dr Andrew Mole said: 'What we're doing in combining AI with high-fidelity turbulence simulations is really exciting! I'm looking forward to continuing this work to increase the efficiency of wind farms and to generate more clean energy."

Find out more here


 

Team presents at Machine Learning for Fluid dynamics, Paris

March 2024

Three members of the AI for Net Zero presented their recent results on real-time digital twins, reinforcement learning, and wind farm optimisation in the first ERCOFTAC Workshop on Machine Learning for Fluid Dynamics at the Sorbonne University in Paris.

The titles of the talks were: 

Real-Time Inference of Model Errors from Experimental Data: Application in Hydrogen-Based Annular Combustors - Andrea Novoa, Nicolas Noiray, James R Dawson and Luca Magri.
Reinforcement Learning of Active Flow Control in Partially Observable Environments - Max Weissenbacher, Anastasia Borovykh, Georgios Rigas.
Multi-fidelity Bayesian Optimisation for the Control of Wind Turbines based on LES and Wake models - Andrew Mole, Sylvain Laizet.

First AI for Net Zero Project Advisory Board Meeting

September 2023

On 26 September 2023, the AI for Net Zero project team gathered in the Department of Aeronautics, Imperial College London to host the first Advisory Board meeting. We were joined by our Advisory Board, made up of representatives from our industrial partners and the UK Government. The investigators gave presentations on their work packages which sparked some excellent discussions, drawing out challenges, opportunities and synergies with the Government and industry thinking. We discussed many areas including the state of preparedness for hydrogen fuel in the aviation industry, the challenges of obtaining wind farm data, promoting the advantages of digital twins in industry, conditions for vehicle wind tunnel experiments and the added value that including policy brings to AI for Net Zero.

The next Advisory Board meeting will be held in April 2024 in Oxford.


 

Environmental Audit Committee Evidence paper 'Sustainable Electrification of the UK Economy'

June 2023

In June 2023, AI for Net Zero submitted an evidence paper to the UK Parliament’s Environmental Audit Committee’s call for evidence entitled ‘Enabling sustainable electrification of the UK economy’. The Inquiry aimed to examine the challenges and opportunities arising from the increasing use of electricity to power the economy of Net Zero Britain. Our evidence paper detailed how AI for Net Zero, in particular our work on wind farms and real time control systems, will contribute to the sustainable electrification of the UK economy.


 

AI for Net Zero awarded funded to help the UK's energy and transport sectors achieve carbon net zero

June 2023

The AI for Net Zero project ‘Real-time digital optimisation and decision making for energy and transport systems’, a collaboration between Imperial College London, University of Edinburgh, University of Oxford and the University of Cambridge, received the good news that it had been successful in its bid for funding. Secretary of State for Science, Innovation and Technology, Chloe Smith MP, announced the successful projects for Engineering and Physical Sciences Research Council (EPSRC) funding as part of a £54 million government investment to develop trustworthy artificial intelligence (AI) research.

Read the full news article on Imperial College London’s news site.


 

Dr Georgios Rigas attended Evidence Week 2022

November 2022

Dr Georgios Rigas attended Evidence Week 2022 and briefed MPs on how Imperial's aerodynamics research will improve future road and air vehicle designs. Dr Rigas’s research was featured in one ‘pod’ of the Evidence Week 2022 exhibition where he spoke to MPs and policy makers about the huge savings that better aerodynamic design will mean for vehicles – both in terms of savings on fuel costs, and on the associated reductions in harmful emissions.

Read the full news article on Imperial College London’s news site.