Experimental study and numerical simulation of flow pattern and heat transfer during steam drying wood

Transcription

1 Experimental study and numerical simulation of flow pattern and heat transfer during steam drying wood Abstract J. Barański 1, M. A. Wierzbowski 2, J. A. Stasiek 3 The high cost of fossil fuel and soaring consumer interest have encouraged people in the wood industry to look for faster and more energy-efficient methods to dry lumber. In this paper results of experimental study and numerical simulation of flow pattern and heat transfer during steam wood drying are presented. Wood species, namely oak (Quercus L.) and pine (Pinus L.), were subject of steam drying process in a laboratory kiln especially arranged for that reason (Wierzbowski et al., 2008). Main focus of those tests was to shorten the time of drying process and afterward to check properties of wood. As results of mechanical properties checking are presented in separate paper, here authors focused on numerical predictions of uniform velocity and temperature profiles through the drying kiln, which is of great importance for drying and also for energy saving. Predicted velocities were used in the laboratory kiln for tests. Satisfactory results were obtained as the time of drying process was significantly reduced. The kiln is equipped with heat exchanger supplied by exhaust gases from furnace allowing spread water to evaporate on its surface. Generated steam, by circulation fan, is distributed between the wood stake. The dryer is for all timber species in terms of final moisture content to 6 % in the high temperature of up to 150 o C. Model of drying chamber was created consistent with existing experimental rig. The flow pattern around and through an array of in-line truncated boards of oak and pine have been simulated numerically. The Renormalization Group k-ε turbulence model and model for the near-wall treatments have been used for simulation. Vortex shedding from in-line boards separated by small gaps have also been numerically investigated. Small gaps between in-line boards have an insignificant effect on the mass transfer. 1 Introduction Reduction of energy consumption and drying processing time are currently two important objectives of timber industry, as drying is one of the most costly consuming steps in terms of energy and time. Extensive researches have been done and are still in progress to determine the optimal drying strategy to achieve the required timber quality at minimum cost. However, most of 1 Senior Research Fellow, jbaransk@pg.gda.pl Mechanical Engineering Faculty, Gdansk University of Technology, Poland 2 Senior Research Fellow, rwierzbo@pg.gda.pl Mechanical Engineering Faculty, Gdansk University of Technology, Poland 3 Professor, jstasiek@pg.gda.pl Mechanical Engineering Faculty, Gdansk University of Technology, Poland

2 experiments and modeling predictions focus on heat and mass transport within the boards, while, in practice, local drying conditions in the kiln strongly interact with heat and mass transport inside wood. Especially these are essential for High Temperature Drying (HTD) schedules with short drying times and uniform drying conditions (air temperature, humidity and velocity) in the kiln. The instrumentation and equipment in the kiln provides, together with a control system, environment, which is apparently different from the conventional drying. One important application of mass and heat transfer coefficients, over in-line boards stacked in an array, is analysis of the external convective transfer processes in drying of timber. Accurate evaluation of the external heat and mass transfer coefficients determines proper design of wood drying kilns and optimum operation of drying processes (Sun et al., 2000). Consequently, it is necessary to revisit this investigation of external transfer over in-line boards. From the other hand thanks to development of the wood drying techniques using saturated or superheated steam and gas-steam mixture flow, waiting time for wood material industry can be reduced and brings economic benefits, such as protection of wood against fungi and fracture, which extends its life. In this paper results of experimental study and numerical simulation of flow pattern and heat transfer during steam drying of wood is presented. Wood species, namely oak (Quercus L.) and pine (Pinus L.) were subject of steam drying process in a laboratory kiln specially designed for the purposes of the research. 2 Experimental background Drying in superheated steam is economically justified because of the shorter processing time and reduced energy consumption in comparison to drying in hot air. Evaporation of free water does not change wood shape and main dimensions during process of wood drying. With the loss of water evaporation zone moves deeper into the wood. The proper conduct of the drying process allows faster extraction of water (Gard et al., 2008). In the initial stage of drying hot water was supplied to the chamber to increase humidity and temperature throughout the material. Exhaust gases flow through the heat exchanger to raise the temperature of the mixture in the chamber. As far as humidity and temperature grow, we start with the drying process. The process of drying the material continues to achieve the assumed wood humidity of 10 % EMC (equilibrium moisture content). The next process was the conditioning of wood - slowly cooled chamber with getting hot water to remove the stress in the material which emerged during the whole process of drying. During drying process of great importance are: physical properties of drying agent, evaporation of water from both timber and free surface, hygroscopic properties of wood (depending on the species), hygroscopic equilibrium of wood, changes inside wood during evaporation. Wood species, namely oak (Quercus L.) and pine (Pinus L.), were subject of steam drying process in a laboratory kiln. The kiln is equipped with heat

