Available online at www.sciencedirect.com
APPLIED ENERGY Applied Energy 85 (2008) 878–895 www.elsevier.com/locate/apenergy
Energy efficiency enhancement of natural rubber smoking process by flow improvement using a CFD technique Perapong Tekasakul a,*, Machimontorn Promtong b a
Energy Technology Research Center and Department of Mechanical Engineering, Faculty of Engineering, Prince of Songkla University, Hat Yai, Songkhla 90112, Thailand b Division of Mechanical Engineering, School of Engineering and Resources, Walailak University, Nakhon Si Thammarat 80160, Thailand Received 9 September 2007; received in revised form 2 February 2008; accepted 4 February 2008 Available online 18 April 2008
Abstract A non-uniform flow and large temperature variation in a natural rubber smoking-room cause an inefficient use of energy. Flow uniformity and temperature variation can be improved by using a computational fluid dynamics (CFD) simulation. The effects of the size, position and number of gas supply ducts and ventilating lids which were at the inlets and the outlets of the smoking-room were investigated. The optimal rubber smoking-room of size 2.6 m 6.2 m 3.6 m contains 154 50 mm-diameter hot gas supply ducts, and four 0.25 0.25 m and four 0.25 0.20 m ventilating lids. The velocity distribution of this model in the rubber-hanging area was rather uniform. The average monitoring temperature of 54 positions was 62.1 °C. This model could reduce the temperature variation by a factor of three from the original room model, i.e., from 15 to 5.5 °C. In a further study, the heat input of an appropriate room model was finely adjusted to obtain a suitable temperature (60 °C) for the smoking process. It was found that an appropriate heat supply at this temperature is 11 kW. At this rate, the temperature variation is 5.3 °C. This improved model should help the rubber smoking cooperatives to achieve at least a 31.25% saving in energy. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: CFD; Rubber smoking; Velocity; Temperature field
1. Introduction Ribbed smoked sheet (RSS) production in about 700 rubber smoking cooperatives at the community level in Thailand is currently experiencing a serious threat from the high rise in fuel cost. The main problem in the rubber smoking process results from low efficiency in fuel utilization. This is caused by the poor design of the smoking-room [1]. The smoking-room used for producing RSS gives large differences of temperature and velocity; these result in non-uniform drying of the rubber sheets. The variations also affect the smoking-time used (which currently takes as long as 4–6 days). This directly influences the amount of fuel used. To improve *
Corresponding author. Tel.: +66 74 287035; fax: +66 74 212893. E-mail address:
[email protected] (P. Tekasakul).
0306-2619/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.apenergy.2008.02.004
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
879
the situation in RSS production of the rubber smoking cooperatives, fuel utilization, RSS quality and smoking period must all be taken into account. These factors can be optimized by improving the flow characteristics in the smoking-room [1–5]. Reduction of temperature and velocity variations will enhance energy efficiency. In the preceding work [1], investigations revealed that problems arose from improper hot gas introduction and exhaust of the present smoking-room. The CFD technique used was validated using experimental results obtained in a rubber smoking-room. The simulation results were in good agreement with the experimental values. In the modeling of the smoking-room, large velocity and temperature variations throughout the room were discovered. The results indicated that the number, size and position of the inlets and the outlets of the smoking-room were not ideal. These parameters are known to significantly affect the circulation of hot gas in the smoking-room. In this study, CFD simulation is again used to improve the flow in the rubber smoking-room for the purpose of energy efficiency enhancement. A suitable configuration of rubber smoking-room and elements which improves flow and temperature distribution is investigated. The effects of size, position and number of supply gas ducts and ventilating lids (which are at the inlets and outlets of the smoking-room) are studied. It should be noted that the original size of the natural rubber smoking-room model 1994 is used for this improvement as its capacity is sufficient for the production at the community level. Moreover, improvement from the basis of an existing infrastructure will save energy costs for the rubber cooperatives. 2. Numerical methods 2.1. Governing equations The equations of mass, momentum and energy conservation are normally used to describe fluid flow–heat transfer characteristics. These equations are solved by appropriate boundary and initial conditions. For natural convection, the governing equations of the flows are given by [1] Continuity equation: o ðq ui Þ ¼ 0 oxi
