Multiscale models for the mechanical behaviour of particleboard

CARINI, ANDREA;VILLA, TIZIANO

2015/2016

---

Abstract

Particleboard is a product composed of chopped wood and it is nowadays widely used in a lot of branches thanks to its low cost, versatility and with a low environmental impact due to the fact that is generally composed of recycled wood. Boards made of particleboard can be considered as an anisotropic and non-homogeneous sandwich panels so the mechanical properties research requires an in-depth study. The studies performed until now, do not consider a finite element model and were only focused on an experimental approach. The main goal of this thesis, instead, was a multiscale modelling of the Particleboard, building a FE model of a Representative Volume Element of the material, which lead to the evaluation of the elastic constants and to the comparison with the ones found using classical experimental methods. To obtain the input data required by the FE environment it was necessary to inspect the Particleboard with a series of different techniques. One of those was the analysis of the Micro-CT images, necessary to inspect the geometrical characteristics and orientations of the particles. These images were also useful for a thresholding algorithm that was developed in order to evaluate the Particleboard volume fractions. In addition, a backlit table was used for the automatic detection of the geometrical characteristics of the raw wood fibres. Moreover, the panel density profiles (obtained experimentally) was analysed for the evaluation of the wood fibres density. All these data were used as input of the FEM modelling environment (Digimat®) and lead to the execution of analysis that give back as results the Particleboard mechanical properties (in particular the elastic moduli), after an appropriate configuration of the software (for example the selection of the right Representative Volume Element generation algorithm). Finally, different models were built in order to compare different approaches and the relative results, considering also the ultimate comparison with the experimental methodology.