Editorial - Mahasarakham Universityjournal.msu.ac.th/upload/journal_file/jfile_no1_45563.pdf ·...

Editorial

The ASEAN stream has, not only fl ooded into all member nations, but also brightened up our world ofacademic activities. Our Maha Sarakham Journal of Science and Technology has published leading research papers in many fi elds to enhance the basic and also the applied knowledge for all ASEAN countries. In this issue of ours, there are papers deal with the health of our ASEAN population, advance in many aspects of science and technology for our ASEAN nations, as well as our ASEAN own background of history and culture. Looking and evaluating ourselves before jumping into ASEAN tides of advancement is still be our main objective. We hope this issue of Maha Sarakham Journal of Science and Technology meets all satisfaction and expectation of our subscribers.

(Prof. Dr. Pradit Terdtoon) Executive Editor

Contents

Original page

Impacts of Air Staging on NO Emission from a Conical Fluidized-bed Combustor Firing Sunfl ower Shells .......................................................................................................................... 1 Kasama Sirisomboon, Piyanat Charornporn, Porametr Arromdee

Modifi ed Form of Low-Reynolds-number Turbulence Model for Predicting Turbulent Heat Transfer .............. 8 Kittisak Khuwaranyu, Sompong Putivisuttisak,

Development of the Typical Meteorological Data for Silpakorn University Sanamchandra Palace Campus Zone .................................................................................................................... 14 Jarut Kunanoppadol

The Concept to Measure the Overall Car Performance ........................................................................................ 24 Jarut Kunanoppadol

Improvement of Dehumidifi cation Effectiveness Prediction Models .................................................................. 30 Thosapon Katejanekarn, Supamas Wichaisin

Infl uence of Operating Parameters on the Performance of a Liquid Desiccant Ventilation Dehumidifi cation System ......................................................................................................................................... 37 Thosapon Katejanekarn, Watanyoo Panangnuwong

Thermal Performance of Solar Water Heater Integrated Vacuum-tube Collector with Loop Thermosyphon ....................................................................................................................... 45 T. Hudakorn, G. Boonyaaroonnate

Culture media for growth of mycelium and for induction of sporulation of Ascosphaera apis, the causative agent of chalkbrood disease of honey bee (Apis mellifera L.)..................................................... 51

Tipaporn Subsomboon, Pittaya Liewseree, Sukum Kositchaimongkol

Biomass Pellet .......................................................................................................................................................... 56 Sonthi Warajanont, Pongsiri Jaruyanon, Nitipong Soponpongpipat

Kinetics of Biomass Decomposition in Pyrolysis and Torrefaction Process .................................................... 64 Unchana Sae-Ueng, Noppong Sritrakul, Nitipong Soponpongpiapat

Properties of Torrefi ed Biomass and Torrefi ed Pellet ........................................................................................... 73 Dussadeeporn Baonongbua, Nitipong Soponpongpiapat, Supachai Wasananon,

Torrefaction Reactors .............................................................................................................................................. 84 Worason Junsatien, Nitipong Soponpongpipat, Sivapong Phetsong

Contents

Original page

Components of the Discussion Section in Biomedical Engineering Research Articles and their Linguistic Characterization ..................................................................................................................... 92 Budsaba KanoksilapathamA Comparison of Learning Achievement Toward cost accounting I Between The Group has Solution of Homework and The Group not has Solution of Homework. ................................. 98 Manop Seeluang

Development of Temperature Controlling System in drying chamber by mixing air inlet ............................ 101 Sivapong Phetsong

Design and Development of an Automatic Swine Feeding Machine Cooperated with RadioFrequency Identifi cation Technique ..................................................................................................................... 106 Saroj Pullteap

The Structural Relationship Model of Factors Affecting Perceived Shared Value for Intention to Stay ....... 112 Napaporn Kantanapa, Montree Piriyakul, Dechaphan Ratsasanasart

The Effect of Equivalent Ratio on the Performances and Emissions of Diesel Engine .................................. 118 Thibordin Sangsawang, Supachai Wasananon, Aminan Amornpattarakiat

Daylight Performance of an Automated Vertical Blinds System ....................................................................... 123 Vichuda Mettanant

คําอธิบายภาพปก : Screw conveyor torrefaction reactor Multiple hearth torrefaction reactorภาพปก : Worason Junsatien. 2556, 84-91

Original

Impacts of Air Staging on NO Emission from a Conical Fluidized-bed Combustor Firing Sunfl ower Shells

Kasama Sirisomboon1, Piyanat Charornporn1, Porametr Arromdee1

Received: 25 January 2012 ; Accepted: 11 April 2012

AbstractIn this experimental study, a conical fl uidized-bed combustor (conical FBC) of 0.9 m inner diameter and 3.4 m height was used to fi re sunfl ower shell at 45 kg/h, excess air about 60% at the secondary to total air ratio (S/T) 20, 40 and 60% for two different locations of secondary air injection: 1.6 and 2.6 m above the air distributor. The axial temperature as well as O

2, CO and C

xH

y concentrations profi les in the reactor as well as at the combustor outlet were investigated.

In the air-staged combustion, fl ue gas temperature was found to be at maximum in the vicinity of fuel injection into the combustor, nevertheless, reduced to the minimum in the upper region (close to a level of secondary air injection). Due to the reduced concentration of O

2 in the bed region during the air-staged combustion of sunfl ower shells, CO

showed signifi cant values in this region, however, drastically reduced in the combustor freeboard above the second-ary air nozzles. As revealed by the experimental data, the percentage of secondary air as well as the location of which secondary air is injected showed signifi cant effects on the axial O

2, and CO concentration profi le, whereas, the

axial CxH

y and NO

x concentration profi le exhibited rather weak effects of this operating variable. Taking into account

the emission characteristics, the percentage of secondary air of 20-40% and the injection level of 2.6 m (above the primary air distributor) seems to be the optimum option for the effective mitigating of NO

x emissions. In the range of

these operating conditions, NOx emissions quantify at minimum value, ~220−227 ppm at the rather low CO emission

in the fl ue gas.

Keywords: air-staged combustion, fl uidized-bed combustor, sunfl ower shells.

IntroductionThailand is one of the agricultural-based countries, many biomass fuels such as rice husk, bagasse and energy crops are available for heat and power generations. While the conventional biomass were traditional utilized in the rice mill, sugar industrial and biomass-fuelled power

plant, the processed waste such as tamarind shells, peanut shells, palm stalks, cotton stalk apricot and date stones, corncob and rice straw were growing

attention as the biomass fuels in the past decade. A large number of the research studies have been mainly focused on the combustion effi ciency and emission performance of the fl uidized-bed combustion system fi ring biomass fuels. The burning of some uncon-

ventional biomass fuels (olive cake, peach and apricot stones) reported to have the high fuel-ash content, big

particle size and high fuel-moisture content, the CO, CxH

y

emissions and the unburned carbon associated with fl y ash showed the elevated values leading to the lower

combustion effi ciency [1-2]. While the CO emissions and unburned carbon in Ref. [2-6] are generally controlled by the optimal operating conditions: excess air, combustion

temperature and fuel feed rate. The NO emission from fl uidized-bed combustor appear to be strongly infl uenced by the nitrogen content of the biomass fuels, however, some studies reported the signfi cant effects of excess air and air-staged combustion on NO emission [2, 7-9].

1 Faculty of Engineering and Industrial Technology, Silpakorn University (Sanam Chandra Palace), Nakhon Phathom 73000, Thailand. Email: [email protected]

Sirisomboon et al. J Sci Technol MSU2

In this research study was aimed at studying the effects of air staging on NO emission when fi ring the sunfl ower shells in the conical fl uidized bed combustor. The axial temperature, O

2, CO and NO concentrations

along the combustor height as well as at the cyclone outlet were investigated for various secondary to total air ratio at different locations at which secondary air is supplied.

Experimental Experimental Set-up Fig. 1 shows the schematic diagram of an experimental set-up with the conical FBC with the cone angle of 40° and inner diameter of 0.25 m at the bottom plane and 0.9 m inner diameter of at the upper part. The

total height of the combustor was 3.4 m consisted of a conical (bottom) section with 1.9 m height and a cylindrical (upper) section with 2.5 m height. The primary air (or fl uidizing air) was supplied through the 13-bubbling-cap air distributor at the bottom part of the combustor by a 25-hp blower. The secondary air was tangentially supplied through a 0.04-m inner diameter pipes at each level of 1.6 and 2.6-m above the air distributor by 5-hp blower. The schematic diagram of the air staging system is also shown in Fig. 1. A screw-type feeder delivered the tested biomass fuel over the bed at 0.6 m above the air distributor.

Table 1 Ultimate and proximate analyses and lower heating value (LHV) of Sunfl ower shells used in experimental studies on the conical FBC

Figure 1 Schematic diagram of the experimental set-up with the conical fl uidized-bed combustor.

Vol 32, No1, January-February 2013 Impacts of Air Staging on NO Emission from a Conical Fluidized-bed

Combustor Firing Sunfl ower Shells3

Silica sand with particle sizes of 0.3-0.5 mm at the static height of 30 cm was used as the bed material in this study. To avoid the bed agglomeration, it was replaced every 18 hours of use [4]. The combustor had gas sampling ports, as well as stationary Chromel-Alumel thermocouples (of type K) for measuring the temperatures in the axial direction inside the rector during the experimental tests. To quantify the gaseous concentrations (O

2, CO and NO) along the

combustor and at the cyclone exit, a model “Testo-350XL” gas analyzer (Testo, Germany) was used to monitor the temperature and gas concentrations.

Fuel Characteristics Table 1 shows the ultimate and proximate analyses as well as the lower heating value (LHV) of the sunfl ower shells used in this work. It can be seen in Table 1 that sunfl ower shells were characterized by a signifi cant amount of volatile matter (VM), a moderate proportion of fi xed carbon (FC), and relatively low contents of fuel-moisture (W) and fuel-ash (A) in the proximate analysis, which resulted in a substantial LHV. Note that the sunfl ower shells contained proportion of fuel-N (1.0 wt.%) and fuel-S (0.1 wt.%), therefore, the SO

2 was not addressed in this study.

2.3 Experimental Planning In the experimental tests, the sunfl ower shells were burned at the fi xed feed rate (FR), 45 kg/h, excess air (EA) about 60% at the secondary to total air ratio (S/T) 20, 40 and 60% for different locations of which secondary air (Z) were supplied: 1.6 and 2.6 m above the air distributor. For the comparative study with the air-staged combustion, the conventional combustion of sunfl ower shells without secondary air injection was also conducted at the fuel feed rate of 45 kg/h and excess air about 60%.

Results and Discussions Combustion Characteristics Fig. 2 shows the effects of air staging and the the location of secondary air injection on the axial temperature and O

2 concentration profi les. As seen in the

fi gure, the location of secondary air injection has shown the signifi cant effects on both axial temperature and O

2

concentration profi les.

Figure 2 Effect of secondary air injection level on the axial temperature (a) and O2 concentration (b) profi les along the conical FBC when

fi ring Sunfl ower shells at EA ≅ 60% and S/T = 40%.


The maximum combustion temperature 1027oC was found at level 1.6 m above the air distributor when the secondary air was injected to the combustor at Z = 2.6 m. The minimum fl ue gas temperature 688oC and the low oxygen consumption was found whensecondary air was supplied at Z = 1.6 m. While the injection at the higher level seems to enhance the good mixing between bed material and the selected biomass fuel (char particles), vice versa, it retarded the oxidation of fuel when the secondary air was injected into the combustor at the lower level. Due to the vortexing fl ow of secondary air in the reactor, it was found that the maximum temperature occurred at the level below the secondary air injection for both casestudies.

Emissions Characteristics Fig. 3 shows the effects of air staging and the location of secondary air injection on the axial CO NO and C

xH

y (as CH

4) concentration profi les. The maximum

CO concentrations in every test were observed in the bed region near the location of fuel injection (at 0.6 m about the air distributor). The highest CO concentration was found in the combustion test with air staging at Z = 1.6 m, mainly

because of the lower combustion temperature, the higher devolatilization (of volatile hydrocarbon to CO) and the

less oxygen consumption (see Fig. 2b) in this region. However, the high CO oxidation occurred in the freeboard

region in the vicinity of secondary air injection into the combustor. Comparing with the air-staged combustion, the CO concentrations in the conventional combustion

were signifi cantly lower, particularly in the bed region, because of the higher rate of char-C oxidation. In contrast to the axial CO concentration profi les, the axial NO concentrations seems to independent from the location at which secondary air was injected. For every experimental test, the highest NO concentrations observed at the same level ∼0.8 m above the air distributor. In the upper region of the combustor (at Z > 2 m), the NO reduction rates in the air-staged combustion tests were found to be higher due to the greater catalytic reduction of NO by CO (on the char surface). As seen in Fig. 3c, the highest C

xH

y concen-

trations in fl ue gas were also found in the bed region at the level in vicinity of the fuel injection. However, the C

xH

y

were effectively mitigated by secondary air, to relativesmall value in the freeboard region (at Z > 2 m), thus leadingto the apparently lower than the C

xH

y concentrations

in the air staging tests.

Major Emissions in Flue Gas Fig. 4 depicts the CO, C

xH

y (as CH

4) and

NO emissions (all presented on dry gas basis and at 6% O

2) from the conical FBC fi red with sunfl ower shells at

45 kg/h, excess air about 60% at the secondary to total air ratio (S/T) 20, 40 and 60% for two different locations

of secondary air injection: 1.6 and 2.6 m above the air distributor.

Whereas the location of which secondary air is injected presented signifi cant effected on the CO

emissions as seen in Fig.4a, it showed slightly effect on the NO emissions. While C

xH

y emissions seemed to

independent from this operating variable.



Due to the higher residence time for C-char oxidation in the conventional combustion (fi ring without

secondary air injection), the CO emissions was found at rather low value (at 232 ppm) compared with the tests

with air staging. In air-staged combustions, the CO emission

seems to be greater when the secondary air was injected at Z = 2.6 m, and the maximum value was 800 ppm at S/T = 40%. In contrast to the CO emissions, NO emissions was found to be reduced to the satisfaction value (220 ppm) for the same operating conditions.

Figure 3 Effect of secondary air injection level on the axial CO, CxH

y and NO concentration profi les inside the conical FBC fi ring Sunfl ower

shells at EA ≅ 60% and S/T = 40%.

From CO, CxH

y and NO emissions in the

range of experimental tests, it can be clearly seen that the main chemical reactions responsible for NO reduction is the catalytic reduction of NO by CO when fi ring the sunfl ower shells in the proposed fl uidized-bed combustor.

Conclusion As revealed by the experimental data, the

percentage of secondary air showed signifi cant effects on the axial O

2, CO and C

xH

y concentration profi le, whereas

the axial NOx concentration profi le exhibited rather weak

effects of this operating variable.


The percentage of secondary air of 20–40% and the injection level of 2.6 m (above the primary air distributor) seems to be the optimum option for effective mitigation of NO

x emissions in the conical FBC.

Under these operating conditions, the NOx emis-

sions can be controlled at a minimum value, about 220

ppm (on a dry gas basis and at 6% O2), whereas the CO

emissions was in the acceptable value, about 800 ppm.

Acknowledgement The authors would like to acknowledge the financial supports from the Thailand Research Fund (contract No. MRG 5380019) and Department of Mechanical Engineering, Faculty of Engineering and Industrial

Technology, Silpakorn University. The authors also sincerely thank Associate Professor Dr.Vladimir I. Kuprianov (School of Manufacturing

Systems an Mechanical Engineering, Sirindhorn International

Figure 4 Effects of air staging on the emission of CO, CxH

y (as CH

4) and NO from the conical FBC fi red with sunfl ower shells.

Institute of Thecnology, Thammasat University) for his

kind advices and providing us the experimental apparatus. Special thanks to Mr. Pichet Ninduangdee for his effective

help in experimental tests.

References1. Kaynak B. Topal H. and Atimtay T., Peach and

apricot stone combustion in a bubbling fl uidized bed, Fuel processing technology, p.1175−1193, 2005.

2. Atimtay A.T. and Voral M., Investigation o co-com-bustion of coal and olive cake in a bubbling fl uidized bed with secondary air injection, Fuel, p.1000−1008, 2009.

3. Kouprianov V. I. Janvijitsakul K. and Permchart W,

Co-fi ring sugar cane bagasse with rice husk in a conical fl uidize-bed combustor, Fuel, p.434−442, 2006.



4. Permchart W. and Kouprianov V. I., Emission performance and combustion effi ciency of a conical fl uidized-bed combustor fi ring various biomass fuels, Bioresource Technology, p.83−91, 2004.

5. Porametr A. Vladimir I. K. Rachadaporn K. Kasama S, Experimental study on combustion of sunfl ower shells in a pilot swirling fl uidized-bed combustor, Energy and Fuels, p.3850−3859, 2010.

6. Demirbas A. Combustion characteristics of different biomass fuels, Progress in energy and combustion science, p.219–230, 2004.

7. Werther J. Saenger M. Hartge E. U. Ogada T. and Siagi Z, Combustion of agricultural residues, Progress in energy and combustion science, p.1−27, 2000.

8. Lyngfelt A and Leckner B. Combustion of wood-chipsin circulating fl idized bed boilers e NO and CO emissions as functions of temperature and air-staging. Fuel, vol. 78, p.1065−1072, 1999.

9. Fang M. Yang L. Chen G. Shi Z. Luo Z. and Cen K. Experimental study on rice husk combustion in a circulating fluidized bed, Fuel Process Technology, vol.85, p.1273−1282, 2004.

10. Chyang CS,Wu KT, Lin CS. Emission of nitrogen oxides in a vortexing fluidized bed combustor, Fuel, vol. 86, p.234−243, 2007.

Original

1 Corresponding author, Faculty of Engineering and Industrial Technology, Silpakorn University (Sanam Chandra Palace), Nakhon Phathom 73000, Thailand. Email: [email protected].

2 Faculty of Engineering, Chulalongkorn University, Bangkok, 10330

Modifi ed Form of Low-Reynolds-number Turbulence Model for Predicting Turbulent Heat Transfer

Kittisak Khuwaranyu1, Sompong Putivisuttisak2


AbstractThis research presents the Finite Volume Method (FVM) for predicting turbulent fl ow and heat transfer. For combing the Low-Reynolds-Number (LRN) k–ω model with a Length Scale Correction (LSC) term and High-Re model, the Baseline model was employed. The new model is so call “BLL model” for complex heat transfer and fl ow phenomena. The modifi ed model is validated with benchmark problems (the fully developed channel fl ow and backward-facing step fl ow and heat transfer) before being applied to the problem of recirculating fl ow and heat transfer over repeated square ribs. Performance of the model is investigated and compared with available experimental data, Direct Numerical Simulation (DNS) data and numerical results using other turbulence models. It is seen that the model gives superior results especially for the near-wall fl ow patterns. The results demonstrate that the performance of the BLL model is much better than that of the High-Re model, LRN model, and it can be shown that this model is the appropriate choice for near-wall fl ow problem.

Keywords: LRN model, k–ω, Turbulent, Heat transfer, Finite volume Method

IntroductionMany form of turbulence models which are used for predicting turbulence effects of heat transfer and fl ow phenomena. The most popular form is the one proposed

by Launder and Spalding1 so call the standard k–ε model. The disadvantage of the standard k–ε model is the inability to predict accurate results for fl ow with adverse pressure gradients. The LRN models shared a weak point with the LRN k-ε model which was the uncertainty of ε specifi cation at the wall2. Pang and Davidson3 implemented the LRN k-ω model by introducing a damping function into the

turbulent kinetic energy term. Their model was compared with two LRN k-ε models and one LRN k-ω model. Several concepts were also used to increase the

model accuracy. Jia et al.4 integrated the reformulated SSG model5 based on the ω-equation and the SST

model6. The obtained result was better than the previous one because the new model had good applicability for complex fl ow fi elds This research presents a concept for a new turbulence model, which combines the Low-Reynolds-Number (LRN) k-ω model with a Length Scale Correction (LSC) term. For combing the LRN the Baseline model was employed. The new model is so call “BLL model” for complex heat transfer and fl ow phenomena. Here, the capability and performance of the BLL models for

calculation of a thermal viscous fl ow will be made.