ð1Þ
Momentum equations:
o o p o oui ouj ð qu j u i Þ ¼ þ l þ qu0i u0j qgi bðT T ref Þ oxi oxi oxj oxj oxi
ð2Þ
Energy equation: o o l oT 0 0 ðqui T Þ ¼ quj T oxi oxi Pr oxi
ð3Þ
where, u is the mean of the velocity components (u; v; w), u0i is the velocity fluctuation, p is the pressure, l is the fluid viscosity and Pr is the Prandtl number. Here, xi is the coordinate axis (x, y, z), q is the density, gi is the gravitational acceleration vector and b is the thermal expansion coefficient. The Boussinesq approximation is employed in the last term of Eq. (2). Here, T ref is the reference temperature (system surrounding temperature is used in this work), T and T 0 are the mean temperature and temperature fluctuation, respectively. The term qu0i u0j is called the Reynolds stress (sij) and qu0i T 0 is the diffusion term for the enthalpy. The determination of these terms requires extra equations. The correlation of these terms with the mean flow field is resolved by turbulence models. As the flow situation occurred within the smoking-room is turbulent, the turbulence model was necessarily employed. This study uses the standard k e model in solving for the additional parameter [6]. The accuracy of this model was validated in previous work [1]. The equations for the kinetic energy of turbulence (k) and its dissipation rate (e) are given in the previous work [1]. The simulation studies were performed using a commercial CFD code, FloVENT Version 5.1, on a Windows-based Pentium 4 Computer (2.8 GHz, 1 GB RAM). It is a special-purpose code for building applications using a standard finite volume method and a rectangular structured grid. Convective terms are discretized
880
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
using a first-order upwind scheme [7]. Diffusion terms are always cast into a central difference form. The discretized system is solved with the SIMPLE solving algorithm [8]. 3. Methodology 3.1. Background and model description The present smoking-room dimensions were 2.6 6.2 3.7 m as in the preceding work [1]. The room floor contained 12 100 mm-diameter inlet ducts which were used for the hot gas supply. Two 0.6 0.6-m ventilating lids were installed at the ceiling for the gas outlet. An 8-m chimney with a 200 mm-diameter was used for gas exhaust, as shown in Fig. 1. From the validation study of the smoking-room [1], it was found that temperature variation could be as large as 15 °C (50–65 °C). Highest temperatures occurred near the rear part of the smoking-room and lowest temperatures occurred near the front part of the smoking-room. This implied that most of hot gas flowed into the smoking-room through the supply ducts near the rear wall. Moreover, most of the hot gas was vented out of the smoking-room through the rear ventilating lid. Back flow of outside air into the smoking-room also took place via the front ventilating lid. These results indicated that the size and positions of the ventilating lids affected the ventilation of the hot gas. Therefore, to obtain a uniform-flow situation for the smokingroom, the unsuitable number, size and position of gas supply ducts and ventilating lids needed to be modified. The boundary and initial conditions for the calculation domain for all model studies were set at the same value as in the previous work [1]. A constant static pressure was set at 101 kPa and the ambient temperature was set at 26.7 °C. The external air velocity was set to zero. System boundary extended by the amount of 26 cm to the side of the room, 80 cm to the front door and oven door, and 95 cm to the top of the chimney exit. Dimensions of the cubical system boundary were then 3.12 9.60 10.05 m. This was necessary to allow fully developed flow conditions prior to entering the room. The air velocity entering the rubber room, as a consequence of the buoyancy forces setup within the room due to the fire (heat source), was calculated by the software. At the beginning, the heat source remained identical to that used in the previous work (16 kW). The heat source was a representation of fuel wood used in the smoking process. In the FloVENT program, the ‘‘Heat Source” module was used. Heat transfer from the heat source to rubber sheets in the smoking-room was caused by natural convection of air which is induced to the furnace door and leaves the room through the ventilating lids. The rubber hanging carts were not included in the simulation. The error should be small because the carts contain only metal bar structure which does not block the flow excessively.
