Project Details
Projekt
Structural design of reinforced concrete based on multiscale modelling and polymorphic uncertainty
Applicant
Professor Dr.-Ing. Michael Kaliske, since 3/2020
Subject Area
Applied Mechanics, Statics and Dynamics
Term
from 2016 to 2021
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 310979660
The realistic modelling of heterogeneous structures is essential for their numerical design and is mainly characterized by the mechanical model and the consideration of the available data by an adequate uncertainty approach. This description requires a realistic representation of the structure as well as the realistic assessment of loading situations, structural and material parameters. An adequate quantification is enabled by the consideration of polymorphic uncertainty. Scope of the research proposal is the numerical simulation and design of reinforced concrete structures (by steel, textile, carbon) with respect to time dependent behaviour and polymorph uncertain parameters. The efficient numerical simulation of reinforced concrete structures requires the inclusion of heterogeneities such as reinforcements and aggregates. Due to the different characteristic length scales of the components, a suitable multi-scale formulation was developed in the first project phase. Based on the acquired data and the conception of the computational approach, interval variables, fuzzy sets and fuzzy probability-based random variables were identified as suitable uncertainty models. The major research goals for the continuation are briefly outlined as follows. Since the realistic computation of mechanical behaviour of reinforced concrete is decisively influenced by the bond formulation between reinforcing components and matrix, an enhanced structural model of a representative volume element (RVE) will be developed, where the existing RVE is extended by the refined modelling of the characteristic topology of e.g. steel bars. Local damage effects are taken into account by the development of a traction-separation law and its implementation into corresponding interface elements. The consideration of uncertain parameters in this context allows a quantifiable assessment of bond qualities. Furthermore, the long-term behaviour of concrete structures with consideration of polymorphic uncertainty on material as well as on structural level, such as conversion of the structural purpose, will be investigated as well. The development of required uncertainty models and analysis methods regarding time dependency are proposed for the second project phase. The developed method for multiscale analysis involves the approximation of the effective mechanical quantities by recurrent neural networks. The development of novel and efficient training algorithms, as well as methods for model order reduction, enable the reduction of the numerical effort. In summary, the proposed developments of the second project phase will lead to a significantly expanded time-dependent analysis of concrete structures with polymorphic uncertain parameters and the clearly quantitative and qualitative extension in design tasks e.g. with regard to minimization of resources or robustness requirements.
DFG Programme
Priority Programmes
Ehemaliger Antragsteller
Professor Dr.-Ing. Wolfgang Graf, until 2/2020