Theoretical formulations The Reynolds-averaging principle is applied to

the Navier-Stokes equations. After performing the aver-aging, the continuity and momentum equations can be shown as follows:

Modifi ed Form of Low-Reynolds-number Turbulence Model

for Predicting Turbulent Heat Transfer9Vol 32, No1, January-February 2013

Turbulence models In the present work, a new model called BLL model which is based on LRN k–ε with LSC and trans-formed k–ε , is proposed. The standard k–ε model7, LRN k–ε model8 and High-Re k–ε model9 are employed for results comparison. The BLL model The basic idea of the new model is based on a combination of an accurate formulation of the Wilcox LRN k–ε model and the concept of a Baseline model to reduce the sensivity to the freestream (in the outer part of the boundary-layer and in free-shear fl ows). In the near-wall region, the Length Scale Correction (LSC) term is employed. The equations of the proposed model are reformulated by multiplying the LRN k–ε model with LSC term by a function (1–F

b) and adding with the

multiplication of the transformed High-Re k–ε equation and a function F

b. F

b is a blending function (a simple

exponential function is used at the beginning) which ensures that the model behaves as a High-Re model away from the surface and as the LRN model in the near-wall

region. For the dissipation rate, the cross diffusion term is removed. To compensate for the inferior performance of the transformed k–ε, the LSC term () is employed. After rearrangement, the new model can be

shown as follows:

where is the turbulent viscosity

and kP is the production of the turbulent energy which can be expressed as:

The model constant is blended by the relation of the model constants in the LRN k–ω model and the transformed k–ε model. If represent constants in the LRN k–ω model ...,, 11 kk f 2 , represent constants in the transformed k–ε and represent the corre-sponding constants of the new model

relation can be written as:

Two sets of model constants are given as

follows:

Khuwaranyu et al. J Sci Technol MSU10

After rearrangement, the new model can be shown as follows:

where is the turbulent viscosity and is the production of the turbulent energy which can be expressed as:

The blending function (Fb) is selected to ensure

asymptotic consistency with the near-wall behavior of the equation of motion. The value of function Fb will be designed to be zero in the near-wall region (activating the LRN model) and set to unity away from the surface (switching to the High-Re model) as shown in Fig. 1. For the present model, the blending function from the LRN two-equation model of Abe et al.8 has been tentatively adopted, as expressed in Eq. (8):

Length Scale Correction term (LSC term) The LSC term has been well known for turbulent separated fl ows. Launder12 applied the Yap correction13 as an extra term to the ε-equation of LRN k-ε model. The Yap correction can be written as the ratio of the compu-tational length scale to the local equilibrium length scale as follows:

where is the local equilibrium length scale is the normal distance to the

nearest wall. Wilcox9 has shown that an extra cross-diffusion term appears in the resultant ε-equation. This term, similarto the so-called “Yap correction”, helps suppressing the rate of the near-wall turbulent length scale. In the present

work, the turbulent length scale isapplied to the Yap correction concept. A new length scale correction for ω-equation (LSC-ω) can be expressed as follows:

where is the near-wall equilibrium length scale yn is the distance from the wall, the turbulence scale constant (C) is equal to and is the von Kárman constant.

Figure 1 Blending function Fb versus y+

Energy equation

For including turbulent effect, the energy equation will have the new unknown terms, so call turbulent

heat fl ux. The terms are expressed by the Boussinesq approximation as:



Numerical procedures The fi nite volume method (FVM) is used to solve the set of governing equations on a staggered grid. To obtain the solution that couples the pressure and velocity,the SIMPLE algorithm14 is employed to calculate the pressure correction terms. Then, the resulting algebraic equations are iteratively solved with a line-by-line TDMA procedure. The convection terms are approximated by the second-order upwind scheme. The hybrid differencing scheme is employed in the turbulence-transport equations to ensure a stable solution procedure. The convergence criterion utilized in this work is that the maximum nor-malized sum of the absolute residual source for all the computed nodes is less than 10-6. For all the investigated cases, the number of suffi ciently small grids is ensured from the grid-independency tests. The inlet boundary values are prescribed for all variables. At the outlet, the streamwise gradients of the fl ow variables are set to zero. The wall boundary condi-

tions are applied: u = v = 0 and k = 0. The use of LRN models requires a fi ne grid to minimize the dependence of the solution on the grid. The boundary condition of εat the fi rst grid point in the near-wall region is given as10

Results and Discussions Turbulent heat transfer and fl ow in smooth wall channel for = 180 The selected fi rst test case is a fully-developed channel heat transfer and fl ow at = 180, for which DNS data exist. The simulating condition, the wall boundary is applied uniform wall temperature, the Reynolds number,

is 180 and Pr is 0.71 . The predicted results are compared with the DNS data from Kim and Moin15 and the numerical results of other turbulence model, ANK model and Wilcox model.

Figure 2 Comparison of the predicted temperature with the DNS,

AKN model and Wilcox model

Figure 2 presents the dimensionless of temperatureat different. The numerical results are in good agreement with the DNS data and the results of ANK model but we can see that the Wilcox model slightly underpredicts for the

region > 20. The Wilcox model result, its is high sensivity in freestream that cause the under predicted result.

Turbulent heat transfer and fl ow over Back-ward facing step

The proposed model predicts the simple fl ow in smooth channel wall. The next suitable test case is the turbulent fl ow over Backward-Facing Step (BFS). The fl ow over BFS is widely use in model test because this fl ow are complex fl ow phenomenon such as sudden expansion

fl ow, recirculating and reattachment. In this research, we have applied the model to selected case of BFS fl ow, for which simulating parameters are available from Avancha

and Plether16 for 5540.

Khuwaranyu et al. J Sci Technol MSU12

In the test case, the boundary conditions and properties are, the Reynolds number, , based on the backward facing step height and free stream velocity is 5540. Where , Uref is free stream velocity at inlet and h is the step height, the confi guration is shown in Figure 3. The test section is 41 mm height (h), Expansion ratio (ER) is 1.5, the channel length behind BFS and inlet section height are 20h and 3h, respectively. The magnitude of 1000 W/m2 uniform heat fl ux applies at bottom wall behind the step. For the heat transfer investigation, the Staton number, ratio of the convection heat transfer and heat fl ux, can be describe as:

Where h is the convection coeffi cient of fl uid and T

w(x) and T

b(x) are the wall temperature and bulb

temperature at different location of x, respectively. Figure 4 presents the distribution of the Staton number along bottom wall behind the step. It can be seen that the results from the k–ω model and the BLL model is better in agreement with the LES data than those of the standard k–ε model. But three models give the Stanton number distribution in similar curve line, it can be explain that heat transfer capacity is low at circulation region and increase over the region behind the backward-facing step. Figure 5 shows the mean temperature, ((T-Tref) / Tref), at the location x/h = 1.0, 2.0, 3.0, 4.0 and 5.0 . The BLL model can be predict in good agreement with the LES data and the k–ε model give a slightly under

prediction of temperature. Moreover, the Standard k–ε model obtains underprediction of mean temperature at all of investigated section, x/h.

Conclusion In this work, the developed model proposed for improving and increasing prediction accuracy. The results demonstrate that the standard k–ε model give good

predicting in simple fl ow. For complex heat transfer and fl ow, the AKN model and the BLL model produce similar results. However, the results obtain that the performance

of the BLL model is much better that that of the High-Re model and LRN model. For improving the model, the

complex heat transfer and fl ow phenomena need to apply for the BLL model.

Acknowledgement This work was supported by the Research Grant for New Scholar (RGNS) from the Commission on Higher Education and Thailand Research Fund (TRF) with Prof. P. Dechaumphai as the project mentor.

References1. Launder B E, Spalding D B. Mathematical Models of

Turbulence. Academic Press, London 1972.2. Peng S H, Davidson L, Holmberg S. A modifi ed

low-Reynolds-number k–ε model for recirculating fl ows. Journal of Fluids Engineering 1997; 119(4): 867-875.

3. Peng S H, Davidson L. Computation of turbulent buoyant fl ows in enclosures with low-Reynolds-number k-ε models. International Journal of Heat and Fluid Flow 1999; 20(2): 172-184.

4. Jia R, Sunden B, Faghri M. A new low Reynolds stress transport model for heat transfer and fl uid in engineering applications. Journal of Heat Transfer 2007; 129(4): 434-440.

5. Spezial C G, Sarkar S, Gatski T B. Modelling the pressure-strain correlation of turbulence : an invariant dynamical systems approach. Journal of Fluid Mechanics 1991; 227: 245-272.

6. Menter F R. Two-equation eddy-viscosity turbulence models for engineering applications. AIAA Journal 1994; 32(8): 1598-1605.

7. Launder B E Spalding D B. The numerical computationof turbulent fl ows. Computer Methods in Applied

Mechanics and Engineering 1990; 3: 269-289. 8. Abe K, Kondoh T, Nagano Y. A new turbulence model

for predicting fl uid fl ow and heat transfer in separating

and reattaching fl ows - I. Flow fi eld calculations.International Journal of Heat and Mass Transfer 1994; 37(1): 139-151.

9. Wilcox D C. Turbulence Modeling for CFD. DCW Industries Inc, California 1993.

10. Wilcox D C. Simulation of transition with a two-equation turbulence model. AIAA Journal 1994; 33(2): 247-255.



11. Jones W P, Launder B E. The calculation of low-Reynolds-number phenomena with a two-equation model of turbulence. International Journal of Heat and Mass Transfer 1973; 16: 1119-1130.

12. Launder B E. The Proceedings of the Second International Symposium - Engineering Turbulence Modeling and Experiments 2, Modelling convective heat transfer in complex turbulent fl ows, Elsevier press, Florence, Italy 1993.

13. Yap C J. Turbulent Heat and Momentum Transfer in Recirculating and Impinging Flows. Ph.D. Thesis, University of Manchester. 1987; 323 pp.

14. Patankar S V. Numerical Heat Transfer and Fluid Flow. McGraw-Hill, Washington, USA 1980.

15. Le H, Moin P, Kim J. Direct numerical simulation of turbulent fl ow over a backward-facing step. Journal of Fluid Mechanics 1997; 330: 349-374.

16. Avancha R V R, Pletcher R H. Large eddy simulation of the turbulent fl ow past a backward-facing step with heat transfer and property variations. International Journal of Heat and Fluid Flow 2002; 23: 601-614.

Figure 3 Mean Staton number

Figure 4 Mean temperature distribution

Original

1 Engineering Business Centre (EBC), Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, SilpakornUniversity, Nakhon Pathom 73000, Thailand. E-mail: [email protected],

Development of the Typical Meteorological Data for Silpakorn University Sanamchandra Palace Campus Zone

Jarut Kunanoppadol1


AbstractA meteorological database is very important for engineering research and design. Normally, the more specifi c location design data always give more accuracy in the calculation result. Therefore, this study aims to generate the typical meteorological data for Silpakorn University Sanamchandra Palace Campus zone by the weather data from Nakhon Pathom Meteorological Station (located at latitude 14° 01' 42.5" N, longitude 99° 58' 12.1" E, elevation 7.46 m). We used the method of Sandia Nation Laboratories to generate the typical meteorological year and investigated the typical weather condition. We found that the outdoor design conditions for the zone are found to be 33.07°C, 94.20 %, 2.56 mm, 6.54 hrs, and 3.31 km/hr for maximum dry bulb temperature, maximum relative humidity, rainfall, sunshine duration, and wind speed respectively. We hope that these results for the zone will give advantages for future research, engineering design, business, and other related applications.

Keywords: Typical meteorological data, Finkelstein-Schafer statistics, Outdoor design condition.

IntroductionA meteorological database is very important for many fi elds of research and design. In general, we do not use the long-term average for engineering calculation but the weather design conditions are normally used (1). A representative database for year duration is known as a

typical meteorological year (TMY), mainly used in USA (2), a test reference year (TRY) or a design reference year

(DRY), mainly used in Europe (3), or other various types of typical weather design condition (4, 5). A typical outdoor

condition consists of individual months of meteorological data sets selected from different years over the available data period. The typical meteorological data have been

considered in many fi elds of research. Some previous studies have been investigated in a design meteorological day (6, 7), outdoor design condition (8), solar heat gain (9), comfortable environment (10), and acceptable wind speed (11) for air-condition and ventilation system design

process (12). Some research have been considered in building simulation (13), such as a fl ow around building (14), an indoor natural ventilation (15), and the effect of

wind to building structure (16). There are some previous studies in wind engineering (17, 18) such as optimizationfor wind system, and other various meteorological conditions(19-21). A typical weather conditions are also benefi cial for research in seasonal parameter for electricity demand analysis (22, 23), and agricultural consideration (24). Typical meteorological data have been used in solarapplication research also (12). Typical meteorological conditions have been generated in various levels upon the scope and objective of the research study; for a country TMY, such as China (25), Hong Kong (4), and Thailand (26); for a city TMY,

such as Athens (Greece) (1), Nicosia (Cyprus) (27), and Damascus (Syria) (2); and for a zone TMY, such as Port Harcourt zone in Nigeria also (28). Basically, the more specifi c location design data always give more accuracy in the calculation result. Therefore, this study aims to gener-

ate the typical meteorological data for Silpakorn University Sanamchandra Palace Campus Zone in Nakhon Pathom, Thailand. We hope that our results will give advantages for future research, engineering design, business, and

Development of the Typical Meteorological Data for Silpakorn University

Sanamchandra Palace Campus Zone15Vol 32, No1, January-February 2013

other related applications. The remaining of the paper is organized as follows; fi rst, we explain the methods used to generate a typical meteorological data; second, we present the results and discussion; and fi nally, this article concludes with discussion.

MethodsIn our study, the meteorological data used to generate the TMY is the secondary data from Nakhon Pathom Meteorological Station with the following geographical data; latitude 14° 01' 42.5" N, longitude 99° 58' 12.1" E, elevation 7.46 m, and 31.4 km from Silpakorn University Sanamchandra Palace Campus as shown in fi gure 1 (29).

The data were monthly summary reports covering the period of 13 years (1998-2010). The long-term average data are as follows; maximum and minimum dry bulb temperature (°C), maximum and minimum relative humidity(%), rainfall (mm), evaporation (mm), cloud (%), sunshine duration (hrs), and wind speed (km/hr) (30). In 1978, Hall et al. developed a TMY method that is one of the most commonly accepted methods to generate typical weather years (31, 32). The method of Sandia Nation Laboratories to generate the typical weather condition consisted of two steps; the fi rst step was to select fi ve candidate years, and the second step was to select the typical meteorology month (TMM) from the fi ve candidate years (33).

Figure 1: Location of (A) Silpakorn University Sanamchandra

Palace Campus and (B) Nakhon Pathom Meteorological

Station

Selection of Five Candidate Years Primarily, the raw meteorological data have to be rearranged for each weather data into calendar months along the thirteen-year periods. The maximum dry bulb temperature data were shown as a sample in table A1. Then, the cumulative distribution function for long-term of all period years (CDF

m) were calculated. The CDF for the

variable x was estimated by Sn(x) as shown in equation

1. (31)

Where ix is the ith order observation (from smallest to largest) and N is the number of observation

on the variable. From defi nition, ( )nS x is a monotonically

increasing step function with step of size 1/N occurring at

ix and is bounded by zero and one (2, 31). The CDFm

of maximum dry bulb temperature is plotted for sample

Kunanoppadol et al. J Sci Technol MSU16

in fi gure A1. Equation 1 was also used to calculate the cumulative distribution function for each year (CDF

y,m).

For each of the twelve calendar month, the procedure involved selecting the fi ve years that were ‘closest’ to the composite of all thirteen years. This was done by comparing the CDF

m with CDF

y,m for each of nine

parameters. The statistic selected to measure the closeness of each year’s CDF to the long-term composite for a given index was the Finkelstein-Schafer (FS) statistic, following an equation 2 (2, 26). The FS of the maximum dry bulb temperature is shown as a sample in table A2.

Then, we calculated the weighted sum (WS) of the nine FS statistics as shown in equation 3 (28, 31). Refer to Hall et al. (1978), choosing the weight factor has not been clear-cut issue but would depend on the ultimate application of the generated typical year. In previous studies, they focused on solar applications, so they weighted solar radiation more than other weather parameters (2, 4, 26, 28). In this study, we would like to develop the typical meteorological data for Silpakorn University Sanamchandra Palace Campus Zone to generate the outdoor design

condition for general engineering applications, then, the weighting scheme used for this study is presented in

table 1.

The weighted sum of weather variables is shown in table A3. Next, the fi ve years with smallest values of WS were highlighted in table and selected as candidate years for the month.

Final Selection of TMM The fi nal selection of the TMM from fi ve candidate years involved examining statistics of the persistence structure associated with monthly values of fi ve weather parameters that were deemed most important in this research; maximum dry bulb temperature, maximum relative humidity, rainfall, sunshine duration, and wind speed (32). We calculated the root mean square difference (RMSD) of the fi ve using meteorological variables as shown in equation 4 (2, 26).

where n is the number of meteorological param-eter, d

i is the difference between the monthly values with

respect to the long-term average of each meteorological parameter. The RMSD of the fi ve selected meteorological variables for the fi ve candidate years is shown in table A4. The TMMs were selected from the smallest values of RMSD of fi ve candidate years, and were highlighted in table A4 also (31).

Results and DiscussionBy applying the method described above, the TMY for Silpakorn University Sanamchandra Palace campus zone

was fi nally formed. Nine meteorological parameters were examined for a period of thirteen years. These parameters

were maximum and minimum dry bulb temperature, maximum and minimum relative humidity, rainfall, evaporation, cloud, sunshine duration, and wind speed. The selected month/year combinations from which the TMY was composed are shown in table 2, and the monthly meteorological data

obtained by TMY are shown in table 3.



Table 1 Weighting scheme for the TMYs

Tmax

Tmin

Hmax

Hmin

R E C Sd

Wspd

2/14 1/14 2/14 1/14 2/14 1/14 1/14 2/14 2/14

where Tmax

is maximum dry bulb temperature (°C), Tmin

is minimum dry bulb temperature (°C), Hmax

is maximumrelative humidity (%), H

min is minimum relative humidity (%), R is rainfall (mm), E is evaporation (mm), C is cloud (%),

Sd is sunshine duration (hrs), and W

spd is wind speed (km/hr).

Table 2 The month/year combinations for the composition of TMY

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Year 2004 2008 2005 2005 2006 2009 2005 2008 2002 2002 2004 2002

Table 3 The monthly meteorological data obtained by TMY


Tmax

31.629 31.883 34.426 36.140 33.958 33.350 33.781 33.681 32.223 32.129 32.357 31.258

Tmin

19.748 22.410 23.335 25.537 24.810 24.613 24.655 23.574 24.290 23.255 22.423 22.152

Hmax

94.677 94.483 94.613 94.467 94.484 95.267 93.613 93.710 93.033 93.871 95.100 93.065

Hmin

48.000 48.379 49.000 49.867 56.387 57.867 57.774 51.000 60.900 57.968 50.100 56.290

R 0.677 1.585 3.481 0.100 3.800 1.597 3.030 4.000 6.200 4.100 0.100 2.000

E 3.294 4.124 5.027 5.257 4.635 3.860 4.394 4.703 3.906 3.710 4.680 3.230

C 25.774 41.552 27.323 38.067 69.710 85.933 83.516 84.323 88.567 61.968 22.667 46.452

Sd

7.513 6.966 8.077 8.257 6.094 4.747 4.706 5.794 4.830 6.700 8.213 6.565

Wspd

1.955 2.469 3.035 3.103 2.261 2.837 4.200 4.332 4.267 3.455 5.370 2.448

where Tmax

is maximum dry bulb temperature (°C), Tmin

is minimum dry bulb temperature (°C), Hmax

is maximum relative humidity (%), H

min is minimum relative humidity (%), R is rainfall (mm), E is evaporation (mm), C is cloud (%), S

d is

sunshine duration (hrs), and Wspd

is wind speed (km/hr).


The fi ve meteorological annual variations for the selected TMY and for the long-term average are shown in fi gure 2 to 6 (26).

Figure 2 Annual variation of maximum dry bulb temperature for

the selected TMY, and long-term average and standard

deviation

Figure 3 Annual variation of maximum relative humidity for the

selected TMY, and long-term average and standard

deviation

Figure 4 Annual variation of rainfall for the selected TMY, and

long-term average and standard deviation

Figure 5 Annual variation of sunshine duration for the selected

TMY, and long-term average and standard deviation

Figure 6 Annual variation of wind speed for the selected TMY,

and long-term average and standard deviation



From fi gures, we found the variance of weather data; therefore, the typical meteorological data have been widely used in engineering design and calculation (1). The outdoor design conditions for Silpakorn University Sanamchandra Palace Campus zone are shown in table 4.

Table 4 The outdoor design conditions

Dry bulb temperature

- maximum 33.07 ± 1.41 °C

- minimum 23.40 ± 1.57 °C

Relative humidity

- maximum 94.20 ± 0.73 %

- minimum 53.63 ± 4.63 %

Rainfall 2.56 ± 1.86 mm

Evaporation 4.23 ± 0.65 mm

Cloud 56.32 ± 25.60 %

Sunshine duration 6.54 ± 1.33 hrs

Wind speed 3.31 ± 1.03 km/hr

Conclusion To develop the typical meteorological data for Silpakorn University Sanamchandra Palace Campus

Zone, we used the weather data from Nakhon Pathom Meteorological Station (located at latitude 14° 01' 42.5"

N, longitude 99° 58' 12.1" E, elevation 7.46 m, and 31.4 km from Silpakorn University) (29, 30). We used the method of Sandia Nation Laboratories consisted two steps to generate the typical weather condition (31, 33).

