Dunlop, a model underneath Sumitomo Rubber Industries, Ltd., and Fujitsu Limited have introduced the event of an AI surrogate mannequin aimed toward predicting tire efficiency with excessive accuracy and diminished evaluation time. This collaboration marks a big step in tire design digital transformation, decreasing evaluation time from 45 minutes to roughly 5 minutes, whereas sustaining accuracy.
The AI surrogate mannequin, developed as a part of Dunlop’s long-term administration technique, was utilized to the structural evaluation of tire deformation involved with highway surfaces. Utilizing Fujitsu’s AI expertise and Dunlop’s tire design experience, the mannequin relies on a Graph Neural Network algorithm, reaching a mean accuracy of 87.7% in comparison with conventional Finite Element Method (FEM) evaluation.
This new expertise runs on Fujitsu’s FUJITSU-MONAKA, an Arm-based CPU designed for top efficiency and vitality effectivity. The corporations plan to optimize inference velocity, accuracy, and energy effectivity by testing the mannequin on a prototype of FUJITSU-MONAKA by December 2026.
The initiative goals to allow quicker decision-making in tire design, decreasing the variety of processes required to find out tire construction and materials specs. By April 2027, Dunlop plans to implement this expertise in sensible operations, accelerating data-driven growth and bettering tire security and environmental efficiency.
Fujitsu intends to use this AI expertise throughout different manufacturing industries, selling carbon neutrality by means of enhanced energy effectivity. The firm will provide the AI inference platform, combining FUJITSU-MONAKA and the Graph Neural Network, by way of its AI platform Fujitsu Kozuchi.
The collaboration aligns with Dunlop’s company technique, R.I.S.E. 2035, and its imaginative and prescient of offering new experiential worth from rubber. Fujitsu, dedicated to sustainability, views this growth as a way to optimize growth processes and improve industry-wide effectivity.