Project start: 1.10.2025.

The rapid development of industry and engineering materials requires faster, more accurate, and more reliable methods for estimation of their properties and mechanical behavior. Traditional methods, which rely on experimental testing and empirical models, are often expensive, time-consuming, and limited to specific testing conditions. In contrast, machine learning techniques offer new opportunities to enhance data analysis, reduce the need for physical testing, and develop estimation models with greater precision.
This project aims to bridge the gap between knowledge of the mechanical behavior of materials and advanced machine learning methods, with the goal of developing models for the reliable estimation of material properties and behavior. Existing approaches will be analyzed, key challenges such as data availability and model interpretability will be explored, and a systematic framework will be set to enable more accurate predictions and better adaptability to industrial needs.
The methodology development will combine experimental data with advanced machine learning algorithms. Special emphasis will be placed on understanding which parameters are critical for describing the mechanical behavior of materials and how they can best be used in modeling. The evaluation of results will be a crucial step in ensuring the reliability of estimations, with additional focus on model interpretability to enable researchers and engineers to make informed decisions based on the results.
In addition to its scientific contribution, the project also holds broader social and economic contribution. The developed tools will enable faster and more efficient material assessment, potentially leading to reduced manufacturing costs, optimized component designs, and increased safety of structures.
In the long term, this project will lay the foundation for the broader application of machine learning in engineering practice, contributing to the development of more efficient, predictable, and environmentally sustainable approaches to estimation of the properties and mechanical behavior of materials.