First, we rearranged the raw data into twelve calendar

months along the thirteen-year periods from 1998 to 2010.

The cumulative distribution function for long-term and for each year were calculated and compared together by the Finkelstein-Schafer (FS) statistic (2, 26). Then, we calculated the weighted sum (WS) of the nine FS sta-tistics by our weighting scheme (31) and the fi ve years with smallest values of WS were selected as candidate years for the month. Second, we calculated the root mean square difference (RMSD) of the fi ve using meteorologi-cal parameters that were deemed most important in this research; maximum dry bulb temperature, maximum relative humidity, rainfall, sunshine duration, and wind speed (32). The typical meteorological months (TMMs) were selected from the smallest values of RMSD of fi ve candidate years (31). The monthly meteorological data obtained by TMY are shown in table 3. Finally, the outdoor design conditions for the zone are found to be 33.07°C, 94.20 %, 2.56 mm, 6.54 hrs, and 3.31 km/hr for maximum dry bulb temperature, maximum relative humidity, rainfall, sunshine duration, and wind speed. We hope that these results for the zone will give advantages for future research, engineering design, business, and other related applications.

Acknowledgement We would like to thank the reviewer and the

editor of STISWB IV for their helpful comment on this article. We also would like to thank the Nakhon Pathom

Meteorological Station, Thailand for benefi cial raw data for research study, and the Department of Mechanical

Engineering, Faculty of Engineering and Industrial Technology, Silpakorn University, Thailand for supporting our research facilities.


AppendixTable A1 The maximum dry bulb temperature data (°C) (30)


1998 33.381 35.454 36.548 37.467 36.971 35.643 34.490 33.955 32.887 32.039 30.400 29.655

1999 30.784 32.196 35.584 34.323 33.500 33.247 33.626 33.123 33.383 31.323 30.323 26.726

2000 31.571 32.559 34.600 34.207 34.490 33.543 33.632 34.123 32.967 32.513 30.060 30.784

2001 32.487 33.707 32.594 36.973 34.323 33.767 33.706 33.042 34.070 31.629 29.277 29.877

2002 30.761 33.622 34.339 36.663 33.868 34.257 33.687 33.055 32.223 32.129 30.780 31.258

2003 30.461 33.711 34.297 36.753 35.152 34.107 32.726 33.577 32.873 31.703 32.513 29.635

2004 31.629 32.562 35.424 37.297 34.310 34.113 34.684 33.668 32.680 32.429 32.357 30.613

2005 31.555 34.825 34.426 36.140 36.106 34.740 33.781 33.735 32.890 31.765 31.060 28.603

2006 31.806 33.196 35.210 35.927 33.958 33.887 33.526 32.777 33.353 32.332 32.800 29.910

2007 30.555 33.282 36.203 35.937 33.513 34.877 33.335 32.877 33.247 31.474 29.560 31.819

2008 31.435 31.883 34.484 35.297 33.939 33.987 33.400 33.681 32.763 32.423 29.857 28.765

2009 29.319 34.539 35.474 35.913 33.945 33.350 32.865 34.374 33.877 32.632 31.533 31.548

2010 31.948 35.357 35.597 37.337 37.116 35.663 34.229 33.345 33.827 31.487 30.640 30.703

Table A2 The FS of the maximum dry bulb temperatureJan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

1998 0.205 0.321 0.231 0.000 0.167 0.263 0.327 0.288 0.199 0.135 0.038 0.042

1999 0.006 0.019 0.125 0.045 0.109 0.006 0.154 0.064 0.038 0.051 0.032 0.000

2000 0.077 0.038 0.202 0.000 0.013 0.077 0.019 0.077 0.045 0.109 0.067 0.096

2001 0.103 0.051 0.051 0.035 0.051 0.013 0.128 0.051 0.026 0.090 0.016 0.013

2002 0.119 0.090 0.038 0.054 0.006 0.013 0.038 0.058 0.013 0.058 0.083 0.026

2003 0.045 0.058 0.019 0.048 0.038 0.064 0.071 0.083 0.019 0.096 0.115 0.035

2004 0.167 0.045 0.064 0.022 0.109 0.154 0.096 0.109 0.019 0.096 0.167 0.099

2005 0.071 0.109 0.224 0.074 0.128 0.186 0.160 0.231 0.122 0.019 0.103 0.003

2006 0.192 0.071 0.045 0.093 0.038 0.019 0.013 0.167 0.032 0.160 0.090 0.061

2007 0.051 0.006 0.067 0.122 0.103 0.115 0.071 0.109 0.083 0.045 0.029 0.032

2008 0.038 0.038 0.019 0.157 0.026 0.032 0.051 0.032 0.006 0.006 0.006 0.010

2009 0.022 0.083 0.077 0.099 0.115 0.109 0.096 0.135 0.179 0.141 0.141 0.064

2010 0.045 0.308 0.256 0.016 0.179 0.186 0.256 0.186 0.250 0.051 0.106 0.071



Table A3 The WS of the meteorological data


1998 0.141 0.209 0.139 0.070 0.154 0.222 0.202 0.231 0.226 0.164 0.159 0.099

1999 0.173 0.211 0.120 0.151 0.218 0.218 0.207 0.136 0.167 0.147 0.104 0.121

2000 0.156 0.175 0.136 0.109 0.132 0.127 0.103 0.115 0.138 0.156 0.095 0.136

2001 0.121 0.120 0.125 0.127 0.141 0.074 0.142 0.065 0.122 0.122 0.098 0.119

2002 0.135 0.136 0.124 0.088 0.094 0.109 0.092 0.062 0.091 0.093 0.089 0.092

2003 0.151 0.146 0.170 0.114 0.046 0.119 0.065 0.109 0.113 0.088 0.063 0.115

2004 0.105 0.073 0.109 0.080 0.092 0.099 0.120 0.094 0.071 0.113 0.086 0.077

2005 0.095 0.093 0.095 0.089 0.053 0.093 0.093 0.100 0.097 0.052 0.103 0.084

2006 0.110 0.099 0.128 0.116 0.085 0.138 0.114 0.082 0.092 0.135 0.139 0.110

2007 0.152 0.056 0.129 0.143 0.120 0.083 0.136 0.117 0.108 0.093 0.083 0.107

2008 0.064 0.079 0.055 0.102 0.100 0.087 0.122 0.088 0.098 0.101 0.084 0.050

2009 0.038 0.084 0.079 0.081 0.061 0.067 0.062 0.077 0.076 0.081 0.104 0.056

2010 0.076 0.109 0.194 0.137 0.190 0.157 0.125 0.178 0.173 0.169 0.174 0.108

Table A4 The RMSD of the fi ve using meteorological variables


1998 2.775

1999 2.218

2000 1.812

2001 1.494 2.418

2002 2.302 1.124 1.443 1.852 0.814 2.519 1.363

2003 1.564 1.266 1.151 2.009

2004 1.602 1.126 1.851 2.562 1.503 2.392 1.614 2.486

2005 1.936 1.762 1.005 2.011 1.440 0.927 0.991 3.409 2.724 2.633

2006 1.010 1.888 3.874

2007 1.732 1.458 1.651 1.973

2008 1.825 1.118 1.863 1.805 0.748 2.082 2.612

2009 2.319 1.810 1.591 2.085 3.758 1.237 1.440 0.810 2.061 3.392 1.944

2010 1.840 2.775


Figure A1 The CDFm of the maximum dry bulb temperature

References 1. Pissimanis D, Karras G, Notaridou V, Gavra K. The

generation of a “typical meteorological year” for the city of Athens. Solar Energy. 1988;40(5):405-11.

2. Skeiker K, Ghani BA. A software tool for the creation of a typical meteorological year. Renewable Energy. 2009;34(3):544-54.

3. Festa R, Ratto CF. Proposal of a numerical procedure to select Reference Years. Solar Energy. 1993;50 (1):9-17.

4. Chan ALS, Chow TT, Fong SKF, Lin JZ. Generation of a typical meteorological year for Hong Kong. Energy Conversion and Management. 2006;47(1):87-96.

5. Petrakis M, Lykoudis S, Kassomenos P. A software tool for the creation of a typical meteorological year. Environmental Software. 1996;11(4):221-7.

6. Gugliermetti F, Bisegna F. Meteorological days for

HVAC system design in Mediterranean climate.Building and Environment. 2003;38(8):1063-74.

7. Hong T, Chou SK, Bong TY. A design day for building

load and energy estimation. Building and Environment. 1999;34(4):469-77.

8. Lam JC, Hui SCM. Outdoor design conditions for HVAC system design and energy estimation for buildings in Hong Kong. Energy and Buildings.

1995;22(1):25-43.

9. Li DHW, Lam JC. Development of solar heat gain factors database using meteorological data. Building and Environment. 2001;36(4):469-83.

10. Sasaki R, Yamada M, Uematsu Y, Saeki H. Comfort environment assessment based on bodily sensation in open air: relationship between comfort sensation and meteorological factors. Journal of Wind Engineer-ing and Industrial Aerodynamics. 2000;87(1):93-110.

11. Penwarden AD. Acceptable wind speeds in towns. Building Science. 1973;8(3):259-67.

12. Kouremenos DA, Antonopoulos KA, Domazakis ES. Solar radiation correlations for the Athens, Greece, area. Solar Energy. 1985;35(3):259-69.

13. Kalogirou SA. Generation of typical meteorological year (TMY-2) for Nicosia, Cyprus. Renewable Energy. 2003;28(15):2317-34.

14. Hirt CW, Cook JL. Calculating three-dimensional fl ows around structures and over rough terrain. Journal of Computational Physics. 1972;10(2):324-40.

15. van Hooff T, Blocken B. Coupled urban wind fl ow and indoor natural ventilation modelling on a high-reso-lution grid: A case study for the Amsterdam ArenA stadium. Environmental Modelling & Software. 2010;25(1):51-65.

16. Nicholls M, Pielke R, Meroney R. Large eddy simulation of microburst winds fl owing around a building. Journal of Wind Engineering and Industrial Aerodynamics. 1993;46–47(0):229-37.

17. Yamada T, Koike K. Downscaling mesoscale mete-orological models for computational wind engineering applications. Journal of Wind Engineering and Industrial

Aerodynamics. 2011;99(4):199-216. 18. Agüera-Pérez A, Palomares-Salas JC, González

de la Rosa JJ, Ramiro-Leo JG, Moreno-Muñoz A. Basic meteorological stations as wind data source: A mesoscalar test. Journal of Wind Engineering and

Industrial Aerodynamics. 2012(0). 19. Diaf S, Notton G, Belhamel M, Haddadi M, Louche

A. Design and techno-economical optimization for

hybrid PV/wind system under various meteorological conditions. Applied Energy. 2008;85(10):968-87.



20. Pallabazzer R. Evaluation of wind-generator potential-ity. Solar Energy. 1995;55(1):49-59.

21. Yang HX, Lu L, Burnett J. Weather data and prob-ability analysis of hybrid photovoltaic–wind power generation systems in Hong Kong. Renewable Energy. 2003;28(11):1813-24.

22. Apadula F, Bassini A, Elli A, Scapin S. Relationships between meteorological variables and monthly electricity demand. Applied Energy. 2012(0).

23. Lam JC, Tang HL, Li DHW. Seasonal variations in residential and commercial sector electricity consumption in Hong Kong. Energy. 2008;33(3):513-23.

24. Shi-jiang Z, Ai-hua S, Zhong-xue Z, Bin W. Multi-variable Linear Regression Equation for Rice Water Requirement based on Meteorological Infl uence. Procedia Engineering. 2012;28(0):516-21.

25. Zhang Q. Development of the typical meteorological database for Chinese locations. Energy and Buildings. 2006;38(11):1320-6.

26. Janjai S, Deeyai P. Comparison of methods for gener-ating typical meteorological year using meteorological data from a tropical environment. Applied Energy. 2009;86(4):528-37.

27. Petrakis M, Kambezidis HD, Lykoudis S, AdamopoulosAD, Kassomenos P, Michaelides IM, et al. Generation of a “typical meteorological year” for Nicosia, Cyprus. Renewable Energy. 1998;13(3):381-8.

28. Oko COC, Ogoloma OB. Generation of a Typical Meteorological Year for Port Harcourt Zone. Journal of Engineering Science and Technology. 2011;6(2):204-14.

29. Google Map. 2012 [cited 2012 12 June]; Available from: http://maps.google.com.

30. Nakhon Pathom Meteorological Station. Meteorologi-cal Station Information, Metadata. 2012 [cited 2012 12 June]; Available from: http://www.mob.tmd.go.th/databass/metnakhonpathom.html.

31. Hall IJ, Prairie RR, Anderson HE, Boes EC. Generationof a typical meteorological year for 26 SOLMETstation. Albuquerque, New Mexico1978.

32. Marion W, Urban K. User’s Manual for TMY2s: Typical Meteorological Years: Derived from the 1961-1990 National Solar Radiation Data Base: National Renewable Energy Laboratory; 1995.

33. Typical Meteorological Year User’s Manual: Hourly Solar Radiation-Surface Meteorological Observations: Nation Climate Center, Asheville, North Carolina; 1981.

Original

The Concept to Measure the Overall Car Performance

Jarut Kunanoppadol1


AbstractThe overall car performance investigating on-road experiments is necessary for research and development in automotive engineering. Car acceleration capability is a fi nal result depending on engine performance, transmission system design, suspension optimization, shape and dimension, aerodynamic, friction reduction technology, driving skill, and other factors. The purpose of this research is to present the concept to measure the overall car performance from accelerationcapacity. We found that this concept is possible and convenient because we can collect digital input signals from an existing electronic control unit and transfer it to additional processor to analyze and display the fi nal result in every mobile display, such as laptop, tablet, and smart phone. The method is cheaper and easier for installation and usage.

Keywords: car performance, acceleration, measurement

1 Silpakorn Automotive Research and Technology (SART), Department of Mechanical Engineering, Faculty of Engineering and Industrial

Technology, Silpakorn University, Nakhon Pathom 73000, Thailand. E-mail: [email protected],

IntroductionThe research and development in automotive engineeringhave been done for long time and are still ongoing as long as we need to take advantage from it. A number of previous research studies have focused on various topics, for example, conceptual development and shape design (1, 2), aerodynamic analysis (3, 4), engine performance improvement(5, 6), brake and suspensions optimization (7-10), and car utility system development (11, 12). There

also have been a number of research studies on emis-sions and alternative fuels (13-15), cost management

in product developing processes (16), and many others. Car performance can be defi ned by several criteria,

such as speed acceleration capacity, brake and control capabilities, etc. For this research, we mainly focus on speed acceleration capability only. There are various

implementations to increase the overall car performance, such as engine performance improvement, transmission and suspension system optimization, lubrication technol-ogy development, aerodynamic design, or driver course training, etc.

Engine performance developments involve increasing the engine outputs; power and torque, and decreasing the engine input; specifi c fuel consumption (17). The engine outputs depend on many operating parameters, such as air-fuel ratio, compression ratio, intake air temperature and pressure, load and engine speed, ignition timing (for spark ignition engine), injection parameter and swirling design (for compression ignition engine) (18). An engine performance map is normally used to describe the effect of operating parameters related to the engine outputs (19). However, the simple way to present correlation between engine power, torque and operating speed is normally shown by engine performance curve (sometimes the specifi c fuel consumption is also shown) (20). Although it

is very useful, the engine performance curve is not usually shown in car specifi cation. Commercially, the engine specifi cation is detailed to consumers only the maximum power, maximum torque, and engine speed at these points. In automotive engineering analysis, there are two

ways to get this curve; fi rst, measured by dynamometer, or second, simulated by calculation (21-23).

The Concept to Measure the Overall Car Performance 25Vol 32, No1, January-February 2013

Car performance is a fi nal result depending on en-gine outputs, transmission selection, tire size, aerodynamic effect, rolling friction, and other factors. Twenty percentages(20%) of indicated power from combustion is sent through transmission and tire system to drive the car forward or so-called driving force, while air resistance and rolling resistance against car motion in the opposite direction (24). So, the car is driven forward with one acceleration value by the net force following the Newton’s second law of motion. Currently, car performance is measured with many types of dynamometer in a laboratory experiment, and it is costly. Our previous research studies have focused on an engine performance development by using offset piston to improve the engine power (5),and a combined turbocharger set to increase a thermal effi ciency (6), and now we are in the process of installing a dynamometer for our experiment. We also have an idea to develop a method for measuring the overall car performance for on-road experiment (25-27). The main objective of this research is to develop and present a concept to measure the overall car perform-ance for the on-road real-time experiment and describe our conceptual framework for future implementation. The remaining of the paper is organized as follows; fi rst, we explain the information of a dynamometer; second, we present the theoretical car performance calculation method; third, the simulation results are shown; fourth, we present the conceptual implementation framework; and fi nally, this article concludes with the discussion.

Dynamometer The laboratory experimental tool to measure the output performance of an engine or a vehicle is a

dynamometer. It can be classifi ed into various typesdepending on the criteria used for consideration. By instal-lation, we separate dynamometers into two types; fi rst, the

engine dynamometer that directly connects an engine to a dynamometer; and second, the chassis dynamometer that can experiment by driving a car on the roller without taking the engine off. Both of them are used to measure and present the output power and torque of the engine

at an operating speed (20). Moreover, we can classify dynamometers by a power transfer method and also split it into two types; the absorption dynamometer, and the transmission dynamometer (28). For the absorption type, dynamometers measure and absorb the engine output power to which they are coupled. The power absorbed is usually dissipated as heat by some means, such as prony brake, rope brake, mechanic or hydraulic friction, eddy-current dynamometer. For the transmission type, the power is transmitted to the load coupled to the engine after it is indicated on some types of scale. These are also called torque-meter. (28) Inertia dynamometer is also included in the trans-mission type. The rolling mass (called drum) is designed to have enough inertia, directly connected to the engine, and loaded of the engine. Then, the engine is run and accelerated from low to maximum speed and measured the angular acceleration and angular velocity of the drum. Angular acceleration results are analyzed with the inertia of drum to calculate the engine torque. Angular velocity results are simply converted to the engine speed. Engine power is calculated from these data and the engine performance curve is presented. The inertia dynamometer is applied to be the chassis dynamometer as well by using the similar method. The concept of measuring the engine torque by acceleration data is applied in this research because it is convenient to install and measure it in a car. However, in the measuring process, the car is driven in maximum acceleration to let the engine work in full load. Therefore, to avoid an accident, the experiment should be done in the safety area such as test drive area, or raceway

only. For future application, we will apply this research to design the equipment and install it in our race car called Formula SAE and measure the overall car performance.

Car Performance To perform the car performance curve, we have to know the engine torque data at every operating speed. These data are informed by the engine performance

curve. But if we do not have the engine curve, calculated simulation is needed (21, 23). We can calculate the output torque and power from the engine, and then simulate the

Jarut Kunanoppadol J Sci Technol MSU26

engine performance curve from the details of car speci-fi cation; maximum power, maximum torque, and engine speed at these points as shown in equation 1. (22)

Then, we use the engine torque and engine speed data to calculate with the transmission system and tire data to fi nd the driving force and car velocity as shown in equation 2. (22)

Driving forces at each speed have to be reduced by resistances that is summarized from air resistance and rolling resistance. Air resistance is related to car square of velocity value, cross-section area, and drag coeffi cient

of the car. Rolling resistance depends on the weight and rolling coeffi cient. The total resistance can be calculated

as shown in equation 3. (22)

After reducing the driving force by total resistance,

we have the net force data. Car acceleration performance can be calculated from the net force and equivalence mass that is depended on gear position. The car

acceleration can be calculated as shown in equation 4. (22)

Finally, the overall car performance curve is represented by accelerate capability curve that presented correlation between accelerate performance related to the engine speed or car velocity.

Simulation Results and Discussions For a better understanding about the concept to measure the overall car performance by the accelerate capability, we presented a case study simulated from specifi cation data of Ford car; model Fiesta 5Dr 1.4L Style AT as shown in table 1 (29).

Table 1 Car specifi cation data (29)

Dimensions & Weight

Overall Width (mm.) 1,722

Overall Height (mm.) 1,496

Weight (kg.) 1,127

Engine

Maximum Power (kW/rpm) 70/5,750

Maximum Torque (Nm/rpm) 126/4,200

Transmission

Gear Ratio 1st Gear 2.816

Gear Ratio 2nd Gear 1.498

Gear Ratio 3rd Gear 1.000

Gear Ratio 4th Gear 0.726

Final Gear Ratio 4.203

Tire Size 185/55 R15

Base on engine specifi cation, we calculated output torque at engine speed from 600 to 7,200 rpm and set

the speed range as 600 rpm. Transmission effi ciency was assumed as 90% in calculating process. Simulated engine performance curve was shown in Figure 1.