Fig. 1. Schematic diagram of the rubber smoking-room model.
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
881
(More details of the domain system of the natural rubber smoking-room can be also seen from our previous study [1].) The total number of grid cells of the system domain covering the room was about 1.4 million. This, however, varies slightly from configuration to configuration. The spacing in each direction is different. A localized grid scheme was used in many positions such as the regions near the ventilating lids, the rubber sheet surfaces and gas supply ducts. The grid spacing in the rubber hanging region was set to 1.5 cm in the direction across the room (between each rubber sheet). The largest spacings used in other directions were coarser but all less than 10 cm. However, the largest spacings in the regions near the ventilating lids and gas supply ducts were varied. The largest grid size in the direction across the room near the gas supply ducts and ventilating lids was 1.5 cm while the largest size in the direction along the room was 2.5 cm. 3.2. Methodology of the study During the removal of moisture from fresh rubber sheets, it is desirable to control the hot gas flow direction to be straight upward so that the moisture can be transported at the highest rate. Uniform upward gas flow and uniform temperature distribution in the entire rubber smoking-room are an ideal combination which will maximize the drying of rubber sheets and minimize the fuel consumption. This condition can be achieved by suitable configuration of hot gas inlets and outlets. To obtain a uniform distribution of the hot gas through the floor, the number of gas supply ducts should be increased. The ideal size and number of the hot gas supply ducts needed to be investigated. The configuration of the outlet of the gas at the ceiling was also crucial to maintaining the ideal flow and temperature distribution in the room. The suitable size, number and position of the ventilating lids were subsequently investigated. This was done via a CFD simulation technique as in the previous work [1]. In addition, the 8-m chimney used for the gas exhaust in air dried sheet (ADS) production was removed. The studies are conveniently considered to comprise four steps: (1) a primary study of the effects of the size and number of gas supply ducts, (2) a study of the effects of the size, position and number of ventilation lids, (3) an investigation of the combined effects of the gas supply ducts and size and position of ventilating lids, and (4) adjustment of the heat supply rate to increase efficiency. Comparison of each improved model was done by investigating the resultant temperature and flow patterns. Temperature contours and flow patterns can be described in three planes as shown in Fig. 2. Moreover, temperatures in the smoking-room can be represented by the values at 27 positions in each plane as also shown in Fig. 2. These positions were selected as the representation of the domain of interest in which the rubber sheets are hung and they are in accordance with the positions used in the previous work [1]. In order to achieve the best outcome, it was necessary to adjust the gas inlets and outlets symmetrically along the middle plane of the room. Hence, there is no need to show the results for all three planes. In this work, results on the middle and right planes are presented.
Fig. 2. Positions of temperature reading in the fully loaded rubber smoking-room by CFD simulation (unit in meter).
882
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
4. Improvement of present rubber smoking-room 4.1. Effects of size and number of gas supply ducts A uniform distribution of hot gas through the bottom rows of rubber sheets was needed. As mentioned in the preceding section, uniform upward flow was desirable. Moreover, the velocity must not be too high. Therefore, the size and position of the gas supply ducts were of great concern in this study. Because the volume of hot gas entering the smoking-room was limited to the proper amount for high efficiency operation, the balance between size and total numbers of the gas supply ducts had to be investigated. Initially, the suitable size of the gas supply ducts was investigated, though their number and position had to be considered simultaneously. Subsequently, the number and position were fine-tuned to obtain the optimal configuration. In this study, the standard sizes of circular ducts; 25-, 50-, 75- and 100 mm-diameters were chosen. From primary investigation, the result of some case studies showed recirculation of hot gas between the smoking-room and the gas supply cavity under the smoking-room. This was the result of using an excessive number of gas supply ducts. For example, when using 75 100 mm-diameter ducts for hot gas supply, the hot gas flowed to the smoking-room from the rear side ducts and recirculated into the gas supply cavity through the ducts at the front part of the room as shown in Fig. 3. In contrast, when using insufficient gas supply ducts, a non-uniform temperature and velocity distribution of hot gas was produced within the smoking-room (as in the case of the present rubber smoking-room that used 12 100 mm-diameter ducts). These results indicate that using the 100 mm-diameter ducts for supplying hot gas makes the control of a uniform flow difficult because their size is too large. When using 75 75 mm-diameter ducts for hot gas supply, the results showed no recirculation, but the velocity and temperature uniformity was poor as a result of the large velocity difference between the front and rear parts of the room. For the case of using 192 50 mm-diameter ducts, the average velocities of hot gas entering the front and rear end ducts were about 0.45 and 0.7 m/s, respectively. The temperature contours within the bottom rows of rubber sheets showed that this duct size gave smaller temperature difference than the 100- and 75 mm-diameter ducts. Many simulations were conducted of the 25 mm-diameter ducts but the calculation required up to 2 million grid cells to obtain accurate results. This required an excessively large calculation time. Moreover, this size of the duct was too small for construction in an actual situation and it could easily
Fig. 3. Temperature contour on the left plane of the example case study.