3 exchanger supplied by exhaust gases from furnace. Water, spread from two nozzles, evaporates on exchanger s surface. Generated steam, is distributed between the wood stake by circulating fan. The dryer is dedicated for all timber species of final moisture content to 6 % in the high temperature of up to 150 o C. Detailed description of laboratory kiln was presented previously (Wierzbowski et al., 2009, Wierzbowski et al., 2008). Figure 1. View of the oak (Quercus L.) pile of boards inside the kiln. Figure 2. View of the pine (Pinus L.) stack inside the kiln. inlet temperature measurement temperature measurement outlet moisture and temperature measurement temperature measurement Figure 3. Model of the drying kiln with measurement points and air supply area.

4 Figure 4. Dimensions of stack of boards and location of probes for measuring temperature and moisture content during experiment. Figure 5. Dimensions of stack of timber and location of probes for measuring temperature and moisture content during experiment. 3 Numerical modelling For separated flows and recirculating flows around the blunt boards in a stack, the renormalization group (RNG) k-ε turbulence model (Yakhot et al., 1986) has been used to solve the turbulent momentum and species transport equations in a three-dimensional geometry. The model equations in their RNG form are similar to those for the standard k-ε model. The RNG k-ε model employs a differential form of the relation for the effective viscosity, yielding an accurate description of how the effective turbulent transport varies with the effective

5 Reynolds number. This allows accurate extension of the model to near-wall flows and low-reynolds-number or transitional flows, e.g., flows in an otherwise quiescent enclosure where the flow is turbulent in regions of limited extent, but is otherwise laminar (Fluent Inc., 2003). In addition, the RNG k-ε model can be used to analyze time-dependent flows with large-scale organized structures (e.g., vortex shedding, shear-layer instability). The drying kiln model is created consistent with existing experimental rig. The grids of boards are shown in Figures 6a and 6b. The space size between the board layers is the thickness of the stickers, which is 20 mm or 30 mm. The grid is tetrahedral in x-, y- and z-direction in the different regions of the air flow system. In the calculation, the SIMPLE algorithm of Patankar (Patankar, 1981) has been used together with the solver of Fluent/Uns (Fluent Inc., 2003) to solve the pressure-velocity coupling equations. The flow pattern around and through an array of in-line truncated boards of oak and pine have been simulated. The Renormalization Group k-ε turbulence model, as mentioned before and model for the near-wall treatments have been used for simulation. In addition to heat transfer analysis the P1 and Discrete Ordinates radiation model were used. Vortexes shedding from in-line boards separated by small gaps have also been investigated, as they have an insignificant effect on the mass transfer. The grid of drying chamber contained 1,824,895 unstructured elements for 30 x 150 mm thin boards and kiln (Figure 6a) and 1,316,655 unstructured elements for 70 x 70 mm lumber and kiln (Figure 6b). a) b) Figure 6. View of the unstructured grid: a) boards inside kiln, b) lumber inside kiln. 4 Boundary conditions and fluid properties The working section of the kiln is 1.55 m long and it has a cross-section of 0.60 m high and 0.85 m wide. The layers of the boards and timber, stacked in the working section of the kiln were shown in Figures 4 and 5. The distance between boards is 20 mm and between timbers is 30 mm. The air velocity between the board layers was 2 to 7 m/s, and the temperature was 100 o C ±2. The non-slip wall condition is applied to the walls of the boards. Uniform conditions are applied at the inlet boundary. These are constant: mixture of gassteam velocities 2.0, 2.5, 3.0, 3.5 and 5.0 m/s and temperatures 80, 90 and 100 o C.