Figure 1 Simulated engine performance curve

For the engine speed lower than 4,200 rpm, the engine output torque correlates with engine speed positively. The maximum torque is equal to 126 Nm at 4,200 rpm as shown in specifi cation and decreases when the engine speed is over 4,200 rpm. However, this engine performance curve is not the exact data because it is calculated by mathematical simulation. It is always better if we have the information from the real performance curve. Then, we used engine torque and operating speed results, with the tire radius of 292.25 mm. to calculate the driving force and car velocity. Total resistance was also analyzed by using assumption parameter by the following values (22); 0.80 for shape factor, 0.023 for air resistance coeffi cient, and 0.015 for rolling resist-ance coeffi cient. After simulating, we performed the caraccelerate capability performance as contour plot between car acceleration (m/s2) and engine speed (rpm) at each gear position as shown in Figure 2. From the fi gure, the areas under the curve line for each gear position were acceleration that the car can move at each gear position and not over the limit lines.

Figure 2 Simulated car performance curve

The overall car performance measured fromacceleration capability is a fi nal result from overall parameters,such as engine output, transmission ratio, transmission effi ciency, tire size, shape and car dimension, friction, electronic control unit, and driver skill. The concept to measure car performance from acceleration data is also feasible for an on-road experiment. Since currently, most cars have an electronic control unit (ECU), this concept is convenient to track digital input signals such as engine speed and car velocity to additional processor to analyze and display the result. Moreover, we can transfer raw data to process and display object on mobile equipment, such as a notebook PC, tablet PC, or smart phone. For future research, we will apply this concept to design and develop an equipment to collect digital input signals from existing ECU, to process the data, and to display the result following the conceptual implementation framework as shown in Figure 3.

Figure 3 Conceptual implementation framework

Conclusion The overall car performance depends on various

operating factors, such as the engine performance, transmission design, suspension optimization, car dimensionand shape design, aerodynamic, friction reduction technology,and driver skill. An on-road experiment is necessary for a designer, driver, tuner, developer, and researcher to

investigate the fi nal result (25-27). Overall, car accelerate

Jarut Kunanoppadol J Sci Technol MSU28

performance speeds up a car within the considering time. Thus, the concept to measure the overall car performance from acceleration capability is possible and convenient because we can collect digital input signals from an exist-ing electronic control unit and transfer it to an additional processor for analyzing and displaying the fi nal result in every mobile display, such as laptop, tablet, and smart phone. This concept is also cost effective and easier for installation and usage.

Acknowledgement We would like to thank the reviewer and the editor of STISWB IV for their helpful comment on this article. We also would like to thank the Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, Silpakorn University, Thailand for providing research facilities.

AppendixTable A Calculation parameters

References1. Tovey M. Intuitive and objective processes in automotive

design. Design Studies. 1992;13(1):23-41.2. Tovey M, Porter S, Newman R. Sketching, concept

development and automotive design. Design Studies. 2003;24(2):135-53.

3. Beccaria M, Buresti G, Ciampa A, Lombardi G, Gentzsch W, Paap H-G, et al. High-performanceroad-vehicle optimised aerodynamic design: Application of parallel computing to car design. Future Generation Computer Systems. 1999;15(3):323-32.

4. Shojaefard MH, Goudarzi K, Fotouhi H. Numerical Study of Airfl ow around Vehicle A-pillar Region and Windnoise Generation Prediction. American Journal of Applied Sciences. 2009;6(2):276-84.

5. Kunanoppadol J, editor. The concept of using offset piston to improve the engine power. International conference on science, technology and innovation for sustainable well-being II (STISWB II); 2010 13-14 August 2010; Quang Binh University, Viet Nam.

6. Kunanoppadon J. Thermal effi ciency of a combined turbocharger set with gasoline engine. American Journal of Engineering and Applied Sciences. 2010;3(2):342-9.

7. Petersen A, Barrett R, Morrison S. Driver-training and emergency brake performance in cars with antilock braking systems. Safety Science. 2006;44(10):905-17.

8. Ramos JC, Rivas A, Biera J, Sacramento G, Sala JA. Development of a thermal model for automotive twin-tube shock absorbers. Applied Thermal Engineering. 2005;25(11–12):1836-53.

9. Summala H, Lamble D, Laakso M. Driving experience and perception of the lead car's braking when looking at in-car targets. Accident Analysis & Prevention.

1998;30(4):401-7.10. Talib ARA, Ali A, Goudah G, Lah NAC, Golestaneh

AF. Developing a composite based elliptic spring for automotive applications. Materials & Design. 2010;31(1):475-84.


11. Gündodu Ö. Optimal seat and suspension design for a quarter car with driver model using genetic algorithms. International Journal of Industrial Ergonomics. 2007;37(4):327-32.

12. Kitada M, Asano H, Kanbara M, Akaike S. Develop-ment of automotive air-conditioning system basic performance simulator: CFD technique development. JSAE Review. 2000;21(1):91-6.

13. Friedrich R, Richter G. Performance requirements of automotive batteries for future car electrical systems. Journal of Power Sources. 1999;78(1–2):4-11.

14. Jahirul MI, Masjuki HH, Saidur R, Kalam MA, Jayed MH, Wazed MA. Comparative engine performance and emission analysis of CNG and gasoline in a retrofi tted car engine. Applied Thermal Engineering. 2010;30(14–15):2219-26.

15. Sørensen B. Assessing current vehicle performance and simulating the performance of hydrogen and hybrid cars. International Journal of Hydrogen Energy. 2007;32(10–11):1597-604.

16. Ibusuki U, Kaminski PC. Product development process with focus on value engineering and target-costing: A case study in an automotive company. International Journal of Production Economics. 2007;105(2):459-74.

17. Sprei F, Karlsson S, Holmberg J. Better performance or lower fuel consumption: Technological development in the Swedish new car fl eet 1975–2002. Transportation Research Part D: Transport and Environment. 2008;13(2) :75-85.

18. Heywood JB. Internal Combustion Engine Funda-

mentals: McGraw-Hill, Inc.; 1988.19. Ferguson CR, Kirkpatrick AT. Internal Combustion

Engine: Applied Thermosciences. 2nd ed: John Wilay & Sons, Inc.; 2001.

20. Pulkrabek WW. Engineering Fundamentals of theInternal Combustion Engine. 2nd ed: Pearson Prantice-Hall, Pearson Education, Inc.; 2004.

21. Hountalas DT. Prediction of marine diesel engine performance under fault conditions. Applied Thermal Engineering. 2000;20(18):1753-83.

22. Kunanoppadol J. Calculation for Automotive Engineering:Mechanical engineering, Faculty of engineeringand industrial technology, Silpakorn university, Thailand; 2009.

23. Yamin JAA, Dado MH. Performance simulation of a four-stroke engine with variable stroke-length and compression ratio. Applied Energy. 2004;77(4):447-63.

24. Crouse WH, Anglin DL. Automotive Engines. 8th ed: Glancoe Division of Macmillan/McGraw-Hill School Publishing Company; 1994.

25. Brundell-Freij K, Ericsson E. Infl uence of street characteristics, driver category and car performance on urban driving patterns. Transportation Research Part D: Transport and Environment. 2005;10(3):213-29.

26. Serrano L, Carreira V, Câmara R, da Silva MG.On-road performance comparison of two identical cars consuming petrodiesel and biodiesel. Fuel Processing Technology. 2012(0).

27. Tan CH, Tan KC, Tay A. Computationally effi cient behaviour based controller for real time car racing simulation. Expert Systems with Applications. 2010;37(7):4850-9.

28. Ganesan V. Internal COmbustion Engine. 2nd ed: Tata McGraw-Hill Publishing, Co., Ltd.; 2004.

29. Ford. Fiesta Specifi cation Data. Thailand: Ford Sales & Service (Thailand) Co., Ltd.; 2012 [cited 2012 May

30]; Available from: http://www.ford.co.th/en/cars/fi esta/specifi cations/spec-data.

Original

1 Building Energy Systems Laboratory, Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, Silpakorn University, Nakhon Pathom 73000, Thailand. E-mail: [email protected],

Improvement of Dehumidifi cation Effectiveness Prediction Models

Thosapon Katejanekarn1, Supamas Wichaisin


AbstractDehumidifi cation effectiveness prediction models for packed bed type liquid desiccant systems are important to assess moisture removal rate which is the key performance. Chung (1994) and Martin and Goswami (2000) have proposed models to predict the dehumidifi cation effectiveness as a function of input parameters including fl ow rate and temperature of the dehumidifi ed air; fl ow rate, temperature, and concentration of the desiccant solution; and specifi c area per volume of the packing. Both works used lithium chloride solution and triethylene glycol as the desiccants and the operating condition in terms of the ratio of the desiccant fl ow rate to the air fl ow rate (L/G ratio) was between 3.5 and 15.4. An experiment was carried out in Nakhon Pathom, Thailand. Forty percent concentration calcium chloride solution was used as the desiccant and the L/G ratio was within the range of 0.277 to 2.771. The actual moisture removal rates from the experiment were compared with those predicted from 3 cases: 1) using constant effectiveness, 2) using the effectiveness from Chung’s and Martin and Goswami’s models, and 3) using the effectiveness from Chung’s and Martin and Goswami’s models with an adjustment of the coeffi cients in the models to suit the experimental condi-tions. The results show that using constant effectiveness is possible if it is the average value over the range of actual operating conditions. For the cases of using the models from the literature, after the adjustment of the coeffi cients the root mean square error (RMSE) and the mean bias difference (MBD) when using Chung’s model were improved from 4.491 g/s and 4.376 g/s to 0.629 g/s and 0.522 g/s, respectively. For the case of using Martin and Goswami’s model, the RMSE and MBD were improved from 265.475 g/s and 265.377 g/s to 2.178 g/s and 2.072 g/s, respectively. It can be concluded that the coeffi cients in the models need an adjustment before use to predict the performance accurately and this study has proposed the suitable sets for two models from the literature for the packed bed liquid desiccant dehumidifi cation system using 40% concentration calcium chloride solution within the L/G ratio range of 0.277 to 2.771.

Keywords: dehumidifi cation system, liguid desiccant, packed bed, effectiveness prediction model, moisture removal rate prediction

IntroductionAir conditioning systems are necessary to create thermal

comfort in a tropical climate such as Thailand.1 The American Society of Heating, Refrigerating and Air Conditioning

Engineers (ASHRAE) specifi es temperature range of 20 to 27oC and relative humidity range of 30 to 60% as the condition at which human beings would feel thermally

comfortable.2 Vapor compression system is mostly used for air conditioning of which the cooling coil handles both cooling and dehumidifi cation. The cooling coil removes heat from

the air causing its temperature to reduce to the dew-point. When the cooling process continues, the moisture in the air passing through the coil will be condensed resulting

in the reduction of the humidity level as well as the temperature of the air. In general, the system cuts off

when the indoor air temperature reaches the setpoint value at the thermostat regardless of how much moisture left in the air. Excessive humidity in the space can cause uncomfortable feeling to the occupants and also the deterioration of materials or mold and mildew problems.3

For the applications that need to control tem-

Improvement of Dehumidifi cation Effectiveness Prediction Models 31Vol 32, No1, January-February 2013

perature and humidity simultaneously (e.g., clean rooms in electronic industry, laboratories in pharmaceutical industry), the air conditioning system has to cool the air well below the setpoint temperature to remove moisture to the desired level and then reheat it back to the setpoint value. This scheme, so-called overcool and reheat, requires a large air conditioning system and causes waste heat at the reheating step.4 A more energy-effi cient option is to separate humidity and temperature control by using a dehumidifi cation system to handle the latent load while the sensible load is still handled by the air conditioningsystem.5,6 The latent or dehumidifi cation load mainly comes from the ventilation air.7 Therefore, mitigating its humidity level before introducing into the conditionedspace is a proper action that will help reduce thecooling load imposed on the air conditioning system. Past researches reported that the size of the system could be reduced by half resulting in less costs and energy consumption.8,9

For the dehumidification system, moistureremoval rate is the key performance parameter. A method to predict the moisture removal rate is to use thedehumidifi cation effectiveness model. Thus, an accurate effectiveness model is in need.

This article presents a study conducted to propose suitable sets of coeffi cients for two dehumidifi -

cation effectiveness models from the literature to predict the moisture removal rate of an existing liquid desiccant

dehumidifi cation system. Comparisons of the results when applying the original and the adjusted models with the

proposed sets of coeffi cients with the actual experimental results are also presented.

Experimental Setup The schematic diagram of the liquid desiccantdehumidifi cation system in this work is shown in Figure 1.

Calcium chloride (CaCl2) solution was used as the desiccant.

The system comprises 3 main parts, i.e., dehumidifi cation, regeneration, and cooling parts.

Figure 1 Liquid Desiccant Dehumidifi cation System in This Study10

The dehumidifi er is of packed bed type. The air is drawn in from the bottom and directly contacts with the desiccant solution sprayed from the top. The dehumidifi ed air is delivered to a heat exchanger to cool down before being supplied into a conditioned space. The weaker desiccant at the outlet will be transported to the regen-erator to evaporate water out by a heat source thus the desiccant becomes stronger and ready to absorb moisture again. However, the temperature of the desiccant will be higher so it needs to be cooled down by cooling water in a heat exchanger before being transferred back to the dehumidifi er. The regenerator in this study is solar parabolic troughs without a sun tracking system. The cooling water is supplied by a 10-ton cooling tower.10

Methodology The actual experiment mentioned in this study used 40% concentration CaCl

2 solution as the desiccant.

The regenerating air fl ow rate was kept constant at 0.0028 kg/s. The dehumidifi ed air fl ow rate was varied at 0.04,

0.06, and 0.08 kg/s. The desiccant fl ow rate was varied at

0.02, 0.07, and 0.12 kg/s. The ratio of the desiccant fl ow rate to the air fl ow rate (or L/G ratio) was found to be in

the range of 0.277 to 2.77. The experiment was carried out during February and March, 2011. The operating data

were collected during 8 a.m. to 5 p.m. The dehumidifi cation effectiveness is defi ned as the ratio of the actual dehumidifying capacity to the

theoretical or maximum one which would occur when the humidity ratio of the outlet air is in equilibrium with the inlet desiccant as shown in Equation 1.

Katejanekarn et al. J Sci Technol MSU32

Where,W

a,in = humidity ratio of inlet air, kg

w/kg

da

Wa,out

= humidity ratio of outlet air, kgw/kg

da

Ws,in

= humidity ratio of air in equilibrium with inlet desic-cant condition, kg

w/kg

da.

Chung (1994)11,12 carried out a study on a liq-uid desiccant dehumidifi cation system using triethylene glycol (TEG) and lithium chloride (LiCl) solutions. The dehumidifi er was of packed bed type. The effectiveness of the equipment has been modeled as a function of fl ow rates of the fl uids, temperature of the fl uids at the inlets, size of the packing, and properties of the desiccant as expressed in Equation 2.

Where,

Ga,in

= fl ow rate of inlet air, kg/s G

s,in = fl ow rate of inlet desiccant, kg/s

Ta,in

= temperature of inlet air, kg/s T

s,in = temperature of inlet desiccant, kg/s

a = area to volume ratio of packing, m2/m3

X = fraction of the vapor pressure depression

of the desiccant solution to the vapor pressure of pure water.

Martin and Goswami (2000)13 conducted a study on a liquid desiccant dehumidifi cation system using a packed bed dehumidifi er and TEG as the desiccant.

A mathematical model to predict the dehumidifi cation effectiveness as a function of relevant input parameters was proposed as shown in Equation 3.

and the coeffi cients are C

1 = 48.3 y = -0.751

k1 = 0.396 m

1 = -1.57

k2 = 0.0331 m

2 = -0.906.

The L/G ratio in the work of Chung and Martin and Goswami was found to be in the range of 3.5 to 15.4. The actual experimental data were taken into a regression analysis to search for the new sets of coef-fi cients that would make both models suitable for the desiccant type and the experimental conditions. The adjusted Chung’s and Martin and Goswami’s models are respectively shown below.

and the coeffi cients are C

1 = 5,716.96 y = -0.4822

k1 = 1.0034 m

1 = -1.647

k2 = 3.9943 m

2 = 3.2062.

Once the dehumidifi cation effectiveness is evalu-ated, the moisture removal rate can then be calculated from Eq. 10.

The following section discusses about the com-parisons of the moisture removal rates predicted by using the dehumidifi cation effectiveness from 3 cases with the actual experimental results. The fi rst case is that when

a constant effectiveness was applied. The second case


is that when the original Chung’s and Martin and Gos-wami’s models were applied. The last case is that when the adjusted Chung’s and Martin and Goswami’s models were used.

Results and Discussion Figures 2 to 4 show the comparisons of the moisture removal rates predicted by applying a constant effectiveness (case 1) and by Chung’s model in cases 2 and 3 with the experimental results. When the dehumidifi -cation effectiveness was treated as a constant value, the predicted moisture removal rates were within the error bars of the actual values. This is due to the fact that the constant effectiveness was the average value over the actual operating conditions. The original Chung’s model gave the predicted values within the error bars of the ac-tual results only at higher air fl ow rates as it can be seen in Figures. 3 and 4. The adjusted Chung’s model yielded the best outcomes since the predicted values were close to the actual results at all operating conditions.

Figure 2 Moisture Removal Rates When Using Original and

Adjusted Chung’s Model at Air Flow Rate = 0.04 kg/s





Table 1 shows the root mean square error (RMSE) and the mean bias difference (MBD) of the mois-ture removal rates predicted by Chung’s model before and after the coeffi cient adjustment and the experimental values. If the values of RMSE and MBD are small and not greater than the measurement errors of the actual results, it is able to say that the model can predict the results accurately. From the table, it is obvious that the RMSE and MBD values were improved signifi cantly when the new set of coeffi cients for Chung’s model was applied. The average RMSE was improved from 4.491 to 0.629 g/s while the average MBD was improved from 4.376 to 0.522 g/s.


Adjusted Martin and oswami’s Model at Air Flow Rate

= 0.04 kg/s


Figure 6 Moisture Removal Rates When using Original and

Adjusted Martin and Goswami’s Model at Air Flow Rate

= 0.06 kg/s


Adjusted Martin and Goswami’s Model at Air Flow Rate

= 0.08 kg/s

Figures 5 to 7 show the comparisons of the moisture removal rates predicted by Martin and

Table 1 RMSE and MBD of Moisture Removal Rate When Applying Chung’s Model before and after the Coeffi cient Adjustment

Desiccant fl ow rate

(kg/s)

Air fl ow rate

(kg/s)

Constant effectiveness Chung (1994) before Chung (1994) after

0.04 0.06 0.08 0.04 0.06 0.08 0.04 0.06 0.08

0.02 RMSE 0.040 2.970 10.395 5.308 3.170 9.230 0.056 0.595 0.247

MBD -3.885 2,857.727 10,249.301 5,264.332 3,058.173 9,084.541 12.413 482.808 101.233

0.07 RMSE 0.486 0.159 8.338 4.924 0.003 14.801 1.153 0.264 0.167

MBD 414.151 54.027 8,338.163 4,851.600 -101.604 14,633.631 1,081.164 151.999 21.972

0.12 RMSE 3.124 2.571 1.690 2.649 0.018 0.316 2.791 0.169 0.215

MBD 3,054.597 2,429.180 1,512.782 2,579.434 -124.057 139.268 2,720.807 57.309 69.443

Table 2 RMSE and MBD of Moisture Removal Rate When Applying Martin and Goswami’s Model before and after the Coeffi cient Adjustment

Desiccant fl ow rate

(kg/s)

Air fl ow rate

(kg/s)

Constant effectiveness Martin and Goswami (2000) before Martin and Goswami (2000) after

0.04 0.06 0.08 0.04 0.06 0.08 0.04 0.06 0.08

0.02 RMSE 0.040 2.970 10.395 0.878 951.130 933.001 0.169 1.777 8.923

MBD -3.885 2,857.727 10,249.301 834.719 951,018.348 932,855.121 125.312 1,665.079 8,777.673

0.07 RMSE 0.486 0.159 8.338 6.839 107.125 1.007 0.988 1.325 0.020

MBD 414.151 54.027 8,338.163 6,767.098 107,020.057 986.344 915.440 1,213.357 -125.083

0.12 RMSE 3.124 2.571 1.690 10.298 0.800 378.202 3.867 1.206 1.330

MBD 3,054.597 2,429.180 1,512.782 10,228.048 657.562 378,024.483 3,797.686 1,094.198 1,184.164


Goswami’s model in cases 2 and 3 with the experimental results. It can be seen that the adjusted model provided the forecast moisture removal rates within the error bars of the actual values at all operating conditions. Therefore, modifying the model’s coeffi cients to suit the actual operating conditions is necessary for accurate prediction. Table 2 shows the RMSE and MBD values of the moisture removal rates predicted by the original and adjusted Martin and Goswami’s model and the experimental values. From the table, it can be seen that the RMSE and MBD values were improved dramatically when the modifi ed Martin and Goswami’s model was applied. The average RMSE was improved from 265.475 to 2.178 g/s whereas the average MBD was improved from 265.377 to 2.072 g/s. This confi rms the necessity of the adjustment of the model’s coeffi cients to suit the desiccant type and the actual operating conditions.

