P_2.2.1 stoch_mod

Wood is an orthotropically describable material that is strongly diffuse, organic and optimised as far as its growth conditions are concerned. The present safety concept for supporting structures made from timber (EC5) is based, as far as the material is concerned, on the 5 % quantile which, especially with strongly diffuse material specific parameters and a multitude of influential parameters, can be extremely sensitively determined using statistics. Laying down firmness specific parameters therefore means assessments based on relations to measurable specific parameters. Additionally, specifying the support models of rod-like or plane-like structures (such as glulam, structural timber, and cross-laminated timber) requires the description of mechanical specific parameters of system-forming single components and their effects and interactions in system structures. So far, any model observations described in the literature have focused on a mechanical description of wood as a material free of growth characteristics (micro- or macro I levels), on a description of selected load situations, on the method of finite elements (FEM), or on a contrasting of experimentally determined technical results of different scale observations. Few publications so far have chosen the approach by which structures are described while the stochastics inherent in the material are taken account of (SFEM in the next project, 2.2.3 MMSM).

It is the objective of this project to compile the statistical characteristics of the relevant mechanical specific parameters, and to describe their interaction as well as their dependences by using multivariate statistics as well as further statistical procedures while looking at different scale dimensions. What is applicable here is the representative statistical distribution model and parameters of the relevant characteristic using the physical situation – these have to be established as a broad statement concerning the properties of the whole, and to examine their relations to other parameters. In doing so, both a large internal and external data pool can be drawn from, as well as some slightly widened experimental technology used in parallel projects. Based on this data, the idea is to study influences while considering classification and overall situation, and to process these using the available statistical models in stochastic finite elements.