Development of theory-guided supervised machine learning models for the characterization and prediction of shrinkage and creep of cement-based materials

With this collaborative project with BASF Construction Solutions GmbH, we are aiming at setting the bases for complementing poromechanics theory with machine learning models to address the general goal of predicting shrinkage and creep properties of cement-based materials.

Among the bases, we focus on building up an infrastructure for creep and shrinkage data curation. We adopt approaches/concepts/tools developed within the framework of the Materials Genome Initiative (MGI) of the USA Govt.

Among all the tools used, the most important is the open source Materials Data Curation System, developed and maintained by the USA National Institute of Standards and Technology (NIST).


Project Period

October 2018 – March 2021

Project Team

Project leader

Beat Münch, Nikolajs Toropovs, Zhangli Hu, Mateusz Wyrzykowski, Patrik Burkhalter, Fabian Bucher, Janis Justs, Pietro Lura