Derome, D.1*, Zillig, W. 2 and Carmeliet, J.3

1Empa, Wood Laboratory
"Uberlandstrasse 129, CH8600 D"ubendorf, Switzerland
dominique.derome@empa.ch

2 Institute for Building Physics, Holzkirchen, Germany
wolfgang.zillig@web.de

3Chair of Building Physics, ETH Z"urich, HIL E 46.3 Wolfgang-Pauli-Strasse 15 CH 8093 Z"urich
Empa, Head Lab. for Building technologies, "Uberlandstrasse 129, CH8600 D"ubendorf, Switzerland
jan.carmeliet@empa.ch
Key words: moisture transport, wood, cellular structure, location dependent properties, multi-scale modeling

ABSTRACT

Moisture transport in wood is complex due to its anisotropic material structure, resulting from the cells anatomy, growth rings where early- and latewood alternate and the presence of vessels, rays and pits. As a result, the moisture transport properties of wooden components are highly dependent on the direction (longitudinal, radial or tangential). Moreover, due to variations in cell geometry, and growth ring structure, the moisture transport properties, such as moisture capacity and vapor diffusion coefficient depend on location. The purpose of this paper is to present a methodology to determine location dependent moisture transport properties and to study their influence on the global hygroscopic behaviour of wood.

In order to derive location dependent moisture transport properties, 2D micro structural images are obtained by Scanning Electron Microscopy of a full growth ring. The image is then thresholded to obtain a binary representation of the cell wall - cell lumen structure. The geometry of the cells is extracted and the cell borders are represented by lines. Based on the line configuration, a triangular mesh is generated. Finite element analyses are performed using a moving window methodology to obtain the position dependent material properties. A vapour pressure difference is imposed at two opposite sides, while the other sides remain vapor tight and the water vapor flux and water vapor diffusion coefficient are determined. These calculations are repeated for different relative humidities. The moisture capacity is determined based on the wall volume in the window. Different sizes of moving window are examined to find a suitable size of the representative elementary volume (area in 2D). In a second step, the location dependent moisture transport properties are used to analyse the water vapour transport at meso- and macroscale. The macroscopic moisture transport properties are then compared with measurements on wood. The water vapour permeability of the wall is determined comparing upscaled and measured water vapour permeability values. Further validation is performed by a 5-days adsorption/desorption test where the moisture distribution in the wood is measured by high-resolution X-ray projection at regular intervals.

The method allows to determine location-dependent density and water vapour permeability. These location-dependent material properties can be then used to model the influence of growth ring structure on the total moisture transport in the wood ignoring the actual geometry of single cells in the growth ring.