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
883
be clogged with the accumulation of smoke particles collected on its surface. Therefore, in all of the subsequent cases studied, only 50 mm-diameter ducts were used in supplying hot gas into the smoking-room.
Fig. 4. Positions of the gas supply ducts and ventilating lids for the case 1 shown from the top view.
Fig. 5. Flow pattern of the fully loaded rubber smoking-room of the case 1 at the (a) right plane, and (b) middle plane.
884
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
4.1.1. Case 1 In this case, 192 gas supply ducts with diameter of 50 mm were used and distributed evenly on the floor of the room. Positions of the gas supply ducts are shown in Fig. 4. The size and positions of ventilating lids of the model room remained unaltered from those in the present rubber smoking-room. 4.1.2. Flow pattern Flow pattern at the right and middle planes are shown in Fig. 5(a) and (b), respectively. As the rubber smoking-room is symmetrical about the middle plane, the pattern at the left plane is not shown. The figure of the right plane shows that high velocity occurs at the rear part of the smoking-room. It is a fact that hot gas at high velocity enters the room through the ducts in this section. Maximum and minimum velocities for the right plane are about 0.5 m/s and approximately zero, respectively. The flow pattern at the middle plane shows that hot gas flows out the room from the rear ventilating lid, while some back-flow of outside
Fig. 6. Temperature contour of the fully loaded rubber smoking-room of the case 1 at the (a) right plane, and (b) middle plane.
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
885
air into the room still takes place at the front ventilating lid, as in the case of the present room [1]. The flow pattern at the middle plane is more uniform than that at the side, ranging from 0 to 0.58 m/s. 4.1.3. Temperature Temperature contours at the right, and middle planes are shown in Fig. 6(a) and (b), respectively. All figures show high temperature regions occurring in the rear part along the length of the smoking-room, particularly at the middle plane. This is because most of the hot gas enters the smoking-room through the ducts located in the rear part. The low temperature regions occur in the space below the front ventilating lid. This then causes a density gradient below the front lid which then induced outside air to flow back into the room, as shown in Fig. 6(b). The temperature representation in the room was obtained from the 54 positions in the simulation as described in Fig. 2. Investigation of the temperature results showed that temperatures at the middle and right planes were between 57–74 °C, and 58–67 °C, respectively (Fig. 7). It can be seen that the largest difference of temperature reading for this model is 18 °C, which occurs on the middle plan. This extremely large temperature difference would result in a very non-uniform drying of the rubber sheets which would affect the quality of the rubber sheet and drying time. Rubber sheets exposed to temperature higher than 60 °C will deteriorate because of air bubbles presented in the sheets. The positions of lowest and highest temperature for this model take place below the front and the rear ventilating lids, respectively. This result indicates that the size and
Fig. 7. Temperature reading of the fully loaded rubber smoking-room of case 1.
Fig. 8. Positions of the ventilating lids and the gas supply ducts of the case 2 shown from the top view.
