Variability in Transport Properties for Blackbutt Timber in New South Wales Within and Between-Tree Variability.код для вставкиСкачать
Dev. Chem. Eng. Mineral Process. 14(1/2), pp. 173-181, 2006. Variability in Transport Properties for Blackbutt Timber in New South Wales: Within and Between-Tree Variability S.J. Cabardo" and T.A.G. Langrish Department of Chemical Engineering, University of Sydney, Sydney, New South Wales 2006, Australia Variability is a key issue in the processing of biological materials, in this case the hying of hardwood timber. This paper reports the measurements of variability of transport properties. which are relevant to the drying of blackbutt, Eucalyptus pilularis Sm, from northern New South Wales. Specifically, within-tree and betweentree variations are reported for two blackbutt regrowth logs. An analysis of variance showed that some timber properties were affected by the board positions within-trees and between-trees. Circumferential and radial efects were significant for the withintree variability of most transport properties. Similarly, radial and circumferential effects were signlficant for most of the transport parameters between trees, but can be tentatively stated because only two regrowth logs were assessed. Timber boards with high initial moisture contents had higher rates of diffirsion and low basic densities using principal components analysis. A possible reason is that if there is less wood material per unit volume, these vacant spaces may be occupied by water, and there is also less resistance for diffusive transport of moisture. Introduction Large quantities of timber from hardwood plantations in Australia are now being harvested, from 1 million cubic metres in 2000-01 to a predicted level of 9.2 million cubic metres in the latter half of the decade as reported by Fer uson et al. [l]. In NSW, the consumption of eucalypts sawn woods was 263,000 m in 2000-01 . In the future, Australian timber drying companies may have to rely on getting most of their timber supplies from these hardwood plantations. However, the companies report a growing difficulty in handling this type of timber for a number of reasons, with the increased amount of variation in timber properties affecting its dried quality being significant. Anecdotally, plantation timber appears to be more variable with K. ~~ * Author for correspondence (firstname.lastname@example.org). I73 S.J. Cabardo and T.A.G.Langrish regards to its properties compared with old growth timber. As a result, it is possible that large variation in intrinsic properties may require better control, hence optimized drying schedules that account for these variabilities. It is probable, on the basis of previous studies in the literature about the drying of hardwoods , that the main transport mechanism of moisture movement within hardwoods is diffusion. This is due to hardwoods consisting mainly of heartwood. Cell wall pits are aspirated within heartwood, thus, water cannot flow by convection through these cells. The diffusion model for hardwoods was also confirmed by the work of Wu (41 because it predicted the same moisture content distributions, i.e. moisture content as a function of distance within the timber, as those observed for Tasmanian eucalypts, at various times during drying. The tree species chosen for this study was blackbutt (E.piZularis Sm.), because blackbutt is the predominant planted hardwood species in NSW, Australia [S]. It is most abundant in many coastal areas from south of Bega up to south-eastem Queensland. In summary, the variation o f the diffusion coefficient (a transport property), in addition to basic density and green moisture content for blackbutt, was investigated in this paper. Experimental Procedure and Apparatus I Preparation of Samples Blackbutt boards (28 mm thick x 108 mm wide x 900 mm long) were cut from two regrowth logs; log 685 that was taken from the Newry Creek region, and log 686 from the Lower Bucca area. These boards were taken from different locations within each log, as shown in Figures 1 and 2. Each ‘ A ’ board (the first 900 rnm from the bottom end) has its corresponding ‘E’ board that was taken from the top end of each log (approximately 3.6 metres up the log). A 20 mm thick sample was taken from each end of each green ‘A’ and ‘E’ board to calculate its corresponding initial moisture content using the oven-dry method. The remaining board length was kiln dried in a drying tunnel, using the conventional drying schedule published in the Australian Timber Seasoning Manual [ 6 ] suitable for mixed sawn blackbutt boards (25 mm thick) as shown in Table 1. Each kiln charge consisted of six boards, resulting in four kiln charges (four experiments). The boards were left to dry, with the schedule changed based on the number of days. Each experiment took 21 days to run. The final moisture contents were calculated by cutting another 20 mm sample from each board at the end of the drying process, and repeating the oven-dry method. Tirnber density was based on the oven-dry weighdgreen volume, where the sample dimensions were measured for green volume. Moisture content change points (%) Dry bulb temperature PC) Wet bulb depression PC) I74 Green 60 40 35 30 25 55 3 55 3 60 4 60 4 65 5 65 8 15tofinal (usually 12) 70 70 10 15 20 Variability in Transport Properties for Blackbutt Timber in New South Wales Figure 1. Cross-section of the bottom end, showing where each ' A ' board was taken within log 685. Figure 2. Cross-section of the bottom end, showing where each 'A' board was taken within log 686. II Drying Tunnel Figure 3 shows a schematic of the overall design of the pilot-scale batch dryer used for drying 800 mm timber boards. This conventional kiln was used to produce controlled drying