Conclusion This study is about the prediction of the moisture removal rate of a packed bed liquid desiccant dehumidi-fi cation system using the dehumidifi cation effectiveness models from the literature. When the effectiveness was treated as a constant value, it would be able to give satisfactory results if it is an appropriate value such as

the average value over the actual operating conditions. When the effectiveness models with modifi ed coeffi cients

to suit the desiccant type and the experimental conditions were applied, the predicted moisture removal rates were

appreciably closer to the actual results implying the importance of the coeffi cient adjustment. This study has proposed the new sets of coeffi cients for Chung’s and Martin and Goswami’s dehumidifi cation effectiveness models suitable for the packed bed dehumidifi er using

40% concentration CaCl2 solution as the desiccant (Eqs.

6 and 7). The air and the desiccant fl ow rates should be

in the range of 0.04 to 0.08 kg/s and 0.02 to 0.12 kg/s, respectively, which yields the L/G ratio range of 0.277 to 2.77.

Acknowledgement This study and the mentioned experimental setup were funded and supported by the Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, Silpakorn University, Thailand.

References1. Chirarattananon S. Building for energy effi ciency.

Bangkok, Thailand: Asian Institute of Technology and Energy Policy and Planning Offi ce; 2005.

2. ASHRAE. ANSI/ASHRAE Standard 55-2004, Thermalenvironmental conditions for human occupancy. At-lanta, Georgia, USA: American Society of Heating,Refrigerating and Air Conditioning Engineers Inc; 2004.

3. Katejanekarn T, Kumar S. Performance of a solar-regenerated liquid desiccant ventilation pre-conditioning system. J Energy and buildings 2008;40(7):1252-67.

4. Katejanekarn T. A Liquid desiccant air conditioning system for buildings in hot and humid climate. DoctoralDissertation, Energy Technology School of Environment, Resources and Development Asian Institute of Tech-nology, 2008.

5. Kanchana C, Praopramote C. A performance study of a liquid desiccant dehumidifi cation system. Bachelor’s Project, Department of Mechanical Engineering Faculty of Engineering and Industrial Technology, Silpakorn University; 2010.

6. Sinotok T, Nomchob S. A performance study of dehumidifi cation and regeneration processes of a liquid desiccant dehumidifi cation system. Bachelor’s Project, Department of Mechanical Engineering,

Faculty of Engineering and Industrial Technology, Silpakorn University; 2012.

7. Brandemuehl MJ, Katejanekarn T. Dehumidifi cation

characteristics of commercial building applications. ASHRAE Transactions 2004;10(Part 2):65-76.


8. Katejanekarn T. Performance of a liquid desiccant dehumidifi cation system using calcium chloride solution. In: Proceedings of The 25th Conference on Mechanical Engineering Network of Thailand (ME-NETT#25); 2011 Oct 19-21, Aonang Villa Resort Hotel, Krabi, Thailand; 2011. p. 231

9. Tuanpongkai K, Chainakorn B, Kanchanasawas P. Design, construction, and performance study of parabolic trough regenerators for use with liquid desiccant. Bachelor’s Project, Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, Silpakorn University; 2009.

10. Wakoi P, Tadsri P, Suklim O. A design, construction, and performance study of a liquid desiccant dehu-midifi er using calcium chloride solution. Bachelor’s Project, Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, Silpakorn University; 2009.

11. Chung TW, Ghosh TK, Hines AL. Dehumidifi cation of air by aqueous lithium chloride in a packed column. Inter J Separation Science and Technology 1993;28(1-3):533-50.

12. Chung TW. Predictions of the moisture removal effi ciencies for packed-bed dehumidifi cation systems. Inter J Gas Separation & Purifi cation 1994;8(4):265-68.

13. Martin V, Goswami DY. Effectiveness of heat and mass transfer processes in a packed bed liquid desiccant dehumidifi er/regenerator. Inter J HVAC&R Research 2000;6(1):21-39.

Original

Infl uence of Operating Parameters on the Performance of a Liquid Desiccant Ventilation Dehumidifi cation System

Thosapon Katejanekarn1, Watanyoo Panangnuwong1


AbstractDehumidifying the ventilation air before being introduced into an air conditioned space is an important issue since it is the main source of latent load. If the ventilation air is dry enough, the size and then the energy consumption of the air conditioning system in the space will be reduced signifi cantly. A liquid desiccant system is an interesting choice due to its many advantages. The key performance indices of the system are the moisture removal rate and the dehumidifi cation effectiveness. This study reports the infl uence of the main operating parameters on the performance of the packed bed type liquid desiccant dehumidifi cation system from the literature as well as a set of experiment carried out in Nakhon Pathom, Thailand. The actual experiment used 40% concentration calcium chloride solution as the desiccant. The results from the experiment were found to be corresponding to the past researches except the infl uence of the air fl ow rate which was still inconclusive. Nevertheless, it can be concluded that the performance of a liquid desiccant system will be high when the fl ow rate and the concentration of the desiccant solution are high and its temperature is low. Also, the performance will be high if the temperature of the dehumidifi ed air is low. The air fl ow rate depends on the required ventilation rate while the input humidity level of the dehumidifi ed air depends on the outside air condition.

Keywords: dehumidifi cation system, liquid desiccant, calcium chloride, packed bed, ventilation air

IntroductionIn a tropical climate like Thailand, it is inevitable to

create thermal comfort for building occupants. Cooling load inside a building comprises sensible and latent

load. The latter mostly comes from the ventilation air.Cooling load can be handled by a variety of air conditioning

systems such as a vapor compression system or an absorption cooling system. Almost all of the systems are temperature-control using a thermostat. When the

inside temperature reaches the setpoint, the system will stop regardless of any excessive humidity level in the space which could cause uncomfortable feeling. One way to resolve this problem is to apply the overcool and reheat scheme. The setpoint temperature will be reduced

to increase the condensation thus the removed moisture until the desired humidity level is attained then an artifi cial heat source is employed to bring the temperature

List of SymbolsSymbolsG = air fl ow rate, kg/sh = enthalpy of air, kJ/kgdam = moisture removal rate, kg/sT = temperature, oCW = humidity ratio of air, kgw/kgdaα = dehumidifi cation effectivenessβ = temperature effectivenessε = enthalpy effectivenessγο = critical surface tension of packing material, N/m

γσ = surface tension of liquid desiccant, N/m

ξ = concentration of liquid desiccant, % by mass

Subscripts a = airab = absorptionin = inletout = outlets = liquid desiccant

1 Building Energy Systems Laboratory, Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology,

Silpakorn University, Nakhon Phathom 73000, Thailand. Email: [email protected],


back to the setpoint value. This method clearly consumes a lot of energy.1-3

Another way to cope with this problem is the use of a liquid desiccant ventilation dehumidifi cation system which has been studied for decades in America and Asia. Liquid desiccant is popular due to its fl exibility and lower regeneration temperature (40-80oC compared with 80-120oC of the solid desiccant).4-5 Triethylene glycol (TEG), calcium chloride (CaCl2), lithium chloride (LiCl), and lithium bromide (LiBr) solutions are commonly used as the liquid desiccant. Confi guration of the dehumidifi er could be spray chamber, spray coil, plate heat exchanger, or packed bed type. The shape may be cubical or cylindrical. When the ventilation air comes into contact with the desiccant, heat and mass transfer takes place and both the humidity level of the air and the concentration of the desiccant are reduced. Key performance parameters of the dehumidifi cation process are moisture removal rate and dehumidifi cationeffectiveness.6 Both values depend on 2 groups of input parameters comprising physical ones (e.g., size of packing and dehumidifi er) and operating ones (e.g., fl ow rate, temperature, and humidity of the air and fl ow rate, temperature, and concentration of the desiccant). This article discusses about the infl uence of the operating parameters on the performance of the dehu-midifi cation process gathered from the literature as well as from an actual experiment. The discussion is limited to the case of packed bed dehumidifi ers using liquid desic-cants.

Dehumidifi cation Process The dehumidifi cation process mentioned in this article is the one that uses liquid desiccant to remove

moisture from the ventilation air before being introduced into a conditioned space. The air may directly contact with the desiccant in a cross-fl ow, counter-fl ow, or parallel-fl ow

manner. The process is a simultaneous heat and mass transfer process. The humidity level of the air and the concentration of the desiccant will be decreased after the process. The performance of the system is determined by the moisture removal rate and the dehumidifi cation

effectiveness.

Infl uence of Operating Parameters on Dehu-midification Process Performance from the LiteratureMoisture Removal Rate This parameter can be viewed as a direct performance indicator of the system since it tells how much moisture can be absorbed by the desiccant. It can be calculated from the difference of the humidity ratios between the inlet and outlet air and the air fl ow rate as shown in Equation 1.

It can also be calculated from the difference of the concentration between the inlet and outlet desiccant and the desiccant fl ow rate as shown in Equation 2.

Dehumidifi cation Effectiveness The dehumidifi cation effectiveness is defi ned as the ratio of the actual dehumidifying capacity to the theoretical or maximum one which would occur when the humidity ratio of the outlet air is in equilibrium with the inlet desiccant as shown in Equation 3.

Since the dehumidifi cation process is a simulta-neous heat and mass transfer process, Equation 3 alone is not suffi cient to identify the performance because it only

refers to the mass transfer capability. In order to cover the heat transfer capability, either heat transfer or enthalpy effectiveness has to be evaluated according to Equations. 4 and 5, respectively.


Chung (1994)6 carried out a study on a liquid desiccant dehumidifi cation system using TEG and LiCl solutions. The system was of packed bed type. The effectiveness of the system has been modeled as a functionof fl ow rates of the fl uids, temperature of the fl uids at the inlets, size of the packing, and properties of the desiccant as expressed in Equation 6. The model was an improved version of that proposed by Ullah et al. (1988).7

Equation 7 shows the relationship between the dehumidifi cation effectiveness and relevant input param-eters developed by Martin and Goswami (2000).8

Where,

Table 1 Infl uence of Operating Parameters on Dehumidifi cation Process Performance from the Literature

Researchers LiquidDesiccant

Confi guration Experiment/Theory

L/GRatio

Findings

Patnaik et al. (1988)9 LiBr Packed bed Experiment 0.90-1.00

Chen et al. (1989)10 LiCl Packed bed Experiment 0.857-1.321

Kavasogullari et al. (1991)11 LiCl, CaCl2 Packed bed Experiment 0.2259

Sadasivam and Balakrishnan (1991)12 LiBr Packed bed Theory 0.5-1.2

Chung et al. (1993)13 LiCl Packed bed Experiment 0.754-8.10 Moisture removal rate ↑ when:

Radhwan et al. (1993)14 CaCl2 Packed bed Theory - - air fl ow rate ↑

Oberg and Goswami (1998)15 TEG Packed bed Experiment - - air temperature ↓

Fumo and Goswami (2002)16 LiCl Packed bed Experiment 4.90-5.63 - air humidity ratio ↑

Abdul-Wahab et al. (2004)17 TEG Packed bed Experiment 2.613-11.54 - desiccant fl ow rate ↑

Chen et al. (2006)18 LiCl Packed bed Theory 4.90-5.64 - desiccant temperature ↓

Chengqin et al. (2006)19 LiBr Packed bed Theory - - desiccant concentration ↑

Liu et al. (2006)20 LiBr Packed bed Experiment -

Gommed (2007)21 LiCl Packed bed Experiment - Dehumidifi cation effectiveness ↑ when:

Katejanekarn et al. (2008)22 LiCl Packed bed Experiment 3.5–15.4 - air fl ow rate ↓

Liu et al. (2008)23 LiCl, LiBr Packed bed Theory 1.25 - air temperature ↓

Babakhani and Soleymani (2009)24 LiCl Packed bed Theory 0.236-2.359 - air humidity ratio ↑

Giovanni (2009)25 LiBr Packed bed Experiment 0.36-1.39 - desiccant fl ow rate ↑

Hassan (2009)26 CaCl2 Packed bed Experiment 1.9-2.0 - desiccant temperature ↑

Moon et al. (2009)27 CaCl2 Packed bed Experiment 1.29-1.38 - desiccant concentration ↑

Tretiak and Abdallah (2009)28 Clay-CaCl2 Packed bed Experiment -

Yin et al. (2009)29 LiCl Packed bed Experiment 0.122-0.8

Kabeel (2010)30 CaCl2 Packed bed Experiment 0.773-1.42

Zhang et.al. (2010)31 LiCl Packed bed Experiment 0.122-0.8

Ge et al. (2011)32 LiCl, CaCl2 Packed bed Experiment 4.125-5.181

Gao et al. (2012)33 LiCl Packed bed Experiment 1.25-1.875


Considering Equations 6 and 7, the important parameters indicating the operating condition of the system are the ratio of the desiccant fl ow rate to the air fl ow rate (L/G ratio or Gs/Ga) and the ratio of the inlet air temperature to the inlet desiccant temperature (Ta/Ts or ha/hs instead in Equation 7). It can be observed from both equations that the effectiveness will be high when the humidity ratio of the air and the fl ow rate and concentration of the desiccant are high, whereas the temperature of the desiccant and the fl ow rate and temperature of the air should be low. Table 1 summarizes the influence of the operating parameters on the performance of the liquid desiccant dehumidification process reviewed from the literature. It is found that the process would perform better (moisture removal rate is higher) when the air fl ow rate, desiccant fl ow rate, air humidity ratio, and desiccant concentration are higher, whereas the air and desiccant temperature should be lower. For the effective-ness, its value would be higher when the air humidity ratio, desiccant fl ow rate, desiccant temperature, and desiccant concentration are higher, while the air fl ow rate and temperature should be lower.

Infl uence of Operating Parameters on Dehumidi-fi cation Process Performance from an Actual Experiment This section discusses the influence of the operating parameters on the performance of an actual dehumidifi cation process. The dehumidifi er used in this experiment was a counter-fl ow packed bed type with a

diameter of 0.68 m and a height of 1.90 m. Outside was insulated with 0.15-m fi berglass. Inside was fi lled with 25-mm Pall ring packings that formed a 0.50-m packed bed.

Figure 1 Dehumidifi er Used in the Experiment3

The concentration of the CaCl2 solution was

controlled at approximately 40% by mass. The dehumidi-fi ed air fl ow rate was varied at 0.04, 0.06, and 0.08 kg/s. The desiccant fl ow rate was varied at 0.02, 0.07, and 0.12 kg/s. The temperature and humidity of the air were at ambient condition. The temperature of the desiccant also depended on the ambient condition because it was cooled by cooling water supplied by a cooling tower. The experiment was carried out during February and March, 2011.5 The L/G ratio of the experiment was within the range of 0.277-2.77.

Infl uence on Moisture Removal Rate Infl uence of Air Flow Rate

Figure 2 Relationship between Moisture Removal Rate and Air

Flow Rate


Figure 2 illustrates the infl uence of the air fl ow rate on the moisture removal rate. It can be seen that at the desiccant fl ow rates of 0.02 and 0.12 kg/s, the moisture removal rate increased with the increase of the air fl ow rate until a certain point then it became nearly constant or gradually decreased. However, the trend did not apply to the case of the desiccant fl ow rate of 0.07 kg/s. Therefore, it should be said that the infl uence of the air fl ow rate on the moisture removal rate was still inconclusive.

Infl uence of Desiccant Flow Rate

From Figure 4, as the system operated along the day the temperature of the air and the desiccant increased from 8 a.m. until 4 p.m. as opposed to the decreasing moisture removal rate. This suggests that the moisture removal rate increases with the decrease of the temperature of the fl uids which is corresponding to the literature.

Infl uence on Dehumidifi cation Effectiveness Infl uence of Air Flow Rate Figure 5 shows the infl uence of the air fl ow rate on the dehumidifi cation effectiveness. It can be seen that at the desiccant fl ow rates of 0.02 and 0.12 kg/s, the effectiveness increased with the increase of the air fl ow rate until a certain point then it became nearly constant or gradually decreased because of the less contact time between the air and the desiccant. Nonetheless, the trend was different for the case of the desiccant fl ow rate of 0.07 kg/s. Therefore, it should be said that the infl uence of the air fl ow rate on the effectiveness was still inconclusive. Infl uence of Desiccant Flow Rate

Figure 3 Relationship between Moisture Removal Rate and Desic-

cant Flow Rate

It can be seen from Figure 3 that the moisture

removal rate increased with the increase of the desiccant fl ow rate. This is due to the fact that the amount of the

desiccant in the system is greater thus the moisture transfer which is corresponding to the previous researches.

Infl uence of Air and Desiccant Temperature

Figure 4 Daily Moisture Removal Rate at a Fixed Desiccant Flow

Rate and Varied Air Flow Rate

Figure 5 Relationship between Dehumidifi cation Effectiveness and

Air Flow Rate

Figure 6 Relationship between Dehumidifi cation Effectiveness and

Desiccant Flow Rate


It can be seen from Figure 6 that the dehumidi-fi cation effectiveness increased with the increase of the desiccant fl ow rate since the amount of the desiccant in the system is greater which corresponds to the past researches.

Infl uence of Air and Desiccant Temperature

the desiccant fl ow rate causes better performance which agrees well with the literature. Finally, it can be concluded that a liquid desic-cant dehumidifi cation process would work best at high desiccant fl ow rate and concentration and low desiccant temperature, while the inlet temperature of the air and the desiccant should be low. The air fl ow rate would be depending on the required ventilation rate whereas the inlet air humidity would be up to the ambient condition.

Acknowledgement This study and the mentioned experimental setup were funded and supported by the Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, Silpakorn University, Thailand.

References1. Katejanekarn T, Kumar S. Performance of a solar-

regenerated liquid desiccant ventilation pre-conditioningsystem. J Energy and Building 2008;40:1252-252.

2. Katejanekarn T. A liquid desiccant air conditioning system: An application for buildings in hot and humid climates. Saarbrucken, Germany: LAP Lambert Academic Publishing GmbH & Co. KG; 2010.

3. Wakoi P, Tadsri P, Suklim O. A design, construction, and performance study of a liquid desiccant dehu-midifi er using chloride solution. Bachelor’s Project, Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, Silpakorn University; 2009.

4. Tuanpongkai K, Chainakorn B, Kanchanasawas

P. Design, construction and performance study of parabolic through regenerators for use with liquid desiccant. Bachelor’s Project, Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, Silpakorn University, Thailand;

2009.5. Kanchana C, Praopramote C. A performance study of

a liquid desiccant dehumidifi cation system. Bachelor’s

Project, Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, Silpakorn University; 2010.

Figure 7 Daily Dehumidifi cation Effectiveness at a Fixed Desiccant

Flow Rate and Varied Air Flow Rate

From Figure 7, as the system ran along the day the temperature of the air and the desiccant increased from 8 a.m. until 4 p.m. as opposed to the decreasing dehumidifi cation effectiveness. This implies that the effec-tiveness increases with the decrease of the temperature of the fl uids which is corresponding to the literature.

Conclusion and Suggestions This article reports the infl uence of the main operating parameters on the performance of the liquid desiccant dehumidifi cation process reviewed from the literature and from an actual experiment. From the literature,

the moisture removal rate increases with the increase of the air fl ow rate, desiccant fl ow rate, air humidity ratio, and desiccant concentration, but with the decrease of the air and desiccant temperature. The dehumidifi cation effective-ness increases with the air humidity ratio, desiccant fl ow

rate, desiccant temperature, and desiccant concentration, but with the decrease of the air fl ow rate and temperature. From the actual experiment, the infl uence of

the air fl ow rate on the performance was not conclusive. One assumption is that the L/G ratio may have an effect which needs further studies. Nevertheless, the increase of


6. Chung TW. Predictions of the moisture removaleffi ciencies for packed-bed dehumidifi cation systems. J Gas Separation & Purifi cation 1994;8:265-8.

7. Ullah MR, Kettleborough CF, Gandhidasan P. Effectiveness of moisture removal for an adiabatic counterfl ow packed tower absorber operating withCaCl

2-air contact system. J Solar Energy Engineering,

Transactions of the ASME 1988;98-101.8. Martin V, Goswami DY. Effectiveness of heat and

mass transfer processes in a packed bed liquid desic-cant dehumidifi er/regenerator. J HVAC&R Research 2000;6:21-39.

