Project Details
Projekt
Data-informed probabilistic prediction of concrete creep considering the hygral and mechanical history of existing structures (CreepStatus)
Applicant
Professor Dr.-Ing. Michael Haist
Subject Area
Structural Engineering, Building Informatics and Construction Operation
Construction Material Sciences, Chemistry, Building Physics
Construction Material Sciences, Chemistry, Building Physics
Term
since 2025
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 562810390
An accurate prediction of the creep strains of concrete is essential for the structural assessment of deflection-sensitive structures such as prestressed box girder bridges. Creep prediction models today are quasi exclusively focused on how a defined concrete will deform during its (future) service life. As at the time of planning the specific concrete composition is unknown, the prediction heavily relies on assumptions on the concretes mean compressive strength and very simplified assumptions on the future environmental and loading conditions. Scatter in the material properties or naturally occurring variations in the environmental (climatic) and/or loading conditions are completely neglected.Following the general goal of SPP 2388 – and here especially the research focus “Status Indicators and Models” – the proposed project “CreepStatus” intends to develop a data informed methodology for concrete creep prediction especially of old, historic concrete structures. Contrary to existing deterministic approaches such as in fib Model Code 2020, the main creep influencing parameters – i.e. concrete properties/composition, climatic and loading conditions – are introduced in a probabilistic manner. In close collaboration with other members of SPP 2388 in a first step, a methodology for calibrating a digital material twin (creep twin) will be derived which will rely on data obtained from minimally-invasive miniature samples, thus protecting the structure. In doing so, local variations in concrete properties and creep characteristics can be identified. In a second and third step, the statistical variations of (cyclic) climatic conditions as well as (cyclic) mechanical loading, respectively, and their influences onto concrete creep are incorporated into the model. Herefore, data from on-site bridge measurements (a.o. obtained in funding phase 1) as well as from weather measurements are analysed for their characteristics with regard to creep-causing deformations and incorporated in the model using probability density functions. With regard to the model the author can built up on physics-informed creep models, developed in previous DFG projects. The overall prediction accuracy of the improved model is finally tested by incorporating the latter into Finite Element Simulations and comparing the predicted deformations with measurement results obtained on the bridge. Further, a close collaboration has been agreed upon with the Dutch TNO, who carry out similar digital twinning in the Netherlands.
DFG Programme
Priority Programmes