conditions, with initial dry-bulb and wet-bulb temperatures of 55°C and 52"C, respectively, and an air velocity of 1.30 m/s through and around the boards. Manipulating the steam flowrate to a steam-injection system enabled control of the wet-bulb temperature to its desired set-point. The steam system consisted of a dryer and six-point steam injection pipe over a 300 mm duct. The control mechanism for the dry-bulb temperature involved the flowrate adjustment, using a control valve, of 100 kPa (gauge) steam to a finned heat exchanger. The total mass of the timber stack was measured using a platform balance. It has a capacity of 200 kg and a resolution of 5 g. This information was transmitted through an RS 232 interface to a computer to estimate the average moisture content of the timber stack. This value was used to determine if the set-points of the dry-bulb and wet-bulb temperatures had to be altered according to the drying schedule shown in Table 1. I 75 S.J. Cabardo and T.A. G. Langrish W a l a r Mains Pump Figure 3. Schematic diagram of timber drying tunnel. 111 Fining Procedure for Diffusion Coefficients The predicted overall drying curve is dependent on the diffusion coefficients and operating temperatures for each experiment (Langrish et al. ). The model is based on solving Fick’s second law of diffusion for mass transfer and the conduction equation for heat transfer. Moisture is assumed to diffuse through the timber and evaporate near the surface. It is often difficult to measure the diffusion coefficient directly, but it can be fitted to experimental data, as here. Schaffner  and Wu  both successfully fitted diffusion coefficients to observed data for eucalypt timbers. Therefore, least-squares parameter fitting was used to adjust the values of the reference diffusion coefficient (Or) and the activation energy (DE) (with T as the average temperature of the timber boards) in an Arrhenius-type equation: D = D, ex& D,/T) ...(1) The predicted parameters, D, and DE,are then used to calculate the diffusion coefficient, D. 176 Variability in Transport Properties for Blackbutt Timber in New South Wales The parameter fitting minimizes the sum of squares of the difference between the predicted and actual average moisture contents for each board. Typical initial values of D,and DEwere 1.13 x los5m2s-' and 3750 K, respectively. The recorded moisture contents that were measured throughout the drying process, and the dry and wet bulb temperatures, were dependent on each board and each experiment. The fitted initial moisture content was set equal to the actual initial moisture content. Fick's second law of diffusion for mass transfer and the conduction equation for heat transfer were then solved simultaneously, with convective boundary conditions at the board surfaces and symmetry at the centerline of each board. Overall, the final adjusted values of D, and DEcharacterized the drymg behaviour of the corresponding timber board. Results and Discussion I Overview The diffusion parameters are listed alongside the initial and final moisture content and basic density for each board in Tables 2 to 5. The standard error is a measure of the difference between the predicted drying rate and the experimental drying rate. All the resulting standard error values were small, and the fitted curves followed the experimental values closely. Since D was dependent on temperature, the average temperature value, T, used was 6OoC, because it reflects the average temperature for each board during the drying process. The parameters D, and DE should be characteristic of the timber and should be applicable to boards of different thickness if diffusion is the dominant moisture transport mechanism for moisture movement. The resulting values of activation energies for all boards were expected not to vary significantly because the activation energy should be a property of the major constituents of the timber matrix. Our values ranged from 3405 K to 3792 K, close to the values of 3800 K and 3773 K reported by Wu [ 4 ] and Langrish et al. , respectively. Table 2. Initial and final moisture contents, basic densities, and fitted diffusion parameters for experiment one. 177 S.J . Cabardo and T.A . G. Langrish Table 3. Initial and final moisture contents, basic densities, and fitted diffusion parameters for experiment two. I I 1 I 1 I Table 4. Initial and final moisture contents, basic densities, and fitted diffusion parameters for experiment three. Table 5. Initial and final moisture contents, basic densities, and fitted d@usion parameters for experiment four. Board ID ~ 5A 5E 12A 12 E 18A 18 E I 78 D, K] [x 1@" rn's-'] Basic density fkg/m3] 3.56 3.64 3.73 3.73 3.68 3.71 2.65 2.06 1.56 1.55 1.79 1.65 613 824 816 822 7 84 819 I@' DE [x l d m2s-'] 1.14 1.13 1.12 1.12 1.11 1.12 [x D Initial moisture content fkghg] 0.94 0.72 0.57 0.52 0.79 0.61 , [kg/kg] 0.05 0.06 0.06 0.07 0.035 0.028 0.022 0.018 0.029 0.022 Variabiliy in Transport Properties for Blackbutt Timber in New South Wales II Between-trees and Within-tree variations Based on the data in Tables 2 to 5 , it can be seen that the initial moisture contents decreased from pith to bark and basic density increased from pith to bark. The small dispersion in final moisture contents at the end of the drying process is a reflection of the lack of variability in equilibrium behaviour, because the drying behaviour at the end of drying is dominated by equilibrium. The conventional kiln schedule used for these drying experiments minimized the dispersion of the final moisture contents, thus satisfying the allowed amount of variation for final moisture contents associated with high quality timber according to Australian Standards . Moreover, an analysis of variance (ANOVA) showed that radial and circumferential