9. Patnaik S, Lenz TG, Lof GOG. Solar open cycle liquid desiccant studies. In: Proceedings of the 10th Annual ASME Solar Energy Conference; 1988. p.121-5.

10. Chen LC, Kuo CL, Shyu RJ. The performance of a packed bed dehumidifi er for solar liquid desiccant systems. In: Proceedings of the 11th Annual ASME Solar Energy Conference; 1989. p. 371-77.

11. Kavasogullari S, Gandhidasan P, Ertas A, Anderson E. Performance of CaCl2 and LiCl liquid desiccants in a dehumidifying packed tower. J ASME Emerging Technology 1991;36:45-9.

12. Sadasivam M, Balakrishnan AR. Analysis of thermal effects in packed bed liquid desiccant dehumidifi ers. J Chemical Engineering Process 1991;30:79-85.

13. Chung TW, Ghosh TK, Hines AL. Dehumidifi cation of air by aqueous lithium chloride in a packed column. J Separation Science and Technology 1993;28:533-50.

14. Radhwan AM, Gari HN, Elsayed MM. Parametric study of a packed bed dehumidifi er/regenerator

using CaCl2 liquid desiccant. J Renewable Energy 1993;3:49-60.

15. Oberg V, Goswami DY. Experimental study of the heat and mass transfer in a packed bed liquid des-iccant air dehumidifi er. J Solar Energy Engineering,

Transactions of the ASME, 1998;120.16. Fumo N, Goswami DY. Study of an aqueous lithium

chloride desiccant system: Air dehumidifi cation and

desiccant regeneration. J Solar Energy 2002;72:351-61.

17. Abdul-Wahab SA, Zurigat YH, Abu-Arabi MK.Prediction of moisture removal rate and dehumidifi ca-tion effectiveness for structured liquid desiccant air dehumidifi er. J Energy. 2004;29:19-34.

18. Chen XY, Li Z, Jiang Y, Qu KY. Analytical solution of adiabatic heat and mass transfer process in packed-type liquid desiccant equipment and its application. J Solar Energy 2006;80:1509-16.

19. Chengqin R, Jiang Y, Yianpin Z. Simplifi ed analysis of coupled heat and mass transfer processes in packed bed liquid desiccant-air contact system. J Solar Energy 2006;80:121-31.

20. Liu XH, Qu KY, Jiang Y. Experimental study on mass transfer performances of cross fl ow dehumidifi er using liquid desiccant. J Energy Conversion and Manage-ment 2006;47:2682-92.

21. Gommed K, Grossman G. Experiment investigation of a liquid desiccant system for solar cooling and dehumidifi cation. J Solar Energy 2007;81:131-8.

22. Katejanekarn T, Chirarattananon S, Kumar S. An experimental study of a solar-regenerated liquid desiccant ventilation pre-conditioning system. J Solar Energy 2009;83:920-33.

23. Liu HL, Jiang Y, Qu KY. Analytical solution ofcombined heat and mass transfer performance in a cross-fl ow packed bed liquid desiccant air dehumidifi er, Inter J of Heat and Mass Transfer 2008;51:17-8.

24. Babakhani D, Soleymani M. An analytical solution for air dehumidifi cation by liquid desiccant in a packed column. J International Communications in Heat and Mass Transfer 2009;36:969-77.

25. Longo GA, Gasparella A. Experimental analysis on

desiccant regeneration in a packed column with structured and random packing. J Solar Energy 2009;83:511-21.

26. Hassan MS, Hassan AAM. Performance of a proposed complete wetting surface counter fl ow channel type liquid desiccant air dehumidifi er. J Renewable Energy 2009;34:2107-16.


27. Moon CG, Bansal PK, Jain S. New mass transfer performance data of a cross-flow liquid desic-cant dehumidifi cation system, Inter J Refrigeration 2009;32:524-33.

28. Tretiak CS, Abdallah NB. Sorption and desorption characteristics of a packed bed of clay–CaCl2 desic-cant particles. J Solar Energy 2009;83:1861-70.

29. Yin Y, Zhang X, Peng D, Li X. Model validation and case study on internally cooled/heated dehumidifi er/regenerator of liquid desiccant systems. Inter J Ther-mal Sciences 2009;48:1664-71.

30. Zhang L, Hihara E, Matsuoka F, Dang C. Experimental analysis of mass transfer in adiabatic structured packing dehumidifi er /regenerator with liquid desiccant. Inter J of Heat and Mass Transfer 2010;53:2856-63.

31. Ge G, Xiao F, Niu X. Control strategies for a liquid desiccant air-conditioning system. J Energy and Buildings 2011;43:1499-507.

32. Kabeel AE. Dehumidification and humidification process of desiccant solution by air injection. J Energy 2010;35:5192-201.

33. Gao WZ, Liu JH, Cheng YP, Zheng XL. Experimental investigation on the heat and mass transfer between air and liquid desiccant in a cross-fl ow dehumidifi er. J Renewable Energy 2012;37:117-23.

Original

1 Faculty of Engineering and Industrial Technology, Silpakorn University (Sanam Chandra Palace), Nakhon Phathom 73000, Thailand.

Email: [email protected].

Thermal Performance of Solar Water Heater Integrated Vacuum-tube Collector with

Loop Thermosyphon

T. Hudakorn1, G. Boonyaaroonnate1


AbstractThis paper investigated a thermal performance of solar water heater integrated vacuum-tube collector with loopthermosyphon. Loop thermosyphon was made of copper tube with the inner diameter of 7.94 mm. The lengths of evaporator and condenser sections were 1,200 and 200 mm, respectively. Water was used as working fl uid with the fi lling ratio of 50% of the evaporator section. Loop thermosyphon was contained into vacuum tube with the internal diameter of 34 mm, while the length of vacuum tube was 1,500 mm. In the experimental setup, the tilt of solar collector was at an angle of 14 degrees above the horizontal. The fl ow rate of inlet water varied from 0.017, 0.033 to 0.067 kg/s. The experimental result showed that when the fl ow rate of inlet water increased from 0.017 to 0.067 kg/s, the heat fl ux was increased from 1,676.41 to 4,028.67 W/m2, and the effi ciency of solar collector was increased from 34.48 % to 55.72 %.

Keywords: solar water heater, loop thermosyphon, thermal performance

IntroductionNowadays the world confronts an energy crisis. Many countries cope with it by launching energy saving campaigns, encouraging people to effi ciently consume natural resources, and placing an emphasis on alternative energy research with the aim of reserving the energy

for the future. Solar energy is another alternative energy that is paid attention to as it is a huge, clean, and safe

energy resource with environmental friendliness. Thailand is located near the equator, getting the heat from the sun

throughout the year. Its average solar radiation is 18.2 MJ/m2/day1. As a result, it can be seen that Thailand has potentiality to develop solar energy. Some examples of solar energy development are solar dryer2-6, solar water

heater7-16, and electricity generation from solar energy. Earlier, solar water heater was in a fl at plate type,

which had the limitation. For example, it could not be used when the solar radiation level is low and when there is scale in its tube. Moreover, while its heat absorbing

capability is low, the heat loss is high, and its effi ciency

decreases according to its age. Later, the thermosyphon fl at-plate solar water heater has been developed10-16, solving the inherent limitation of the fl at plate solar water heater. This type of solar water heater can be used on the day the low solar radiation intensity as the vacuum tube can receive both of ultraviolet and infrared. There is no scale problem as there is heat transfer at the upper part of the tube (condenser section). This type of solar water heater provides more heat than the fl at plate type. However, it has some drawbacks including the limitation of thermosyphon that transfers the heat from the evaporator section to the condenser section. This limitation is called

fl ooding limit caused by the high axial heat fl ow leading to a high relative velocity between the counter current vapor and liquid fl ows, and consequently an increase of

the shear stresses at the vapor/liquid interface. Thus, instability of the liquid fl ow is created, which leads to entrainment of liquid. The entrained liquid is transported to the condenser by the vapor fl ow and is collected there. The high shear stresses can cause the returning condensate

Hudakorn et al. J Sci Technol MSU46

fl ow to be completely stopped. Then the condensate fl ow breaks up at the fl ooding point. In any case, the intense entrainment of fl ooding causes an insuffi cient liquid sup-ply to the evaporator. This leads to a local dryout and a complete dryout of the evaporator. The research studies the thermal performance of solar water heater integrated vacuum tube collector with loop thermosyphon. Loop thermosyphon is a heat transfer device with very high thermal conductance and no additional power input to the system. Thermosyphon is operated by heat transfer from latent heat of vaporiza-tion of working fl uid inside the tube. The working fl uid was vaporized by receiving heat from heat source and con-densed by removing the heat from heat sink. Its effi ciency is higher than that of the thermosyhon because there is no limitation of counter fl ow effect between the vapor from the evaporator section and the liquid condensation from the condenser section.

Experimental Figure 1 shows solar water heater integrated vacuum tube collector with loop thermosyphon, consisting of vacuum tubes with the external diameter of 47 mm., internal diameter of 34 mm., and the length of 1,500 mm. Loop thermosyphon was made of copper tube with the internal diameter of 7.79 mm. The evaporator section length was 1,400 mm. while its condenser section length was 200 mm. (shown in Figure 2). The working fl uid was water with the fi lling ratio of 50% of the evaporator section. The solar collector was faced south and inclined 14o up from the horizontal plane since Nakhon Pathom province

in Thailand, the location of the test set, is at 14oN latitude and 100oE longitude. The schematic diagram of experimental setup is

shown in Fig. 3. The setup was comprised of a vacuum-tube collector, a loop thermosyphon, a pump, a water tank, a fl ow meter, and a data logger (Yokogawa MW-100,

accuracy ±0.1oC. Eleven Chromel-Alumel thermocouples (Omega, type K, accuracy ±0.5oC) were connected to the

data logger in order to measure positions as follows: Te,1

-T

e,3 in the middle of the evaporator section, T

c,1-T

c,3 in the

middle of the condenser section, and Ta outside the collec-

tor to measure ambient air temperature. In addition, four

thermocouples were used to measure the temperature of the cooling water: two at the inlet and two at the outlet section of manifold header to monitor the temperature difference in order to calculate the heat transfer rate per unit area of the condenser section by using the calorifi c method, as shown in the following equation:

The experimental procedure was started byestablishing the experimental setup as shown in Fig. 3, then recording temperature at each part of the solar

collector in every minute. The water temperature differs between inlet and outlet water was used to calculate heat transfer rate per unit area of the condenser section as

indicated in equation (1). The experiments were conducted during 9.00 a.m. – 5.00 p.m. by varying the inlet water fl ow rate of 0.017, 0.033 and 0.067 kg/s.

Thermal Performance of Solar Water Heater Integrated

Vacuum-tube Collector with Loop Thermosyphon47Vol 32, No1, January-February 2013

Figrue 1 Components of solar water heater integrated vacuum

tube collector with loop thermosyphon

Figrue 2 Loop thermosyphon

Figrue 3 Experimental setup

Results and DiscussionsThe results show the thermal performance of solar water heater integrated vacuum tube collector with loop thermosyphon. The fl ow rate of inlet water was set to be 0.017, 0.033, and 0.067 kg/s. The experiment was conducted from January 15 to March 10, 2012 on the eight fl oor of the Faculty of Engineering and Industrial Technology,Silpakorn University from 9.00 am to 5.00 pm. Figure 4 shows the results on January 16, 2012. The solar radiation was unstable with the average of 652.54 W/m2, while the highest solar radiation was at 891.29 W/m2 measured at 12.50 pm. From the graph, between 9.00 am and 12.50 pm, the solar radiation tended to increase, which made the temperature of the evaporatorsection, condenser section, and outlet water increase. Later, from 12.50 pm to 5.00 pm, the solar radiation tended to decrease, which reduced the temperature of evaporator section, condenser section, and outlet water as well. The day’s average temperature of the evaporator section was 68.90 ºC, with its highest temperature measured at 2.03 pm of 79.10 ºC. The day’s average temperature at the condenser section was 51.40 ºC, with its highest temperature measured at 2.07 pm of 72.40 ºC. The day’s average temperature of the outlet water was 44.30 ºC with its highest temperature measured at 2.04 pm of 53.50 ºC.

Figrue 4 The relationship of temperature, solar radiation, and time

at the fl ow rate of 0.017 kg/s.


It could be seen that the temperature at various sectors of solar water heater integrated vacuum tube collector with loop thermosyphon increased because of the solar radiation falling on the vacuum tube. This tube absorbed the heat from the solar radiation and sent the heat to the evaporator section of loop thermosyphon contained in the vacuum tube. When the evaporatorsection of the thermosyphon received the heat, the working fl uid in the tube whose state was saturated fl uid changed to vapor and fl owed up to the condenser section located in the manifold header. The heat in the condenser section was then transferred to the water in the manifold header. After the working fl uid with vapor had exchanged the heat, it was condensed to fl uid and fl owed back to the evaporator section by the gravity force. As seen in fi gure 4, form 2.00 pm to 5.00 pm. when the solar radiation decreased, the temperature of outlet water from solar heater did not change in line with the solar radiation because the water inside the manifold header collected heat. The temperature difference between inlet and outlet water of the solar water heater in Figure 4 could be calculated for heat transfer rate per unit area of the condenser section by calorifi c method as shown in Figure 5. According to Figure 5, it was found that from 9.00 am to 12.50 pm, the heat transfer rate per unit area of the condenser section increased as the solar radiation increased. However, the heat transfer rate per unit area of condenser section kept increasing although solar radiation decreased because during this time the temperature of evaporator and condenser sections of the thermosyphon

and outlet water was stable and did not change with the decrease of the solar radiation. This was because there was the heat collecting of the water in manifold header.

Figrue 5: The relationship of heat transfer rate per unit area, solar

radiation, and time at the fl ow rate of 0.017 kg/s.

Figrue 6: The relationship between heat transfer rate per unit area of the thermosyphon’s condenser section and the water fl ow rate

Figure 6 shows the results of heat transfer rate per unit area of the condenser section at the various water

fl ow rates with the average solar radiation ranging from 554.193 to 665.908 W/m2 with the reliability of 99.9%. The graph indicates that when the inlet water fl ow rate increased, heat transfer per unit area of the condenser section of solar water heater increased. The inlet water fl ow rate increased from 0.017 to 0.067 kg/s, causing the heat transfer rate per unit area of the condenser section to

increase from 1,676.41W/m2 to 4,028.76 W/m2. Because

of the increase of the inlet water fl ow rate of solar water heater, the heat releasing at the condenser section of thermosyphon was more effi cient, which increased the heat transfer rate per unit area of the condenser section in line with the increase of inlet water fl ow rate of the solar

water heater.

Thermal Performance of Solar Water Heater Integrated

Vacuum-tube Collector with Loop Thermosyphon49Vol 32, No1, January-February 2013

From Figure 7, it was found that when the inlet water fl ow rate increased, the thermal resistance decreased. The fl ow rate increased from 0.017 to 0.067 kg/s, contributing to the decrease of thermal resistance from 1.812 to 0.085 K/W.

Figrue 7 The relationship between thermal resistance of solar water heater integrated vacuum-tube collector with loop thermosyphon and water fl ow rate.

Figrue 8 The relationship between the effi ciency of solar water

heater and water fl ow rate

From Figure 8, it was found that when the inlet water fl ow rate increased, the effi ciency increased. The fl ow rate increased from 0.017 to 0.067 kg/s, contributing to the increase of effi ciency from 34.48 to 55.72%. The increase of inlet water fl ow rate contributed to the increase

of effi ciency as the increase of the inlet water fl ow rate could cause the increase of the heat transfer rate of the solar water heater. This contributed to the higher effi ciency

of the heater.

ConclusionFrom investigating the thermal performance of solar water heater integrated vacuum-tube collector with loop thermosyphon, it was found that when the fl ow rate of inlet water increased, the heat transfer rate per unit area of the condenser section and the effi ciency of solar water heater increased, while the thermal resistance of the solar water heater decreased.

Acknowledgement The author would like to thank Associate Professor Serm Janchai, Department of Physics, Faculty of Science, Silpakorn University, for supporting the solar radiation data. Considerable appreciation is also owed to the Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, Silpakorn Univer-sity, for providing the funding of this study.

References1. Janchai, S. and Laksanaboonsong, J., Solar Potential

Map from Satellite Data for Thailand, Jirangratchata Printing, Bangkok, 1999.

2. Hossain, M.A. and Bala, B.K., Drying of hot chilli using solar tunnel drier, Solar Energy, Vol.81, pp.85-92, 2007.

3. Hudakorn, T., A Study on Performance of a V-groove Flat Plate Solar Collector for a Solar Dryer, the 23th

Conference of the Mechanical Engineering Network of Thailand, Chiang Mai, Thailand, 2009.

4. Montero, I., Blanco, J., Miranda, T., Rojas, S. and Celma, A.R., Design, construction and performance testing of a solar dryer for agroindustrial by-products, Energy Conversion and Management, Vol.51, pp. 1510-1521, 2010.

5. Banout, J., Ehl, P., Havlik, J., Lojka, B., Polesny, Z. and Verner, V., Design and performance evaluation of a Double-pass solar drier for drying of red chilli

(Capsicum annum L.), Solar Energy, Vol.85, pp.506-515, 2011.


6. Singh, S. and Kumar, S., Testing method for thermal performance based rating of various solar dryer designs, Solar Energy, Vol.86, pp.87-98, 2012.

7. Soo Too, Y.C., Morrison, G.L. and Behnia, M., Performance of solar water heaters with narrow mantle heat exchangers, Solar Energy, Vol. 83, pp.350–362, 2009.

8. Chong, K.K., Chay, K.G. and Chin, K.H., Study of a solar water heater using stationary V-trough collector, Renewable Energy Vol.39 pp.207-215, 2012.

9. Kumar, R. and Rosen, M.A., Integrated collector-storage solar water heater with extended storage unit, Applied Thermal Engineering, Vol. 31, pp.348-354, 2011.

10. Hussein, H., Transient investigation of a two phase closed thermosyphon fl at plate solar water heater, Energy Conversion and Management, Vol.43, pp.2479-2492, 2002.

11. Esen, M. and Esen, H., Experimental investigation of a two-phase closed thermosyphon solar water heater, Solar Energy, Vol.79, pp.459-468, 2005. Chen, B., Chang, Y., Lee, W. and Chen, S.,

12. Long-term thermal performance of a two-phase ther-mosyphon solar water heater, Solar Energy, Vol.83, pp.1048-1055, 2009.

13. Huang, J., Pu, S., Gao, W. and Que, Y., Experimental investigation on thermal performance of thermosy-phon fl at-plate solar water heater with a mantle heat exchanger, Energy, Vol.35, pp. 3563-3569, 2010.

14. Chien, C.C., Kung, C.K., Chang, C.C., Lee, W.S., Jwo, C.H. and Chen, S.L., Theoretical and experi-mental investigations of a two-phase thermosyphon solar water heater, Energy, Vol.36, pp.415-423, 2011.

15. Taheriana, H., Rezania, A., Sadeghi, S. and Ganji, D.d., Experimental validation of dynamic simulation of the fl at plate collector in a closed thermosyphon solar water heater, Energy Conversion and Management, Vol. 52, pp.301-307, 2011.

16. Redpath, D., Thermosyphon heat-pipe evacuated tube solar water heaters for northern maritime climates, Solar Energy, Vol.86, pp.705–715, 2012.

Original

1 Senior Scientist, 2 Assist. Prof., 3 Assist. Prof., Faculty of Engineering and Industrial Technology. Silpakorn University. Sanham chandra

palace, Nakorn Pathom 73000, Thailand.* corresponding author;Tipaporn Subsomboon, Faculty of Engineering and Industrial Technology. Silpakorn University. Sanham chandra

palace, Nakorn Pathom 73000, Thailand. [email protected].

Culture media for growth of mycelium and for induction of sporulation of Ascosphaera apis, the causative agent of chalkbrood disease of honey bee (Apis mellifera L.)