886
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
position of ventilating lids significantly affect the flow in the smoking-room. However, the temperature difference at the right plane is only 9 °C. This model was further studied by varying the size, position and number of the ventilating lids. However, since the temperature in the rear part of the smoking-room was excessively high, the number of gas supply ducts in the rear part needed to be reduced.
4.2. Effects of size, position and number of ventilating lids Velocity distributions at the side and middle planes of the previous studies were non-uniform although there was improvement of temperature distribution. The preceding results indicate that the size of the ventilating lids was too large as back-flow still took place at the front ventilating lids (results not shown). In addition, their positions were unsuitable as most of gas flowed out of the room through the rear ventilating lid.
Fig. 9. Flow pattern of the fully loaded rubber smoking-room of the case 2 at the (a) right plane and (b) middle plane.
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
887
Therefore, in this study, the model room of case 1 was modified. The size, position and number of the ventilating lids for ideal gas venting were investigated. Since the number of the present ventilating lids (2) seemed to be insufficient, and their size (0.6 0.6 m) was probably too large, the study began by reducing their size to 0.25 0.25 m and increasing their number to 12. It was found that this total area was not suitable as back-flow still took place at the ventilating lids near the front part of the room (results not shown). To decrease this effect, their number was reduced from 12 to 8. It should be noted that in this study only the model room having a minimum number and suitable position of the ventilating lids was introduced. 4.2.1. Case 2 In this case, eight ventilating lids with the size of 0.25 0.25 m were used for venting gas at the ceiling. Two rows of four lids were placed as shown in Fig. 8. To decrease the high temperatures occurring at the rear part of the room, the ventilating lids were shifted toward the front part of the room leaving the rear part containing
Fig. 10. Temperature contour of the fully loaded rubber smoking-room of the case 2 at the (a) right plane and (b) middle plane.
888
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
a fewer number of the lids. The distance between the ventilating lids in the row along the length of the room was 1.2 m. To direct the flow of hot gases toward the front part of the room, six gas supply ducts of two middle rows under the carts were removed and one column of the gas supply ducts was added to the front part of the smoking-room as also shown in Fig. 8. Total number of gas supply ducts for this model was 158. 4.2.2. Flow pattern Flow patterns at the right and middle planes of the model room are shown in Fig. 9(a) and (b), respectively. At the right plane, the high velocity region occurs at the middle part of the smoking-room. The velocity for this plane varies from 0 to 0.45 m/s, but the velocity at the front and the rear areas of right plane are very low. This may result from the influence of the ventilating lid positions which were shifted toward the front part of the room. However, the velocity distribution at the middle plane is quite uniform. The velocity for this plane varies from 0 to 0.35 m/s. Flow pattern in this case was improved from the previous case but the velocity at both ends of the sides of the room was still too low. Improvement of velocity distribution in these areas to have more uniformity was needed. 4.2.3. Temperature Temperature contours at the right and middle planes of the model room are shown in Fig. 10(a) and (b), respectively. At the right plane, the high temperature region was slightly shifted to the center part of the room
Fig. 11. Temperature reading of the fully loaded rubber smoking-room of case 2.
Fig. 12. Positions of the ventilating lids and the gas supply ducts of the case 3 shown from top view.
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
889
as a result of moving the ventilating lids toward the front. At the middle plane, high temperatures occur in the center part of the model room. This indicates the influence of high velocity of the hot gas entering the center part of the smoking-room. The temperatures at 54 positions at the right and middle planes were between 66–72 °C and 68–75 °C, respectively (Fig. 11). It can be seen that the maximum temperature variation has been reduced drastically from 19 °C (Case 1) to 9 °C, but it was considered to be still too large. However, the velocity distribution at the right plane is rather non-uniform. The results also show that sufficient ventilating lids are available because no back-flow of outside air into the room occured. Therefore, to reduce the high temperature at the middle part of the room, the distance between the ventilating lids was increased. Some gas supply ducts on the middle rows at the center of the model room were removed to prevent the high temperature occurrence.
