Yokohama Rubber develops AI-powered tire mold design support system

Yokohama Rubber Co., Ltd. has independently developed a tire mold design support system that integrates simulation (FEM: Finite Element Method) and AI technology, which will be implemented in April 2026. This system will provide information that complements the knowledge and experience of engineers (trends in tire characteristics changing with changes in mold design factors) based on a vast amount of virtual experiments. As a result, even less experienced engineers will be able to easily design molds, leading to improved development speed, cost reduction, and mold design with fewer rework issues. Furthermore, by understanding the relationship between mold design factors and tire characteristics from multiple perspectives, new insights and inspirations can be gained, leading to the development of higher-performance products.

This system was developed based on the AI utilization concept "HAICoLab*" formulated by the company in October 2020, with the aim of further innovating the development process. In mold design, which affects tire characteristics, conventionally, trial and error through prototyping and evaluation was necessary to understand the relationship between mold design factors and tire characteristics, which required a great deal of time and cost. In addition, there were challenges such as individual differences in the accuracy and time required for mold design, as it relied heavily on the knowledge of experienced engineers.
*This term is a neologism based on "Humans and AI collaborate for digital innovation," and also embodies the meaning of a joint research institute between humans and AI

. This system combines "automation of simulation" and "prediction and visualization by AI" to solve these challenges. First, it automatically generates numerous tire FEM models with varying mold shapes and performs tire characteristic calculations in a virtual space. Next, it uses the calculation results as training data to construct an AI model (surrogate model) that instantly predicts the relationship between mold design factors and tire characteristics. By applying XAI (eXplainable AI) technologies such as SHAP (SHapley Additive exPlanations) and PDP (Partial Dependence Plot) to this AI model and quantitatively visualizing the influence of mold design factors on tire characteristics, engineers can easily obtain clear guidelines on "which mold design factors to change and by how much to achieve the target characteristics."

Yokohama Rubber has previously developed an environment in which engineers can easily obtain design guidelines through tire characteristic value prediction systems (2021) and tire design support systems (2024). The tire mold design support system developed this time is a subsequent initiative based on "HAICoLab." We will continue to promote further enhancements to this development environment to accelerate innovative tire development.