Tipaporn Subsomboon1*, Pittaya Liewseree2, Sukum Kositchaimongkol3


AbstractAn investigation of media suitable for growth and development of mycelium of the fungus, Ascosphaera apis, the causative agent of chalkbrood disease (CBD) in honey bee, Apis mellifera L, was carried out. Five culture media: Sabouraud dextrose agar + 0.2% Yeast extract (SDYEA), Malt yeast extract + 20% dextrose agar (MY 20), Jasmine brown steam rice (JBSR) and SDYEA + 5% bee hemolymph (SDYEAH) were used to demonstrate the fungal growth compared with that of culture on the standard medium, potato dextrose agar (PDA). Diameter of the mycelium colonies were compared. Results showed that among the tested media the diameters of mycelium colonies were signifi cantly different (P ≤ 0.05). The sizes of mycelium colonies on MY20 were the largest, followed by those on SDYEAH, SDYEA, PDA and JBSR, respectively. The thicknest hyphae or highest biomass was found on the SDYEAH, followed by MY20, SDYEA, PDA and JBSR, respectively. However, JBSR could prolong the hyphal longevity through 90 days without subculturing. Medium that suitable for induction of sporulation was MY20. Amount of spore produced on MY20 was more than those from malt yeast extract + 30% dextrose agar (MY30) and SDA, but size of ascocarp and ascus in MY30 was bigger than those on MY20 and SDA signifi cantly (P ≤ 0.05), however size of ascospores were signifi cantly different (P ≤ 0.05).

Keywords: Ascosphaera apis, chalkbrood, honey bee, culture media, induction of sporulation.

IntroductionChalkbrood disease is honey bee epidemic cause by A. apis. After infection, the fungal hyphae grow to cover the bee larvae which is then called mummifi ed larvae. The fungal can rapidly spread to other bee larvae, bring the bee cultivation fails. For studying the biology and pathogenicity of A. apis, the culture has to be maintained

by transfer to fresh growth media. By using standard media for fungus such as

Potato dextrose agar (PDA) and Sabouraud dextrose agar (SDA), is necessary to subculture every two weeks which is time consuming. Raffi nengo, et al1 found that integral rice kernel (IRK) prolonged the fungal hyphal longevity through 60 days without subculturing whereas

Rosalind and Buckner2 found that plant oil and bee larvae lipid increased spore germination from 50% to 70-80%.

In this study A. apis was grown on different growth media, some were modifi ed from standard media.

The ability, prolong mycelia longevity and the ability to induce mycelia develop to be ascospore of these media

were compared.

Experimental Material

Take samples (mummifi ed larvae) of A. apis from Petchabun province (A. apis, Petchabun isolate) was chosen in this research.

Subsomboon et al. J Sci Technol MSU52

Method Material for resting the ability to prolong mycelial longevity.

Potato agar (PDA) was used as standard media to compare with other 4 media: Sabouraud dextrose agar + 0.2% yeast extract (SDYEA), Malt yeast extract agar + 20% dextrose (MY20), Jasmine Brown Steam Rice (JBSR) and SDYEA +5% Bee hemolymp, H (SDYEHA). The mummifi ed larvae soak in sterile distilled water for 5 min to remove the dirt and wipe to dry then cut it to small pieces and place on the prepared media one piece on each plate. Incubate at 30±5oC for three days, 0.5 cm. diameter cork borer was used to bore at the colony edges. A bored piece was taken to place at the middle of new media plate. (Five plates for each type of media.) Incubate at 30±5oC and record the fungal colony diameter, and mycelia density. The difference between media was analyzed by using ANOVA. The media testing the ability to induce myc-elial develop to ascospore. Three types of media which different in amount of carbon source (glucose) wereinvestigated for suitable for sporulation: Malt yeast extract agar + 20% dextrose (MY20), Malt yeast extract agar + 30 % Dextrose (MY30) and Sabouraud dextrose agar (SDA)was standard media.

The 0.5 cm diameter cork borer was used to cut at colony edge to place on new media plate. Incubate at 30±5oC for more than 10 days. In each plate, cut the sporadic area three pieces sequencing from the center of the plate by using cork borer. Scrape the agar surface containing spores and put in 15 ml. centrifugal tube, fi lling 2 ml. sterile distilled water were plus 0.01% triton X-100. Count the ascocarp amount by haemocytometer. The size of ascocarp, ascus and ascospore were determined on cork borer disk by staining with lacto phenol cotton blue under light microscope. Each media were experimented fi ve replicates. The difference between media was ana-lyzed using ANOVA.

Results and Discussion The media testing the ability in prolong mycelia longevity

The test for ability in prolonging mycelia longevity was done by cutting agar at the edge of 3-day fungal colony and placed on fi ve types of media SDYEA, MY20, JBSR, PDA and SDYEHA, after incubation at 30±5oC, the colonies’ diameters were measured for 7 days. It was found that the mycelia of A. apis Petchabun isolate most rapidly grew on media MY20 with the largest colony, following by SDYEHA and SDYEA with comparably grew and PDA, JBSR respectively. (Table 1)

Table 1 Diameter of Ascosphaera apis, Petchabun isolate colonies, on fi ve types of agar in day 1 to day 7.

Media Colony diameter of Ascosphaera apis (cm.)

Day1 Day2 Da3 Day4 Day5 Day6 Day7

PDA 0.694a

±0.0431.69a

±0.2592.754b

±0.2893.858b

±0.4764.868b

±0.5335.472b

±0.5466.364b

±0.611

SDYEA 1.184c

±0.1053.292c

±0.1394.934c

±0.1736.374c

±0.1757.352c

±0.2418.082c

±0.0508.5c

± 0.000

SDYEAH 1.444d

±0.0433.318c

±0.1895.124c

±0.1146.544c

±0.2577.684c

±0.1968.5c

±0.0008.5c

±0.000

JBSR 0.860b

±0.1511.136b

±0.1091.36a

±0.0502.212a

±0.0753.104a

±0.0823.824a

±0.2835.026a

±0.842

MY20 1.464d

±0.0654.162d

±0.1696.53d

±0.3468.092d

±0.4008.336d

±0.2338.368c

±0.1978.5c

±0.000

Culture media for growth of mycelium and for induction of sporulation of Ascosphaera

apis, the causative agent of chalkbrood disease of honey bee (Apis mellifera L.)53Vol 32, No1, January-February 2013

The growth increased each day from day 1 to day 10, according to fi gure 1. After culturing for 3 and 5

Figure 1 The Ascosphaera apis, Petchabun isolate mycelia growth on 8.5 cm. plate. The colonies’ diameter’ were measured for 10 days

on JBSR and PDA. 7 days measurement on SDYEA, MY20 and SDYEHA due to full mycelial growth on plate.

days on the media SDYEHA, the mycelia was more dense and fl uffy than other media. (Table 2)

Table 2 Ascosphaera apis, Petchbool isolate mycelial fl uffi ness on fi ve type of media agar on day 3 and day 5.

MediaFungal fl uffi ness of Ascosphaera apis

Day3 Day5

Sabouraud dextrose + yeast agar (SDYEA) 2 2

Sabouraud dextrose + yeast agar + hemolymph (SDYEHA) 3 3

Malt yeast extract agar (MY 20) 2 2

Potato dextrose agar(PDA) 1 1

Jasmine brown stream rice (JBSR) 1 1

Table 3 Ascospore densities of Ascosphaera apis, Petcahbun isolate on bored agar of three types of media.

Media Amount of ascospore(spores/ml)

Sabouraud dextrose agar (SDA) 1.55 a x 106±0.716

Malt yeast extract agar (MY 20) 3.20 c x 106±0.647

Malt yeast extract agar (MY 30) 2.45 b x 106±0.647

Subsomboon et al. J Sci Technol MSU54

Table 4 Size of ascocarp, ascus and ascospore of Ascosphaera apis Petchabun isolate on three types of media.

Media

Size of ascocarp, ascus and ascospore (μm)

ascocarp ascusascospore

(wide x length)

Sabouraud dextrose agar (SDA)70.7 a±2.171(62.5 - 80)

12.8 a±1.333(7.5 -17.5)

2.2 a±0.196 x 5.1a±0.048

Malt yeast extract agar(MY 20)

82.9 b±3.273 (67.5 - 92.5)

14.1 b±1.080( 10 -17.5 )

2.4 b±0.116 x 5.3 b±0.082

Malt yeast extract agar(MY 30)

90.1 c±2.908(75 -102.5)

17.4 c±1.398( 10 - 20)

2.5 c±0.116 x 5.3 b±0.164

This may be due to the unnecessity of the mycelia for adaptation to the new media. The reason was the media had the composition of hemolymph, which was corresponded to the research of Kimbrough and Atkinson3 . The fungal Hymenoscyphus caudatus could grow and sporulated well on the imitated media composed of host tissue, however, the concentration percentage of hemolymph had no different effect, with signifi cance atP ≥ 0.05, on the growth of mycelia on solid media.

The media testing the ability to induce mycelia develop to be ascospore.

Spore density.The area spore grew was cut from the center of

the plate sequence for three pieces in one plate by using 0.5 cm. diameter cork borer. The spore area on the agar

was scraped and put in 15 ml. centrifugal tubes, fi lling with 2 ml. sterile distrilled water were plus 0.01% tritonX-100.

Ascospore amounts were counted by haemocy-

tometer under light microscope The ascospore number on media MY20 was more than MY30 and SDA; 3.20 x 106, 2.45 x 106 and 1.55 x 106 spore/ml. respectively, which differed signifi cantly with P ≤ 0.05 (Table 3). The result was corresponded to those of Wu and Youssef4. They found the media with glucose resulted in more spore formation than the media with ammonium sulfate more

than 0.25% or without sugar. However, in media MY30, the ascospore amount of A. apis Petchabun isolate was 2.45x106 spore/ml, which was less than media MY20,

but higher than media SDA. This may be due to too high carbon source (more than 20%), which unsuitable for

sporulation.

Ascocarp, ascus and ascospore sizes.The average size of ascocarp, ascus and

ascospore on three types of media were measured. Ascocarp were found to different signifi cantly at

P ≤ 0.05. However, ascus sizes were different signifi cantly at P ≤ 0.05 (Table 4). On media MY30, the ascocarp sizes were the biggest, ascus sizes were comparable and the average sizes were 90.1 μm. and 17.4 μm., respectively. On media MY20, the average sizes were 82.9 μm. and 14.1 μm., respectively. On SDA media, the average sizes were 70.7 μm. and 12.8 μm., respectively. The results were corresponded to that of Ruffi nengo, et al1. They found A. apis cultured on MY2 with glucose only 2%. Ascospore which had the fi gure of oval-long shape

had comparable which on SDA, MY20 and MY30, were different signifi cantly at P ≤ 0.05 (Table 4). On media MY30, the average size was 2.5 x 5.3 μm. on MY20 was

2.5 x 5.3 μm. and on SDA was 2.2 x 5.1 μm. thus differed from on MY20 and MY30 signifi cantly at P ≤ 0.05, but between on MY30 and MY20 were not different signifi -cantly at P ≥ 0.05 in length.

ConclusionSuitable media for complete mycerial growth

were SDYEHA, MY20 and SDYEA. The media for the

Culture media for growth of mycelium and for induction of sporulation of Ascosphaera

apis, the causative agent of chalkbrood disease of honey bee (Apis mellifera L.)55Vol 32, No1, January-February 2013

most rapid mycelia growth was MY20. The media which mycelia grew the most dense was SDYEHA. The media able to the most prolong mycelia longevity more than 90 days was JBSR without subculture. This made the media suitable for genetic maintenance.

The media suitable for sporulation induction was MY20 and the media producing the largest ascocarp and ascus was MY30.

AcknowledgmentsThis research was supported by the Research

and Development Institute of Silpakorn University and Department of Biotechnology, Faculty of Engineering and Technology, Silpakorn University, Thailand.

References1. Ruffi nengo S, Pena NI, Clement G, Palacio MA,

Escande A. Suitability of culture media for the production of ascospores and maintenance of Ascosphaera apis. J Api Res 2000;39(3-4):143-148.

2. Rosalind RJ, Buckner J. Lipids stimulate spore germination in the entomopathogenic ascomycete Ascosphaera aggregate. Mycopathologie 2004;158: 293-302.

3. Kimbrough JW, Atkinson M. Cultural features and imperfect stage of Hymenoscyphus caudatus. Amer J Bot 1972;59(2):165-171.

4. Wu XM, Youssef NN. Synchronous sporulation of Ascosphaera proliperda. Mycologia 1993;85(6):1028-1034.

Original

Biomass Pellet

Sonthi Warajanont,1 Pongsiri Jaruyanon,1 Nitipong Soponpongpipat1


AbstractThis article reviews biomass pellet by focusing on biomass and its potential as an alternative energy, biomass’s properties,biomass pellet production processes, fuel properties of biomass pellet, biomass pellet standards, and economics of biomass pellet production.

Keywords: biomass pellet, biomass pellet production processes, economic, biomass pellet standards

IntroductionNowadays the world is confronting the decrease of petroleum quantity. The amount of biomass from agricultural residues is adequate to be used as alternative energy. However, the use of raw biomass, which is the biomass without undergoing the improvement process,has some disadvantages. For example, the high moisture content in raw biomass causes instability of combustion. A problem also occurs when raw biomass is moved from its source to the storage area. With the big size and heavy weight of raw biomass as well as its different shape, length, and size, the transportation takes a lot of time, while each time of transportation, a small quantity of biomass can be moved. Moreover, there is a problem in the design of combustion chamber. However, all mentioned problems can be solved with pelletization that is a way to improve raw biomass by using enough compressive force on particles contributing to the form of pellet. Pelletization causes improved biomass to have the same size and shape, contributing to the decrease

of its volume, storage area, and a cost of transportation.This article explains the details of biomass potential as an alternative energy, its properties, production process, its

fuel properties, biomass pellet standards and economics of biomass pellet production.

Biomass Biomass is renewable materials from plants or animals. The term “Biomass” also includes agricultural

residues such as wood scrap from lumber industry; animal excrement; and residues from agricultural processing, communities, and production process in agriculture industry.27 Crude oil, natural gas, and coal that come from the deposit of dead plants and animals are not biomass. Agricultural crop includes sugarcane, cassava, corn, and oat4. Agricultural residues include ricestraw, cassava root, ear of corn, soybean trash7 grass19, 28 wheat straw28 barley straw28 stalk corn13 grape component16 sorghum trunk13 and palm olive.17 Wood and wood residues include fast growing trees and standing timber, residues from lumber factory and paper factory 27 Scotch pine6,10 Norway sprout6 molt bark12 chestnut16 eucalyptus16

wood residues from forest15 palm olive17 sawdust16. Agricultural waste from industry includes husk from

rice miller, sugar trash and sugarcane husk from sugar factory, waste from palm oil extraction,27 coffee trash16

from coffee extraction, rice bran from rice milling,8 and

bunch of palm fruit from crude palm oil extraction.21 There are also garbage and excrement from animals and other living things. Biomass is comprised of cellulose (C

6H

10O

5) and

lignin (C40H

44O

6) and can be divided into 2 types namely

woody biomass such as branch, wood scrap, sawdust, and paper pulp residue; and non-woody biomass including husk, sugarcane husk, rice straw, spike, and dry animal excrement.29 Each kind of biomass can have different components that are cellulose, hemicelluloses, lignin, flour, and protein. Trees and herbs have the main

1 Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, Silpakorn University (Sanam Chandra

Palace), Nakhon Pathom 73000, Thailand. Email:[email protected],

Biomass Pellet 57Vol 32, No1, January-February 2013

components that are cellulose, hemicelluloses, and lignin, while cereal is comprised of fl our as major component.27 Moreover, composition of each biomass was different depending on its origin and environment. By considering the excessive amount of biomass to be alternative energy, it was found that the earth could store biomass on the ground for 1,800 billion ton and in the water for 4 billion ton. Ground biomass is compared to the energy content of 33x1012 GJ that is 80 times of the world’s annual energy consumption.27 This can be shown in Figure 1. By considering the potential of available biomass, it was found that the large portion of available biomass came from forest, especially wood residues from tree cutting and timbering with the energy content of 15x109 GJ which is 36% of biomass resources27 as shown in Figure 2. It could be seen that the earth consists of biomass as alternative energy apart from petroleum and coal. For the optimum benefi t, there is a need to develop effi cient technology of biomass.

Figure 1 The amount of excessive biomass in the world.

Figure 2 Availability of excessive biomass in the world.

Analysis of Biomass Properties As mentioned, there are many kinds of biomass, and the main components of biomass are carbon,

hydrogen, oxygen, nitrogen, sulfur, and chlorine (the components of carbon, hydrogen, and oxygen account for 97-99 percent of its mass). There are two methods to analyze biomass properties namely proximate analysis and ultimate analysis. Proximate analysis method analyzes biomass components by considering on moisture (M), ash (A), volatile matter (VM), and fi xed carbon (FC).26 The property test follows the ASTM D3172 standard. The sample of fuel will be dried in the oven at the temperature of 105-110oC until its weight is constant. The lost weight compared to the origin weight of the fuel is the moisture of the sample. The dried sample is then heated in the container with the temperature of 900oC in closed container to get rid of combustible volatile matter. After that, the sample is combusted at the temperature of 750oC until the weight is stable, and it is the weight of its ash, while the lost weight is the amount of fi xed carbon. The proximate analysis can be used for fuel comparison29 and can be shown as the relationship equation below.20

FC + VM + M + A = 100 % of mass (1) Ultimate analysis method analyzes biomass properties by considering on component quantity of the biomass. The analysis focuses the quantity of nitrogen (N), carbon (C), hydrogen (H), sulfur (S), moisture (M), ash (A) and oxygen (O), and the relationship equation can be shown as20

C+H+O+N+S+M+A = 100% of mass (2)

The elemental analysis is similar to ultimate analysis, but there is minor difference. Elemental analysis

method analyzes the quantity of sodium (Na), magnesium (Mg), silicon (Si) phosphorus (P), potassium (K) calcium (Ca), cesium (Sc), chromium (Cr), manganese (Mn),

iron (Fe), copper (Cu), zinc (Zn), arsenic (As), cadmium (Cd), mercury (Hg), and lead (Pb). These components are calculated to mass percentage as well26. The data of Thai agricultural residues were shown in Table 1-3.

Warajanont et al. J Sci Technol MSU58

Production Process of Biomass Pellet Biomass pellet production process is developed from the process of the animal food industry. The process starts from putting biomass in the ring die press machine with the roller pressing the biomass through the small holes around the ring die. Later, the pressed biomass will be cut to pieces with the same length and become biomass pellet.22 Nowadays, the pelletization process requires many steps namely gathering the biomass from agricultural area and other sources, dehumidifying, biomass milling, biomass modifying before pelletizing, pelletization, cooling, screening, and pellet’s packaging.3 The pelletization process and its machine are shown in Figure 3-4 and can be explained in details as follows. Gathering the raw biomass from agricultural area and other sources.16 This step uses tractor or agricultural utility cart to store it in the form of cubic, or sometimes it is rolled and stored in the raw material storage area. In case of many kinds of biomass existing together, it is categorized and stored in each container. Each will be mixed together with the consideration on the possibility and suitability before dehumidifying. Dehumidifying means removing of the water from the biomass by controlling its moisture at 10-15 % by weigth.3 This step requires the drying machine with the temperature range of 60 - 140oC depending on drying technique, drying media, raw material, temperature, and residence time. There are many kinds of drying machines such as rotary drum dryer, fl ash dryer, and spouted bed dryer.14

Biomass milling is a way to reduce the biomass size with the milling machine whose screen size is ¼

inches or about 3 millimeters. The size of biomass after being milled must not be bigger than the pellet’s diameter. However, if the size is too small, the fuel pellet might be

contaminated when it is formed.3

Pelletization is to form the milled biomass with

pelletization machine. There are two kinds of this machine namely fl at-die and vertical mounted ring-die. Before the pelletization starts, the water must be added to stimulate

lignin in the biomass to absorb the water. After the lignin absorbs the water, the small particles of biomass will be

combined together. The compressive force between roller and pellet mill die is required to form the pellet.3

Figure 3 (a) The ring die pellet press (b) The fl at die pellet press23

Cooling is to reduce the temperature of the pelletized fuel as it has the high temperature and softness. It is necessary to reduce the temperature so that the fuel pellet will be hard and durable for the transportation. The temperature will be reduced from 90oC to 25oC. Screening is a process to separate the fuel pellet from the dust that will be returned to the producing proc-ess. The separated fuel pellet will be moved to storage bag.3Thehigh quality biomass pellet contains the dust not more than 0.5% of the pellet. Pellet’s packaging is done automatically. This process is the last stage of pelletization before the fuel

is distributed.

Figure 4 The chart of pelletization process


The parameters affecting the biomass pellet qualities From the previous study, it was found that moisture,and particle size affected the qualities of biomass pellet.22Moreover, there is a lot of research further studying the effect of compressive force, temperature, storage time, die thickness, and hammer mill screen sizeon biomass pellet properties. The moisture content of raw biomass has an infl uence on net calorifi c value, combustion effi ciency,15 bulk density, and durability index. In the case of bulk density, a higher moisture content cause a lower bulk density of biomass pellet. Moreover, durability index of biomass pellet decrease when moisture content of biomass pellet increase.13

It was found that a large particle size causes cracks and fracture in pellet.18

The hammer mill screen size has a role in speci-fying the size of the biomass after being milled. The pellet formed from the ground biomass has more mechanical durability than the pellet formed from the coarsely ground biomass.13

The die thickness refers to the thickness of the pelletization machine’s part that is an iron plate with holes like the honeycomb. The die is used for pressing the raw biomass through to become pellet. There are two kinds

of frame that are circle frame for vertical pelletization machine and ring die for radius pelletization machine.