Fig. 13. Flow pattern of the fully loaded rubber smoking-room of the case 3 at the (a) right plane and (b) middle plane.
890
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
4.3. Combined effects of position, number, and size of gas supply ducts and ventilating lids Results of the preceding model room show that the temperature variation in the smoking-room was less than 9 °C but velocity distribution was not sufficiently uniform. To obtain better temperature variation, and uniformity of velocity distribution, fine-tuning of the position and number of the gas supply ducts and the size and position of the ventilation lids simultaneously was needed. In this study, the distances between the gas supply duct rows, and the distances between the ventilating lid columns were adjusted to be unevenly distributed. This model of the rubber smoking-room was then used in the following study. 4.3.1. Case 3 In this case, four ventilating lids with the size of 0.25 0.25 m were placed towards the front part and another four ventilating lids with the size of 0.25 0.20 m were placed towards the rear part of the room. Distances between the columns of ventilating lids were, from the front part, 1.2, 1.4 and 1.6 m, respectively;
Fig. 14. Temperature contour of the fully loaded rubber smoking-room of the case 3 at the (a) right plane and (b) middle plane.
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
891
154 gas supply ducts were used. Some gas supply ducts were removed from the preceding model room. The distances between the rows of the gas supply ducts were unevenly distributed along the length of the room. Positions of the ventilating lids and the gas supply ducts are shown in Fig. 12. 4.3.2. Flow pattern Flow patterns of hot gas at the right and middle planes are shown in Fig. 13(a) and (b), respectively. At the right plane, velocity distribution is significantly improved from the previous model rooms. But low velocity occurs near the front and rear ends of the room. However, most of these areas are empty space outside the rubber hanging carts; therefore this region was not studied in this research. The velocity at the right plane varies from 0 to 0.40 m/s. However, the velocity distribution at the middle plane is quite uniform as in the preceding case. The velocity for this plane varied from 0 to 0.25 m/s. 4.3.3. Temperature Temperature contours at the right and middle planes of the model room are shown in Fig. 14(a) and (b), respectively. Results from investigation of temperature at 54 positions of the right and middle planes show that temperatures at the right and middle planes were between 67–72 °C and 68–73 °C, respectively (Fig. 15). In conclusion, this model room (case 3) gave good results of temperature and velocity distributions. Maximum temperature variation in the model room was found to be only 6 °C, while the velocity distribution was sufficiently uniform. However, average temperature in this model room was about 70 °C which is higher than the suitable temperature used for rubber sheet smoking process. In general, the temperature for rubber sheet drying should not exceed 60 °C to prevent deterioration of the product. This high temperature was a result of the removal of the exhaust draining (draft) tube. All of the hot gas then flowed into the smoking-room without any loss to outside. The heat source of this model was still the same as that of the original rubber smokingroom (16 kW). Therefore, in the following case study, the heat source input was reduced so that the average temperature of the smoking-room did not exceed the suitable temperature (60 °C). This resulted in the reduction of fuel usage. 4.4. Adjustment of the heat source input to the new model room The acceptance of a new model of the rubber smoking-room depends on temperature and velocity distributions for suitable rubber smoking process. In general, the maximum temperature during the smoking
Fig. 15. Temperature reading of the fully loaded rubber smoking-room of case 3.
892
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
process should be controlled at about 60 °C [2]. In this study, the heat source input was iteratively reduced so that the average temperature was about 60 °C. 4.4.1. Case 4 In this simulation, the geometry of the smoking-room remained unchanged from the case 3. The heat source was reduced from 16 kW to 12 kW (i.e., a 25% reduction). 4.4.2. Flow pattern Flow patterns of hot gas at the right and middle planes of the model room are shown in Fig. 16(a) and (b), respectively. Characteristics of velocity distribution in this case are the same as in case 3. At the right plane, low velocity occurs near the front and rear ends of the room, which is empty space outside the rubber hanging carts. The variations of velocities at both planes are identical to that in case 3.
Fig. 16. Flow pattern of the fully loaded rubber smoking-room of the case 4 at the (a) right plane and (b) middle plane.
