Model fitting for visco-elastic creep of Pinus radiata during kiln drying

Citation
Mn. Haque et al., Model fitting for visco-elastic creep of Pinus radiata during kiln drying, WOOD SCI TE, 34(5), 2000, pp. 447-457
Citations number
14
Categorie Soggetti
Plant Sciences","Material Science & Engineering
Journal title
WOOD SCIENCE AND TECHNOLOGY
ISSN journal
00437719 → ACNP
Volume
34
Issue
5
Year of publication
2000
Pages
447 - 457
Database
ISI
SICI code
0043-7719(200012)34:5<447:MFFVCO>2.0.ZU;2-#
Abstract
This work examines the applicability of mathematical models for correlating the visco-elastic strains during kiln drying of radiata pine (Pinus radiat a D. Don) sapwood at various temperatures and moisture contents. The eventu al aim is to use a mathematical model incorporating these strains to optimi se the drying schedules and minimise the degradation. Data sets from previo us experiments (Keep 1998) obtained at temperatures from 20 to 140 degreesC for sapwood at 5, 15 and 20% moisture contents (dry basis) were analysed. The data were fitted for various theoretical models, namely the Maxwell, Ke lvin and Burgers models, and the Bailey-Norton equation. The parameter valu es and standard errors for the above models over the range of experimental data have been determined. The results indicate that the Maxwell model did not fit the experimental data well, having only one parameter. In most case s, the Bailey-Norton equation was inadequate, as it is a power-law model wi th a predicted continuous increase in creep with time and does not predict a plateau in the creep strain, as has been observed for most of Keep's (199 8) data. The Kelvin model gave a better fit than the Bailey-Norton equation for most of the data sets. From visual inspection of the plots for the exp erimental data and the model predictions with time, it was found that both the Kelvin and Burgers models fitted the data satisfactorily. However, the three-parameter Burgers model was not a significant improvement over the tw o-parameter Kelvin model at the 0.01 level of significance, as shown by an analysis of variance.