6 The flow exit is treated to be a constant pressure (zero gauge pressure) outlet boundary. All walls of the kiln have got the same temperature. Because of no heat exchange assumption between kiln and surroundings, heat losses are set to be equal to zero. Numerical predictions of uniform velocity and temperature profiles through the drying kiln were used in the laboratory kiln for tests. 5 Results Experiments were carried out with Pomeranian region lumber of oak and pine. Probes to measure moisture content inside wood are placed in the material so that it was possible to measure moisture content in a number of characteristic points of the kiln, i.e. in the middle of the boards or in the outer layers of the stack Figures 3, 4 and 5. Following figures presents the results of numerical and experimental works for oak and pine lumbers. Figure 7 shows temperature distribution inside the kiln. Fig. 7a presents temperature values along the kiln. It is quite uniform and difference between maximum and minimum value is about 3K. Similar situation is on Fig. 7b, which presents temperature values in cross-section of the kiln. Here, temperature differences are greater, about 12 K. During experiment this difference is minimised by influence of reversible fan. a) 372 Temperature 371 Temperature [K] b) 366-0,8-0,6-0,4-0,2 0,0 0,2 0,4 0,6 0,8 Chamber's Length [m] Temperature [K] Temperature -0,5-0,4-0,3-0,2-0,1 0,0 0,1 0,2 0,3 0,4 0,5 Chamber's Width [m] Figure 7. The results of numerical modelling: a) temperature along kiln, b) temperature in cross section of kiln.

7 Figure 8 presents the results of numerical modelling of pine boards drying. Calculated values of velocity inside kiln (Fig. 8b) led to uniform temperature inside the kiln (Fig. 8c, 8d). Between boards velocity achieve values of about m/s. The highest value of velocity is at the outlet of the kiln. Predicted drying parameters (velocity and temperature) were used during experimental works. a) b) c) d) Figure 8. Results of pine numerical simulations: a) velocity vectors of gassteam flow in vertical cross section, b) velocity vectors of gas-steam flow in horizontal cross section, c) temperature field of gas-steam flow in vertical cross section, d) temperature field gas-steam flow inside kiln. Figure 9 presents the results of numerical modelling of oak lumber drying. Calculated values of velocity (Fig. 9a) inside the kiln led to uniformity of temperature inside the kiln (Fig. 9b and 9c). In layers between boards, velocity achieved values of about m/s. The highest values of velocity are at the outlet of the kiln.

8 a) b) c) d) Figure 9. Results of oak numerical simulations: a) velocity vectors of gassteam flow in horizontal cross section, b) temperature field of steam flow in horizontal cross section, c) temperature field of steam flow inside kiln d) pathlines of gas-steam flow inside kiln. Hot water was supplied to the chamber to increase humidity and temperature throughout the material in the initial stage of drying. Exhaust gases flow through the heat exchanger to raise the temperature in the chamber. As far as humidity and temperature grows, we start with the drying process. The process of drying the material continues to achieve the assumed wood humidity of 10 % EMC (equilibrium moisture content). The next process was the conditioning of wood - slowly cooled chamber with getting hot water to remove the stress in the material which emerged during the whole process of drying. Results of moisture content during oak lumber drying process are presented on Figure 10. In this case overall time was extended due to achieve proper level of moisture inside wood as hot water was directed on a part of pile.

9 22,5 20,0 Moisture content Moisture content [%] 17,5 15,0 12,5 10,0 7,5 5,0 2,5 0, Time [h] Figure 10. The results of 7.0 cm x 7.0 cm oak lumber (Quercus L.) drying process using gas-steam mixture (time dependence of moisture content). Figure 11 presents a view of dried oak lumbers. Because of high temperature and long time of drying, structure and colour of wood were changed. Fractures of wood after this kind of drying process can be also observed. Figure 11. The results of 7.0 cm x 7.0 cm oak lumber (Quercus L.) drying process using gas-steam mixture (fractures and colour changes of wood). Thus, there was need to change the position of hot water supply. Finally, water nozzles were placed near the heat exchanger, which allows water to evaporate directly to steam during kiln s operation. In new arrangement of water nozzles the drying time was remarkably shortened (Fig. 12, 13, 14 and 15).

