An artificial intelligence-based (AI) tire performance prediction system has been developed by Nexen tire.
With the system, the tire maker will utilize machine learning technology in the concept design stage, enabling the fast and accurate anticipation of primary performance indicators during tire development, including noise, handling and fuel efficiency.
To ensure big volumes of high-quality data are available for machine learning, Nexen has created data pre-processing technology capable of detecting and replacing irregularities in protected data. The substantial amount of learning data will enable the tire maker to secure a forecasting model with good predictive performance for insufficient data.
Being able to forecast tire performance earlier in the development stage will affect the quantity of prototypes produced and the development time. Usually, finite element analysis (FEA) is used to forecast the performance of tires, namely the evaluation of high-precision performance estimates but the system can take a long time to calculate figures, making it inefficient.
By utilizing the new tire performance prediction system, Nexen aims to make its tire design processes faster, more accurate and with an increased number of performance improvements during the pre-production stage.
“We aim to finalize the development of the Virtual Brain Loop System, a tire development system based on our own virtual design technology, and apply it to OE and RE goods,” explained Seong Rae Kim, researcher at the Nexen univerCITY, the company’s central R&D institute. “Through combined industry-academic research, we intend to increase talent training and R&D skills.”
