Projekt je zaměřen na tvorbu obecného algoritmu pro identifikaci únavové křivky s využitím adaptivního statistického vzorkování a metody rozvoje polynomiálního chaosu. V průběhu projektu bude na české straně vyvinut numerický algortimus a na německé straně budou provedeny laboratorní experimenty pro validaci matematických modelů.

Description in English
During the proposed project, a novel methodology for the adaptive sequential construction of characteristic fatigue curve of material or product, reflecting uncertainties affecting the experimental program. The proposed method will be based on the theory of probability and polynomial chaos expansion. The developed theoretical method will be further implemented into a Python numerical algorithm. The numerical algorithm will serve for the identification of optimal experimental design for sequential improvement of the accuracy of identified fatigue curve, leading to a significant reduction of the number of laboratory experiments while maintaining the accuracy of identification. Laboratory experiments for the construction and verification of the proposed algorithm will be conducted by German team members, who have rich experience in this topic. Initially, the optimal number and design of initial laboratory experiments will be determined in cooperation with Czech team, reflecting specific conditions of experiments and variability of material parameters of specimens leading to uncertainty in the fatigue curve. The obtained results together with data from the literature will be utilized for the development of the method by Czech team. Once available, the algorithm will be verified using laboratory experiments for the identification of a fatigue curve and results will be compared to the existing approach.The theoretical background of the developed method and its implementation to the numerical algorithm will be presented in 2024 at international conference and published in conference proceedings indexed by Scopus/WoS – a result type D according to IS VaVAI – RIV. Experimental applications and verification of the method will be presented similarly in 2025 – a result type D

Keywords
aktivní učení, rozvoj polynomiálního chaosu, únava materiálu, kvantifikace nejistot

Key words in English
active learning, polynomial chaos expansion, material fatigue, Uncertainty Quantification