Simulation of timber structures under consideration of material and structural inhomogeneities

• Applicant: Professor Dr.-Ing. Michael Kaliske
• Subject Area: Applied Mechanics, Statics and Dynamics
• Term: from 2011 to 2015
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 195530091

The final aim of the applicant in the area of timber mechanics is a realistic numerical analysis of timber structures, by means of FEM. In previous projects, models for the detailed description of the mechanical behaviour of timber have been developed. These models are related to homogenous timber without discontinuities and require deterministic material properties. The naturally grown material wood is affected by structural (branches and varying grain course) and material inhomogeneities (naturally varying material properties) which influence the mechanical behaviour of structural members made of timber. In the first project period, models to consider growth inhomogeneities in terms of branches and methods to capture the natural variation of material properties by a stochastic data model have been developed and will be developed until the preliminary project end.Branches as natural weak points in wood often initiate failure. To enable a realistic analysis of timber structures the existing methods to model failure and damage in timber should be enhanced and combined with the models to simulate growth inhomogeneities. Under compression, failure in timber can be described very well by plasticity formulations. Therefore, plasticity models will be combined with the methods to describe branches in the first part of the continuation project. Tensile and shear failure can be simulated more adequately by fracture mechanics and cohesive zone models compared to plasticity. Thus, in the second part of the continuation project, the existing cohesive elements and the related material description will be enhanced. To capture longitudinal tensile failure initiated by branches, the cohesive material model has to be extended. For the modelling of crack paths starting from branches, an adaptive meshing procedure for the cohesive elements needs to be developed and implemented. In the third period of the continuation project, the methods to describe material inhomogeneities will be enhanced. Although it is generally agreed that the variation of material properties can be mapped appropriate by stochastic distributions, the information about the parameters of these distributions are varying considerably. This additional uncertainty cannot be captured by randomness adequately. To use the advantages of a stochastic data model and be able to consider non-stochastic uncertainty, the development of a fuzzy-stochastic FEM for the simulation of timber structures is intended.

• DFG Programme: Research Grants