Numerical benchmark campaign of COST Action TU1404 – microstructural modelling

• Published in: RILEM Technical Letters Vol. 2; pp. 99 - 107
• Main Authors: Wyrzykowski, Mateusz, Sanahuja, Julien, Charpin, Laurent, Königsberger, Markus, Hellmich, Christian, Pichler, Bernhard, Valentini, Luca, Honório, Túlio, Smilauer, Vit, Hajkova, Karolina, Ye, Guang, Gao, Peng, Dunant, Cyrille, Hilaire, Adrien, Bishnoi, Shashank, Azenha, Miguel
• Format: Journal Article
• Language: English
• Published: International Union of Laboratories and Experts in Construction Materials 21-09-2017; RILEM Publications SARL
• Subjects: Benchmark; Civil Engineering; COST TU1404; Engineering Sciences; Mechanics; Microstructure; Numerical modelling; Benchmark; COSTTU1404; Numericalmodelling; Microstructure
• ISSN: 2518-0231; 2518-0231
• DOI: 10.21809/rilemtechlett.2017.44

This paper presents the results of the numerical benchmark campaign on modelling of hydration and microstructure development of cementitious materials. This numerical benchmark was performed in the scope of COST Action TU1404 “Towards the next generation of standards for service life of cement-based materials and structures”. Seven modelling groups took part in the campaign applying different models for prediction of mechanical properties (elastic moduli or compressive strength) in cement pastes and mortars. The simulations were based on published experimental data. The experimental data (both input and results used for validation) were open to the participants. The purpose of the benchmark campaign was to identify the needs of different models in terms of input experimental data, verify predictive potential of the models and finally to provide reference cases for new models in the future. The results of the benchmark show that a relatively high scatter in the predictions can arise between different models, in particular at early ages (e.g. elastic Young’s modulus predicted at 1 d in the range 6-20 GPa), while it reduces at later age, providing relatively good agreement with experimental data. Even though the input data was based on a single experimental dataset, the large differences between the results of the different models were found to be caused by distinct assumed properties for the individual phases at the microstructural level, mainly because of the scatter in the nanoindentation-derived properties of the C-S-H phase.