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
893
Fig. 17. Temperature contour of the fully loaded rubber smoking-room of the case 4 at the (a) right plane and (b) middle plane.
4.4.3. Temperature Temperature contours at the right and middle planes of the model room are shown in Fig. 17(a) and (b), respectively. Both figures show a small temperature difference. Results from investigation of the temperature at 54 positions on the right and middle planes show that temperatures at the right and middle planes are between 59.0–63.5 °C and 61.0–64.5 °C, respectively (Fig. 18). This model gives good results of temperature and velocity distributions. Average velocity of this model is about 0.25 m/s. The temperature variation of this model room is about 5.5 °C and the average temperature predicted by the model is about 62.1 °C, which is close to the suitable level (60 °C). To obtain the temperature at the optimal level, the heat source was then further adjusted. It was predicted that a suitable heat source is 11 kW. This means a 31.25% reduction of energy consumption attainable for rubber smoking process. The temperature variation in this case was 5.3 °C and the average temperature was 60.4 °C.
894
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
Fig. 18. Temperature reading of the fully loaded rubber smoking-room of case 4.
5. Conclusion Results from computer simulation aimed at improving the rubber smoking process indicate that size, position and numbers of gas supply ducts and ventilating lids significantly affect the temperature and velocity distributions. The most suitable rubber sheet smoking-room with the size of 2.6 6.2 3.6 m should contain 154 50 mm-diameter hot gas supply ducts, and four 0.25 0.25 m, and four 0.25 0.20 m ventilating lids. The predicted heat source input is 11 kW. The reason for choosing this size of the smoking-room is that it is suitable for general rubber smoking cooperatives located throughout the country. It should be kept in mind that the simulation here assumes that steady-flow conditions occur. This is usually not the case in the actual smoking process. However, this improved model should help the rubber smoking cooperatives to save energy by at least 31.25% because it has shown that the heat source input can be reduced by that percentage from the original rubber smoking-room. The village entrepreneurs who operate such a cooperative will in turn receive more profit from the energy saving by implementing this model. Moreover, the quality of rubber sheets will be improved. In the future work, effects of moisture content will be studied as this also affects the drying environment. Drying characteristics of the rubber sheets will be included. Accuracy of the simulation will then be increased. Acknowledgments This research was financially supported by the Graduate School, Prince of Songkla University, Thailand. Thanks also go to Michael Allen and Alan Geater, for correcting the use of English and helpful suggestions. References [1] Promtong M, Tekasakul P. CFD study of flow in natural rubber smoking-room: I. validation with the present smoking-room. Appl Therm Eng 2007;27:2113–21. [2] Prasertsan S, Kirirat P. Factor affecting rubber sheet curing. RERIC Int Energy J 1993;15:77–87. [3] Pomvisaid J. Affecting of air velocity and temperature for rubber sheet drying. Project No. ME 6/2537, Department of Mechanical Engineering, Faculty of Engineering, Prince of Songkla University, Thailand; 1984. [4] Tonsattayaleard S. Parametric optimization of rubber sheet drying. Project ME.25/2542, Department of Mechanical Engineering, Faculty of Engineering, Prince of Songkla University, Thailand; 1999.
P. Tekasakul, M. Promtong / Applied Energy 85 (2008) 878–895
895
[5] Kalasee W, Pongpariyawatee Y, Tantaprapa W, Suwanjaras S, Kirirat P, Tekasakul P. Fuel efficiency enhancement in rubber smoking for rubber cooperative in Southern Thailand. In: The 17th conference on mechanical engineering network of Thailand, Prachinburi, Thailand, 2003, pp. 560–564. [6] Launder BE, Spalding DB. The numerical computation of turbulent flows. Comput Meth Appl Mech Eng 1974;3:269–89. [7] Roe P. Approximate Riemann solvers, parameters vectors, and difference schemes. J Comput Phys 1981;43:357–72. [8] Patankar SV. Numerical heat transfer and fluid flow. Washington: Hemisphere; 1980.
