The research findings, which explore the potential of large language models (LLMs) for data analysis, modelling, and predicting the time-dependent behaviour of key coating system parameters, were introduced by Kolarik in a presentation entitled Predictive Analytics of Aluminide Diffusion Coatings Using Machine Learning to Forecast Their Aging and Service Life. Further details are expected to be published in an upcoming scientific paper.
Aluminide diffusion coatings offer a reliable and cost-effective way to protect steel from high-temperature corrosion in harsh environments. These coatings can be applied in the form of aluminium slurries using various deposition methods, including spraying and brush application, before undergoing heat treatment to create a diffusion layer. Predictive analytics based on machine learning methods offers significant potential for forecasting the ageing and service life of the coating under operating conditions.
“Our research demonstrates that machine learning methods are highly effective for analysing complex material systems influenced by multiple parameters, where conventional physical models often struggle to capture relationships and the significance of individual parameters,” said Kolarik.
ICMCTF is one of the world’s leading international conferences dedicated to thin-film deposition, characterisation, and advanced surface engineering. The event brings together scientists, technologists, and industry representatives from around the globe to share the latest research findings, expertise, and emerging trends in the field.
Text: Martina Šaradínová
Photo: Archive V. Kolarik