effects were significant sources of the within-tree variation of the diffision coefficients, initial moisture contents and basic densities. A similar result was found for the variability of the same properties between-trees. However, the ANOVA for between-trees can only be tentatively stated because only two regrowth logs were dried, resulting in only one degree of freedom. Nevertheless, these results support the previous statement that the behaviour of these three parameters changes across the radius of the log, i.e. from pith to bark. The observation that the variability of these parameters is caused by the same effects suggests that there may be a correlation between all three parameters. In addition, the within-tree variability of the initial moisture content was affected by the height, the radial x circumferential interaction, and the height x circumferential interaction. III Principal Components Analysis Boards from experiments one to four showed an apparent correlation between high initial moisture contents, higher diffusion coefficients, and low basic densities. Timber boards with high densities indicates that there is more wood material per unit volume. This may either be due to large cell volumes and/or more cell-wall material, leaving less space for the water to occupy, which explains the low initial moisture contents for timber boards with high densities [3, 101. In addition, more wood material means a higher resistance to diffusive transport of moisture . Therefore the diffusion coefficient is expected to be lower in high density wood. A principal component analysis (PCA) [ l I] was performed on the three parameters. If these parameters are closely correlated, then the PCA should result in only one large principal component, and one large eigenvalue and eigenvector. The results of the PCA showed that the first and largest eigenvalue accounted for 92% of the total amount of variation within these parameters. Figure 4 shows that boards with high initial moisture contents have low basic densities, and thus high diffusion coefficients. One board (24E) from log 685 was considered an outlier; although it had a high basic density, it also had a high diffusion rate which was possibly due to the crack present throughout the length of the board. However, the presence of cracks has not increased the apparent diffusion coefficients by an order of magnitude. I 79 S.J. Cabardo and T.A. G. Langrish I \- 24E \ Basic density 1 (1000s of kgiin3) \ liiitiai inoisture content (kglkg, dry basis) Figure 4. Three-dimensional plot of the parameters, together with the line of maximum variation. Conclusions Within a tree, initial moisture contents decreased from pith to bark,, and basic densities increased from pith to bark. ANOVA analysis showed that circumferential and radial effects were likely to be significant for the variability of the diffusion coefficient, initial moisture content and the basic density, both within a tree and between trees. However, results from the ANOVA of between-trees c.an only be tentatively stated because only two regrowth logs were dried, whereas a thorough analysis was conducted for within-tree effects. In addition, timber boards with high initial moisture contents had low basic densities and high diffusion coefficients. This observation may be connected with the link between less wood material, more space for water in the green timber, and the increased ease with which moisture can diffuse through the timber board. 180 Variability in Transport Properties for Blackbutt Timber in New South Wales Acknowledgements This work is supported by the Australian Research Council, Forests NSW - a special mention to Dr Ross Dickson and Mr Bill Joe, and J. Notaras and Sons, Grafton. Nomenclature D D, DE T Diffusion coefficient, m2s-' Reference diffusion coefficient, m2 s.' Activation energy for diffusion, K Average temperature of the timber boards, K References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10 II Ferguson, I.S., Fox, J., Baker, T., Stackpole, D., and Wild, 1. 2002. National and Regional Plantation Wood Availability 2001-2044. Consultmi S Reporf for National Forest Inventov, Burenu of Rural Sciences, Canberra. ABARE. 2001, Australian Forest and Wood Products Statistics, March and June quarters 2001. ABARE, Canberra, p.13. Keey, R.B., Langnsh, T.A.G., and Walker, J.C.F. 2000. Kiln Drying of Lumber, pp.65-115, 175-181. Springer Verlag, Berlin. Wu, Q. 1989. An Investigation of Some Problems in Drying of Tasmanian Eucalypt Timbers. M. Eng. Sci. Thesis, University of Tasmania, Hobart, pp.140-141. Boland, D.J., Brooke;, M.I.H., Chippendale, G.M., Hall, N., Hyland, B.P.M., Johnston, R.D., Kieinig, D.A., and Turner, J.D. 1989. Forest Trees of Australia, CSIRO Melbourne. hitp://data.brs.gov.nw'mapsew/pfnnt/species.php?speciesid=14,accessed 1112412003. Mills, R. 1991. Australian Timber Seasoning Manual. Australian Furniiure Research and Development Insiiiute Limited. pp.160-I 66. Schaffner, R.D. 1981. Fundamental Aspects of Timber Seasoning, M. Eng. Sci. Thesis, Faculty of Engineering Science, Faculty of Engineering, University of Tasmania, Australia, pp. 130-135. Langrish, T.A.G., Brooke, AS., Davis, C.L., Musch, A,, and Barton, G.W. 1997. An Improved Drying Schedule for Australian Ironbark Timber: Optimisation and Experimental Validation. Drying Technology, 15( 1 ), 47-70. AustralianMew Zealand Standard ASMZS 4787:2001: Timber - Assessment of Drying Quality, Standards Australia, Standards New Zealand, accessed 12/3/2003. Walker, J.C.F. 1993. Primary Wood Processing. Chapman and Hall Publishers, London, pp.68-74. Smith, L.I. 2002. A Tutorial on Principal Components Analysis. hiip://www.cs.oingo.nc.nz/cosc4S3~student~tu1orials/principal~component~.pd~ accessed 19/3/2004.