60 Second Spotlight on Georgios Mavros, Loughborough University
Ahead of the Tyre Testing & Modelling Seminar at HORIBA MIRA on 20 March, we caught up with Georgios Mavros to hear his thoughts on the topic.
Briefly explain your current role.
I am a professor of vehicle dynamics at Loughborough University. My research focus is tyre modelling and parameter identification, covering many different aspects of tyre behaviour including vibration, friction and rolling resistance. I work within a group of five academics with complementary research interests and our aim is to apply cutting-edge engineering science to real-life problems. I also teach vehicle dynamics, tyre modelling and vehicle design to our automotive students.
Why do you believe tyre modelling is critical for automotive innovation today?
The tyre is the only vehicle component that directly and significantly affects so many performance attributes of modern vehicles including fuel efficiency (or battery range), driveability, handling, NVH (noise, vibration and harshness) and safety. Today’s electric vehicles come with ultra-refined, high-torque electric powertrains and with an additional weight penalty due to battery packs. Such features have placed additional pressure on tyres to emit less rolling noise and to wear less. The desire for a longer driving range has also renewed our interest in tyre rolling resistance. As the majority of automotive OEMs work towards 100% virtual vehicle development, being able to predict all the above aspects of tyre performance in simulation is a key requirement. Especially when the aim is to optimise vehicle design, there is a need for tyre models to be both computationally efficient and of appropriate fidelity.
What is the most exciting development or trend in your area of tyre research or testing?
Tyre research is inherently a multi-physics affair underpinned by a large variety of modelling and experimental approaches. While efficient high fidelity multi-physics tyre models are now generally accessible, one area of physics where our modelling lacks despite the introduction of some promising theories in the past 30 years, is friction. In this respect, any further advancement in predicting friction as a function of rubber properties and road surface roughness is exciting. Besides that, tyre parameter identification is another very important and exciting topic. Models need parameters and parameters require testing, but testing is expensive, and every time testing is performed, the very purpose of creating a simulation model is somewhat defeated. For many years researchers have used machine learning techniques to estimate tyre parameters intelligently on-the-fly by monitoring vehicle states with cheap sensors, without the need to perform dedicated (and expensive) tyre testing. The renewed excitement around AI that has been sparked by large language models is contagious to tyre modelling and even more relevant now, since autonomous vehicles and ADAS systems depend on the on-line estimation of parameters such as tyre-road friction.
What do you see as the biggest challenge in tyre modelling and simulation?
This is a very difficult question since most aspects of tyre physics are challenging to predict. From a philosophical viewpoint perhaps, the greatest challenge would be the creation of a single truly multi-physics super fidelity model that is able to accurately capture the effect of any design or material parameter on any tyre performance attribute, from comfort and noise to friction and even the prediction of the size distribution of rubber particles emitted through wear. Such a model does not exist and if it did it would be impossible to simulate using current resources. Recent advances in quantum computing (for instance the Willow chip developed by Google) may greatly assist tyre engineers in this respect, in the medium future.
How do you see tyre modelling evolving over the next 10 years?
I suspect that, solid, yet incremental steps will continue to be made in all aspects of physical tyre modelling, but machine learning will play an increasing role in producing data-based tyre models. One of the problems with machine learning is that obtaining the necessary amount of data from dedicated tyre testing is prohibiting due to the enormous costs and time involved. But millions of vehicles drive millions of miles every year and in doing so large amounts of data are recorded. Tapping into such a resource from a tyre modelling perspective, for example by creating data-based tyre digital twins, is likely to be game-changing for tyre scientists and engineers.
How does sustainability influence your work or research in tyre testing and modelling?
Tyre testing and modelling usually refers to the process of testing and modelling those tyre attributes that affect vehicle performance. In this sense, other aspects of tyre engineering with a potential bearing on sustainability – for example the vulcanisation process or the recycling/retreading of tyres – are specific and interesting to the tyre manufacturing industry only. Adopting the conventional, vehicle-focused definition of tyre modelling/testing, sustainability is relevant mostly from a rolling resistance and tyre wear perspective. Unfortunately, rolling resistance and wear compete with other attributes such as grip, making tyre design a balancing act. For the traditional simulation engineer, optimising this balance is not within scope, but quantifying these attributes properly can be important. With Euro 7 regulations being in effect soon, we are now more interested in wear and, crucially, the size/shape distributions of the generated rubber particulates, as well as their chemical composition. Identifying how such distributions depend on tyre loading, driving style and road surface characteristics is one of the research streams in our group.
How important is collaboration between academia and industry in advancing tyre technology?
It is difficult to overstate the importance of collaboration between academia and industry in the field of tyre engineering. Our industrial partners are at the forefront of developments around tyre modelling and testing and a constant source of inspiration and real-life problems for us. In the past 20 years we were fortunate to collaborate with JLR, Michelin, Ford and Formula 1 teams and this has brought great insights and knowledge into our group. In return, our industry partners have benefited from the exploratory approach inherent in research but also from access to a pool of young talent, highly skilled in tyre modelling. We are about to start two new exciting projects with JLR and Formula 1 and we hope that these will continue the tradition of cutting-edge industry-relevant research.