Modelling color changes in wood during conventional drying

Authors: Liu, Wenhua
Advisor: Fortin, YvesStevanovic-Janezic, TatjanaTremblay, Carl
Abstract: Wood discoloration during drying is diminishing quality and value of end product, increasing production costs and decreasing yield. Wood discoloration during drying is a complex process which is difficult to predict. The development of a wood color model can save material and time by simulating color changes for any specific drying conditions. A set of six experiments were performed in this study on paper birch and sugar maple sapwood to measure wood color changes during conventional drying at three different levels of dry-bulb temperature (40, 60 and 80˚C) and two levels of wet-bulb depression (4 and 15˚C). Statistical wood color models for lightness L* were proposed for each species, both for the wood surface and through the board thickness, to predict the wood color changes during conventional drying using mixed regression models with the board sample taken as the random effect. Three temperature (3T) (40, 60 and 80˚C) and two temperature (2T) (60 and 80˚C) models were developed. Finally, the statistical wood color models were integrated into an existing heat and mass transfer numerical model (DRYTEK) in order to simulate, for any conventional dynamic wood drying schedule, wood color changes on the surface and through the board thickness. The numerical model parameters (moisture content-water potential relationship, effective water conductivity, convective mass transfer coefficient) were experimentally determined for paper birch and sugar maple at the three drying temperatures. Both types of wood color predictive models were then validated by means of an independent drying run conducted at the dry-bulb temperature of 70°C and the wet-bulb depression of 10oC. The results of the wood color measurement tests show that at the dry-bulb temperatures of 60 and 80˚C, the L* values of both species below the surface decrease rapidly with a decrease of the moisture content (M) from the green state to the fiber saturation point. Then, the L* values decrease slowly until about 15 - 25% M where they may even start to increase. Wood color changes at 40˚C were found very small, either positive or negative. The L* values at the surface also decrease as the moisture content decreases and, except for the temperature of 40oC, the changes in color increase with the drying temperature. In general, the higher is the dry-bulb temperature, the greater is the decrease of the L* values through the board thickness. Conversely, the higher is the wet-bulb depression at a given dry-bulb temperature, the smaller are the color changes. The color components a* and b* follow a similar behavior as L* with respect to drying temperature and wet bulb depression. However, contrarily to the L* values, the a* and b* values increase with a decrease of M. The comparison of the predicted L* values obtained from the statistical models with the experimental L* values obtained from the validation tests shows a very good agreement between both types of results in the case of sugar maple. For paper birch, a fairly large discrepancy is observed during the first part of drying between predicted and experimental results but a better agreement is found at the end of drying. The results of the moisture content-water potential relationship determination show that the water potential increases with temperature at a given moisture content. A characteristic plateau was found in the water potential range of -2,000 and -6,000 J kg-1. As expected, the effective water conductivity increases with moisture content and temperature and it is higher in the radial direction than in the tangential direction. The convective mass transfer coefficient increases with dry-bulb temperature at a given wet-bulb depression, whereas it decreases with an increase of web-bulb depression at a given dry-bulb temperature. Finally, the comparison of the wood color measurements during the validation tests on paper birch and sugar maple with the wood color values simulated with the integrated statistical/numerical models shows a very good agreement between both types of results in the case of sugar maple. As it was observed for the statistical models, the fit was poorer in the case of paper birch, especially above the fiber saturation point where the initial moisture content seems to be an important factor in the color changes behavior of wood during drying.
Document Type: Thèse de doctorat
Issue Date: 2011
Open Access Date: 18 April 2018
Grantor: Université Laval
Collection:Thèses et mémoires

Files in this item:
Description SizeFormat 
28403.pdfTexte5.8 MBAdobe PDFThumbnail
All documents in CorpusUL are protected by Copyright Act of Canada.