10 22,5 20,0 Moisture content Moisture content [%] 17,5 15,0 12,5 10,0 7,5 5,0 2,5 0, Time [min] Figure 12. The results of 7.0 cm x 7.0 cm pine lumber (Pinus L.) drying process using steam-gas mixture (time dependence of moisture content). Figure 12 presents the results of experimental work of moisture content changes for pine lumber drying. The temperature of drying agent was about 100 o C and inlet velocity during heating stage was about 4.5 m/s. During drying and conditioning process, inlet velocity was reduced by control system to about 2.5 m/s. This was necessary to achieve low velocity between wood layers to avoid fractures of wood. Overall process took about 2.5 days. In Figure 13 photo of dried pine lumber is presented. Slight colour changes and no fractures were reported. Figure 13. The results of 7.0 cm x 7.0 cm pine lumber (Pinus L.) drying process using gas-steam mixture (fractures and colour changes of wood).

11 Similar conditions were obtained during drying process of pine boards. Temperature was also about 100 o C and inlet velocity was about 4.5 m/s. Figure 14 shows results of moisture content changing of pine boards drying process. This drying process took about 3 days. 22,5 20,0 Moisture content Moisture content [%] 17,5 15,0 12,5 10,0 7,5 5,0 2,5 0, Time [min] Figure 14. The results of 3.0 cm x 15.0 cm pine boards (Pinus L.) drying process using gas-steam mixture (time dependence of moisture content). In Figure 15 photo of dried pine boards is presented. Slight colour changes and no fractures were also reported. Figure 15. The results of 3.0 cm x 15.0 cm pine boards (Pinus L.) drying process using gas-steam mixture (fractures and colour changes of wood).

12 6 Conclusions The results obtained from tests shows that drying time shortens of about [%] what justifies further experiments. With the time shortening to 2,5-3 days, it is assumed that also energy consumption for drying process of soft wood, such as pine, will decrease. Next steps are planned with the use of coniferous and leafy lumber. 7 Acknowledgment The financial assistance of Ministry of Science and Higher Education, Poland, Grant N /3058 is kindly acknowledged. References Gard W.F., Riepen M.: Super-heated drying in Dutch operations. Conference COST E53, Delft, The Netherlands, October Fluent Incorporated (2003). Fluent Incorporated, Centerra Resource Park, Lebanon, NH Langrish, T. A. G., Kho, P. C. S., & Keey, R. B.: Experimental measurements and numerical simulation of local mass-transfer coefficients in timber kilns. Drying Technology, 10, , Langrish, T. A. G., Keey, R. B., Kho, P. C. S., & Walker, J. C. F.: Timedependent flow in arrays of timber boards: Flow visualization, mass-transfer measurements and numerical simulation. Chemical Engineering Science, 48 (12), , Pang S., Simpson I.G., Haslett A.N.: Cooling and steam conditioning after hightemperature drying of Pinus radiata board: experimental investigation and mathematical modeling. Wood Science and Technology 35 (2001), Springer Verlag 2001, s Patankar, S. V.: Numerical heat transfer and fluid flow. New York: McGraw-Hill Book Company, Sun, Z. F., Carrington, C. G., & Bannister, P.: Dynamic modeling of the wood stack in a wood drying kiln. Chemica Engineering Research and Design, Transactions Institution of Chemical Engineers, Part A, 78, , Syrjanen T., Oy Kestopuu: Heat treatment of wood in Finland-state of the art Wierzbowski M., Barański J., Stąsiek J.: Gas-steam mixture wood dryling. COST E53 Meeting ''Quality Control for Wood and Wood Products'' : EDG Drying Seminar ''Improvement of Wood Drying Quality by Conventional and Advanced Drying Techniques'', Bled, Slovenia, April 21st-23rd, 2009.

13 Wierzbowski M., Barański J., Stąsiek J.: Suszenie drewna mieszaniną parowogazową. Termodynamika w nauce i gospodarce. Wrocław, 2008 (in polish). Yakhot, V., & Orszag, S. A.: Renormalization group analysis of turbulence. Basic theory. Journal of Scientific Computing, 1, 1-51, 1986.