Yokohama Rubber develops AI-based tire mould design system

Hiratsuka, Japan — Yokohama Rubber Co. (YRC) has developed a proprietary tire mould design support system using simulations and artificial intelligence to improve development efficiency and reduce reliance on trial-and-error processes.

The system combines finite element method (FEM) simulations with YRC’s own AI technology and is designed to “supplement the knowledge and experience” of development staff, said the Japanese group 30 April.

According to YRC, the system makes it easier for less-experienced staff to design new moulds by providing information on “changes in tire characteristics due to different mould design factors.”

The information will be based on a large number of virtual experiments, YRC explained.

The platform is expected to “increase mould development speed, while reducing costs and the reworking often required to realise a new mould design.” 

YRC noted that mould design greatly affects the key characteristics of tires.

Until now, obtaining a thorough understanding of the relationship between mould design factors and tire characteristics has required “a costly and time-consuming trial-and-error process including trial production and follow-up evaluations.”

The process, YRC explained, has relied heavily on the know-how of experienced staff to design and develop moulds accurately and in a timely manner.

The new system, however, automatically generates a large number of FEM tire models with different mould shapes and calculates tire characteristics in a virtual space.

The results are then used as training data to construct an AI “surrogate model” that “instantly predicts the relationship between mould design factors and tire characteristics.”

YRC said the application of explainable AI technologies, including SHAP and Partial Dependence Plots (PDP), enables development staff to “quantitatively visualise the impact of mould design factors on tire characteristics.”

This, it noted, allows the designers to determine how much to adjust each factor to achieve target performance.

The system was developed under YRC’s HAICoLab – human and AI collaborate – programme, launched in October 2020.

YRC previously used the HAICoLab concept to develop a system to predict key tire characteristics in 2021 and a tire design support system in 2024.