DUNLOP and Fujitsu Cut Tyre Analysis Time by 90% With AI

DUNLOP and Fujitsu have developed an AI surrogate model that can predict tyre deformation behaviour far faster than conventional finite element method analysis. The companies say a proof of concept reduced structural analysis time by about 90%, cutting computation from around 45 minutes to five while handling approximately 600,000 mesh elements.

A faster route through tyre simulation

The project focuses on one of the most demanding stages in tyre development: modelling how a tyre deforms when it contacts the road surface. DUNLOP and Fujitsu used DUNLOP’s design expertise and real design data to train a graph neural network model, allowing it to predict results normally obtained through finite element method analysis.

The companies said the model reached an average accuracy of 87.7% when predicting contact shape against conventional FEM analysis. In practical terms, the work could help tyre engineers reach earlier decisions on structure, materials and contact characteristics. It may also reduce the number of repeated design cycles before physical validation begins.

The development sits alongside wider industry interest in AI-enabled design. Tyre News has recently reported on Yokohama’s use of AI to speed tyre mould design, while broader sector coverage has examined how tyre manufacturers are using AI across design, production and logistics.

‍

‍

‍

‍

Why it matters for DUNLOP

For DUNLOP, the timing is significant. Sumitomo Rubber Industries has been rebuilding and repositioning the DUNLOP brand across key regions after acquiring important four-wheel tyre rights from Goodyear. Tyre News previously reported that the transaction transferred Dunlop rights across consumer, commercial and specialty tyre segments in specified markets.

This makes faster development capability more than a laboratory gain. It could support a wider product renewal cycle, including tyres for electric vehicles, premium passenger cars and future original equipment programmes. Recent Tyre News coverage of the Blue Response TG summer touring tyre noted its role as an early European product under Sumitomo’s DUNLOP stewardship.

From proof of concept to design tool

DUNLOP and Fujitsu now plan to develop the technology into a design support tool for practical use by DUNLOP designers. The target for implementation is April 2027. Before then, the companies plan to verify the model on a prototype of Fujitsu’s next-generation Arm-based FUJITSU-MONAKA CPU by December 2026.

The companies said the system is intended to improve inference speed, accuracy and power efficiency. Fujitsu also plans to use the project as a route into wider manufacturing applications, including large-scale FEM analysis for automotive and other industrial sectors.

DUNLOP said the work supports its long-term R.I.S.E. 2035 direction and its aim to create “new experiential value” from rubber technologies. Fujitsu said the initiative could also support lower-energy AI inference platforms for manufacturing.

The tyre trade relevance

For tyre manufacturers, the main value lies in faster engineering decisions rather than headline AI claims. Advanced tyres now need to balance wet grip, rolling resistance, noise, durability, EV load demands and material efficiency. Each added requirement increases the pressure on simulation, testing and development cost.

If the model proves reliable beyond the proof-of-concept stage, it could help reduce bottlenecks in early-stage structural analysis. That would matter to manufacturers seeking shorter development cycles without weakening validation standards.

‍

‍

Tagged with: DUNLOP tyres, Fujitsu, AI tyre analysis, tyre structural analysis, FEM analysis, CAE simulation, graph neural network, tyre design, Sumitomo Rubber Industries, EV tyres, tyre development, FUJITSU-MONAKA

‍