The thicker die results in enhancing of durability, the evidence was found in pellet of wheat straw, corn stover,

and sorghum stalk.13

Compressive force is the force from the press-

ing head of pelletizing machine on the biomass. From the study, it was found that when the compressive force increased, the length and the density of the pellet were

increased.31

The die temperature for the pelletization must be

prepared before the process starts. The die is warmed beforehand at the temperature of 90oC to ensure that lignin starts melting and mix with other particles of the biomass, contributing to the ability to form the pellet.10,19,22

Storage time of raw biomass is another important

factor for pelletization. The raw biomass with long storage time causes more energy consumption to form the pellet than the fresh biomass with the short time of storage. It was also found that the pellet from the long time storage biomass had lower mechanical durability and higher bulk density than the pellet formed from the fresh biomass because the amount of fat and resin affecting the bulk density decreased.24,30

Specifi c Properties of Biomass Pellet The specifi c properties of biomass pellet are shown in Table 4.From the table, each defi ned specifi c property is important and can tell its quality as follows: The length and the diameter of the pellet is the pellet shape value which estimated amount of pellet. Moreover, the length and diameter can prevent jamming while the biomass is processed, packed, and transported. The ash content reflects the maintenance frequency of the equipment taking the biomass pellet to be combusting fuel since the ash from the combusting process is composed of potassium, chlorine, and sulfur dioxide attached to the wall of the pipe and combustion room, causing the erosion of the tools.2

The moisture content represents the amount of water inside the pellet. The moisture plays an important role in the fuel combustion. If the moisture is high, the

heat resulting from the fuel combustion will lose with the water evaporation.2 Moreover, the moisture affects the storage of the pellet. If the moisture of the pellet is high, the pellet will be easily damaged by biodegradation. For example, there might be fungus, or rotten.

Bulk density indicates the hardness and energy content of the fuel pellet. Heating value is the amount of heat which can be obtained when the pellet was combusted.

Standard of Biomass Pellet The countries that develop the pelletization are in Western Europe. These countries are located in the cold

climate region such as France, Italy, German, Sweden, and North America, so the biomass pellet is widely used for heaters in households. As a result, there is standard


of biomass pellet in each country such as PFI in U.S.A., ONORM M1735 in Australia, SS 187120 in Sweden, DIN 51731 in German, CTI-R 04/5 i n Italy, and ITEBE in France. The standard in each country is benefi cial for commercial production for exporting the biomass pellet to overseas. Table 5 summarized the specifi cation of pellet in each country.

Economics of biomass pellet production Biomass pellet production is a biomass improve-ment technology, which is a way to produce a low energy density biomass in the form of higher energy density. To invest in the biomass pellet industry, there is the necessity to consider on the costs of raw material, labor, drying process, pellet production, transportation, and other services. From the study on the cost of biomass pellet production, the production cost can be proportioned to raw material 23%, drying process 8%, labor 18%, pelletizing machine 5%, energy 12%, maintenance 6%, packaging 13%, and other services 15% as shown in Figure 5.22

Considering on the size of biomass pellet factory, a small factory requires higher cost (production capacity: more than 10 ton/hour) as a big factory has high effi ciency and capacity of the machine contributing to cost saving. In contrast, the small factory needs cheap raw material and labor cost but long term profi t.11 There is other research

showing that the suitable size of the factory depends on

the amount of raw biomass, while the production cost relates to transportation cost.4 Therefore, the best way to manufacture fuel pellet from biomass is to construct the factory where the biomass is available and the biomass pellet is demanded; also, the size of the factory must be suitable for the amount of available biomass.

Conclusion Biomass pellet is biomass or agricultural residues that undergo the processing to be small solid fuel with higher energy effi ciency. Biomass pellet is produced by milling agricultural residues such as rice straw, sawdust, and leaves of sugarcane, palm, and other plant residues before forming them to fuel pellet. When the biomass pellet and the raw biomass are compared, biomass pellet has higher energy density and bulk density. To manufac-ture the effi cient fuel pellet from biomass must consider on two things. The fi rst one is factors affecting the formation of the pellet such as compression force, temperature, stor-age time, die thickness, and hammer mill screen size, and the second one is investment. From the study, it has been found that the suitable size of the biomass pellet factory depends on the amount of biomass, while the production cost relates to the transportation cost. Therefore, there should be further research on fuel pellet from biomass for the alternative energy benefi ts.

Table 1 Proximate analysis of Thai agricultural residues26

Element Rice husk Rice straw Mize stalk Sugar cane Palm stem Palm branch

N (%) 0.37 0.60 1.30 0.80 0.40 0.76

C (%) 38.00 44.40 44.20 44.90 42.97 43.31

H (%) 4.73 5.00 5.80 5.90 5.58 5.44

S (%) 0.09 0.10 <0.01 <0.01 <0.01 <0.01

O (%) 50.20 30.80 43.50 - - -


Table 2 Ultimate analysis of Thai agricultural residues26

Element Rice husk Rice straw Mize stalk Sugar cane Palm stem Palm branch

N (%) 0.37 0.60 1.30 0.80 0.40 0.76

C (%) 38.00 44.40 44.20 44.90 42.97 43.31

H (%) 4.73 5.00 5.80 5.90 5.58 5.44

S (%) 0.09 0.10 <0.01 <0.01 <0.01 <0.01

O (%) 50.20 30.80 43.50 - - -

Figure 5 The proportion of biomass pellet production cost22

Table 3 Elemental analysis of Thai agricultural residues26

Element Rice husk(mg/kg)

Rice straw(mg/kg)

Mize stalk(mg/kg)

Sugar cane(mg/kg)

Palm stem(mg/kg)

Palm branch(mg/kg)

Na - 3,641 9.56 16.93 121.17 118.29

Mg 0.55 1.35 2.00 1.17 0.74 1.95

Si 6.66 4.30 2.68 5.30 0.73 3.01

P 43.13 54.96 111.77 45.42 29.30 81.84

K 2,395 2,799 9,611 5,141 1,481 2,952

Ca 570 1,221 1,954 1,767 2,075 3,560

Sc - 0.24 - - - -

Cr 997.33 559.42 112.26 63.49 - -

Mn 3.2.04 852.27 43.98 125.86 5.89 12.11

Fe 2,127 2,594 446 356 8.31 35.65

Cu 210.76 6.70 7.50 4.26 1.64 4.07

Zn 38.23 29.46 65.73 40.98 19.26 21.21

As 0.76 1.04 0.04 0.49 - 0.03

Cd - 0.06 - 0.01 - -

Hg 0.98 - 0.04 - 0.84 -

Pb(%) 32.89 1.55 0.45 1.88 0.27 -


Table 4 The specifi c properties of biomass pellet3

Item Properties

1. diameter of pellet 6-10 millimeters

2. length of pellet 10-30 millimeters

3. heating value 16.9-18.0 MJ/kg

4. moisture content 7-12 percent of mass

5. ash content Approximately 0.5 percent of mass

6. bulk density 650-700 kg/cubic meter

Table 5 Comparison of biomass pellet standard in each country1, 5

Country Standard Heating value(MJ/kg)

Bulk density

(kg/m3)

Highest Moisture

(% wt)

Ash amount

(% wt)

Europe CEN/TS 14961 16.9 - 10 0.7

Australia ONORM M7135 18 - 10 0.5

German DIN 51731 17.5-19.5 - 12 1.5

Sweden SS 187120 ≤16.9 ≤500 12 1.5

U.S.A PFI - 641.7 8 1

Italy CTI-R 04/5 >16.96 620-720 ≤10 ≤0.7

France ITEBE >17.02 >650 <10 <10

Acknowledgments The authors gratefully acknowledge the fi nancial support from the department of Mechanical Engineering,Faculty of Engineering and Industrial Technology, Silpakorn University.

References1. Garcia-Maraver A, Popov V, Zamorano M. A review

of ruropean standard for pellet quality. Renewable Energy 2011;36:3537-40.

2. Williams A, Jones JM, Ma L, Pourkaashanian M. Pollutants from the combustion of solid biomass

fuels. Progress in Energy and Combustion Science 2012;38:113-37.

3. Anton Leander Schuller. Developing a Wood Pellet

Fuel Sector in South Yorkshire; 2004.

4. Sultana A, Kumar A, Harifi eld D. Development of agri-pellet production cost and optimum size.

Bioresource Technology 2010;101: 5609-21.5. Brkic M, Janic T. Standards for Pelleted and Briquet-

ted Biofuels. Cont. Agr. Engng 2009;35:219-80.6. Rhen C, Ohman M, Gref R, Wasterlund I. Effect of

raw material composition in woody biomass pellets on combustion characteristics. Biomass and Bioenergy 2007;31:66-72.

7. Sheng C, Azevedo JLT. Estimating the higher heating value of biomass fuels from basic analysis data. Biomass and Bioenergy 2005;28:499-507.

8. Chou C-S, Lin S-H, Lu W-C. Preparation and characterization of solid biomass fuel made from rice straw and rice bran. Fuel Processing Technology 2009;90:980-7.


9. Chaiyote C. An Investigation on Green Fuel Production from Crop Residues Mixed with Manila Grass. Thesis M.S. in Renewable Energy, Naresuan University, Pisanulok, Thailand, 2009.

10. Bergstrom D, Israelsson S, Ohman M, Dahlqvist S-A, Gref R, Boman C, Wasterlund I. Effect of raw material particle size distribution on the characteristic of Scots pine sawdust fuel pellets. Fuel Processing Technology 2008;89:1324-9.

11. Nilsson D, Bernesson S, Hansson P-A. Pellet produc-tion from agricultural raw materiala-A systems study. Biomass and Bioenergy 2011;35:679-89.

12. Relova I, Vignote S, Leon MA, Ambrosio Y. Optimisa-tion of the manufacturing variables of sawdust pellets from the bark of Pinus caribaea Morelet: Particle size, moisture and pressure. Biomass and Bioenergy 2009;33:1351-7.

13. Theerarattananoon K, Xu F, Wilson J, Ballard R, Mckinney L, Staggenborg S, Vadlani P, Pei ZJ, Wang D. Physical properties of pellets made from sorghum stalk, corn stover, wheat straw, and big bluestem. Industrial Crops and Products 2011;33:325-32.

14. Stahl M, Granstrom K, Berghel J, Renstrom R. Indus-trial processes for biomass drying and their effects on the quality properties of wood pellets. Biomass and Bioenergy 2004;27:621-8.

15. Zamorano M, Popov V, Rodriguez ML, Garcia-Mar-aver A. A comparative study of quality properties of pelletized agricultural and forestry lopping residues. Renewable Energy 2011:1-8.

16. Gil MV, Oulego P, Pevida C, Pis JJ, Rubiera F. Me-

chanical durability and combustion characteristics of pellets from biomass blends. Biomass Technology 2010;101:8859-67.

17. Carone MT, Pantaleo A, Pellerno A. Infl uence of process parameters and biomass characteristics on

the durability of pellets from the pruning residues of Olea europaea L. Biomass and Bioenergy 2011;35: 402-10.

18. Kaliyan N, Morey RV. Factor affecting strength and durability of densifi ed biomass products. Biomass and Bioenergy 2009;33:337-359.

19. Gilbert P, Ryu C, Sharifi V, Swithenbank J. Effect of process paprameters on pelletisation of herbaceous crops. Fuel 2009;88:1491-7.

20. Nag PK. Power Plant Engineering Third Edition, Tata McGraw – Hill Publishing Company Limited, New Delhi, 2008.

21. Razuan R, Finney KN, Chen Q, Sharifi VN, Swithen-bank J. Pelletised fuel production from palm kernel cake. Fuel Processing Technology 2011;92:609-15.

22. Samson R, Duxbuy P, Drisdelle M, Lapointe C. As-sessment of pelletized biofuels, 2000.

23. Leaver RH. Wood pellet fuel and the residential market. Andritz, Inc.

24. Samuelsson R, Thyrel M, Sjostrom M, Lestander TA. Effect of biomaterial characteristics on pelletizing properties and biofuel pellet quality. Fuel Processing Technology 2009;90:1129-34.

25. Mani S, Sokhansanj S, Bi X, Turhollow A. Econom-ics of Producing Fuel Pellets from Biomass. Applied Engineering in Agriculture 2006;22(3):421-6.

26. Garivait S, Chaiyo U, Patumsawad S, Deakhuntod J. Physical and chemical properties of Thai biomass fuels from agricultural residues. The 2nd Joint In-ternational Conference on “Sustainable Energy and Environment (SEE 2006),1-23 November 2006.

27. Yokoyama S. The Asia Biomass Handbook: A Guide for biomass production and utilization, 2008.

28. Mani S, Tabil LG, Sokhansanj S. Effects of compressive force, particle size and moisture content on mechanical properties of biomass pellets from grasses. Biomass and Bioenergy 2006;30:648-54.

29. Chukgai S. Combustion, Active Print Company Lim-

ited, Bangkok, Thailand, 2004.30. Filbakk T, Skjevrak G, Hoibo O, Dibdiakava J, Jirjis

R. The infl uence of storage and drying methods for

scots pine raw material on mechanical pellet prop-erties and production parameters. Fuel Processing Technology 2011;92:871-8.

31. Stelte W, Holm JK, Sanadi AR, Barsberg S, Ahren-feldt J. Fuel pellets from biomass: The importance of

the pelletizing pressure and its dependency on the processing conditions. Fuel 2011;90:3285-90

Original

Kinetics of Biomass Decomposition in Pyrolysis and Torrefaction Process

Unchana Sae-Ueng,1 Noppong Sritrakul1, Nitipong Soponpongpiapat1


AbstractThis paper reviewed the mass loss kinetics of biomass in torrefaction process which is well-known as biomass’s properties improving process. According to consideration of torrefaction process as mild pyrolysis process, this process was also reviewed. The weight loss prediction models, which are established from decompositionmechanisms of raw biomass, were reviewed. In addition, the experimental set up both isothermal and non-isothermal conditions in nitrogen atmospheres and oxidative atmospheres were presented in detail. Finally, parameters affecting kinetics of both torrefaction and pyrolysis such as particle size of raw biomass and heating rate were explained.

Keywords: Biomass, Kinetics, Mild pyrolysis, Thermogravimetric analysis, Torrefaction

IntroductionIn recent years, the biomass is regarded as an interested renewable energy source because it is part of the neutral carbon cycle. In the world, there is a large quantity of biomass and it is environmentally friendly. However, the use of raw biomass is challenged by its poor properties such as low calorifi c value, low energy density, high moisture content, and low grindability index. These drawbacks result in many problems; for example, trans-portation, storage and application of raw biomass.1-5 These problems can be solved by improvement of raw

biomass’s properties. There are several processes for biomass properties improvement such as densifi cation,

torrefaction, pyrolysis, and carbonization process. The pyrolysis process is general name of biomass thermal

decomposition, and composes of carbonization, torrefaction and others process. The carbonization, which is known as slow pyrolysis, starts at temperature of 4000C6 while torrefaction, which is known as mild pyrolysis, starts

at temperature of 200 - 3000C.1-3, 7 It is noted that both pyrolysis and torrefaction process is commonly used in practices. The understanding of biomass decomposition of both processes is need for developing of operation procedure and equipment design.

Pyrolysis Process Generally, the pyrolysis is a thermo-chemical decomposition of biomass in inert atmosphere. Pyrolysisresults in decomposition of complex hydrocarbonmolecules biomass into simpler molecules of gas, liquid, and char.6 The biomass pyrolysis compose of four procedures: moisture evolution at temperature of lower than 2200C, hemicellulose decomposition in temperature range of 220-3150C, cellulose decomposition in range of 315-4000C, and decomposition of lignin in range of 160 - 9000C.8, 9

The reaction mechanisms of pyrolysis The study of biomass kinetics can be conducted by means of the reaction mechanism consideration. The reaction mechanism is defi ned as the sequence of chemical change. This mechanism is an important

key for prediction of the weight loss of biomass due todecomposition reaction. The kinetics model of biomass decomposition was studied base on decomposition of

hemicelluloses, cellulose, and lignin. The detail of each model can be described as follows.

1 Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, Silpakorn University

(Sanam Chandra Palace), Nakhon Pathom 73000, Thailand. Email: [email protected],

Kinetics of Biomass Decomposition in Pyrolysis and Torrefaction Process 65Vol 32, No1, January-February 2013

One-step global model

A One-step global model shows the conver-sion of biomass to char and volatile matters as shown in Figure 1. According to literature,10-12 it was found that the conversion prediction of this model was faster than that of the actual experiment, particularly at higher pyrolysis temperature regions. For one step model, the softening effect and formation of an intermediate were not taking into account.10, 11

Figure 1 One-step global model11

Two-step consecutive-reaction model There are two reaction mechanisms of this model as shown in Figure 2. From the fi gure, raw biomass decomposes into intermediate and primary volatile. Intermediate, later, decomposes into fi nally char and secondary volatile. From previous researches,10, 11, 13-15 the result shown that a two-step consecutive-reaction models provided excellent agreement between the experimental data and the conversion prediction.

Figure 2 Two-step consecutive-reaction model11

Semi-global reaction model

There are three reaction mechanisms of this model as shown in Figure 3. From the fi gure, raw biomass decomposes into intermediate (I) and volatile (I). Next, intermediate (I) decomposes into intermediate (II) and volatile (II). Finally, intermediate (II) decomposes into

fi nally char and volatile (III).16

Figure 3 Semi-global reaction model16

Experimental setup The experiment on pyrolysis kinetics can be conducted by several methods but the most popular technique is the thermogravimetric analysis (TGA).10-15,

17-19 The TGA continuously measures and records the changes in weight of solid biomass on a high precision thermobalance. The typical TG curve is shown in Figure 4 for non-isothermal experimental setup and 5 for isothermal experimental setup.20 Figure 4 represents the weight loss of biomass in percentage versus temperature. The curves can be divided into three different ranges: drying, the releasing of some light volatiles, and the releasing of other heavy components22 while Figure 5 represents the weight loss of biomass in percentage versus time. It is noted that there are difference between isothermal and non-isothermal TG curve on horizontal axis.

Figure 4 The non-isothermal TG curve for wheat straw21

Figure 5 The isothermal TG curve for wheat straw21

Sae-Ueng et al. J Sci Technol MSU66

The DTG is derivative thermogravimetry, which is conducted by TGA equipment. The typical DTG curve is shown in Figure 6. The peak of the DTG curves shows reactivity of main composition of biomass to volatile matters. Mostly, the TG and DTG technique were used together.10, 11, 14, 23

In addition, the experiment on pyrolysis can also be applied by laboratory pyrolysis reactor. As same as in TGA, the data gathering of weight loss is needed. Branca and Blasi16 studied the decomposition of beech and willow at isothermal condition by quartz reactor. The experimental condition of previous researches was shown in Table 1.

Figure 6 The DTG curve of rice husk24

1.3 Parameters affecting pyrolysis kinetics 1.3.1 Heating rate The effect of heating rate on decomposition of solid biomass in pyrolysis process was shown in Figure 7. It can be seen from the fi gure that residual weight fraction

of raw biomass decreases with increasing of heating rate. The heating rate also affects kinetics parameters. When the heating rate increases, frequency factor (A) and rate constant (k) increase but the activation energy (E) decreases.23 Moreover, the increasing of heating rate results in increasing of initial decomposition temperature.10

Figure 7 The effect of heating rate on palm shells decomposition11

Particle size of biomass The effect of particle size on decomposition rate of solid biomass was shown in Figure 8. It was seen that, the decreasing of particle size of biomass results in the decreasing of residual weight fraction. Guo and Lua,11 studied by isothermal TGA, they found that the pyrolysis temperature and the reaction rate is signifi cantly increased when particle size is small. The non-isothermal experi-ment with the lager particle size was also conducted. It was found that the mass loss rate decreased due to decreasing of released volatile and increasing of residual solid state.23 The high activation energy and diffi cult decomposition occurred with the large particle size; the decomposition was controlled by the kinetics and heat transfer mechanism. However, the less activation energy and easy decomposition occurred with small particle size; it was only controlled by kinetics mechanism.10, 11, 17

Figure 8 The effect of particle size on palm shells decomposition11


The data of pyrolysis kinetics parameters was summarized as shown in Table 2.

Torrefaction Process The torrefaction is biomass’s properties improvement technology. Torrefaction is also called mild pyrolysis.1, 3, 7, 28 It was carried out at temperature range of 200-3000C, in inert atmosphere, and low heating rate.3, 7

Reaction mechanisms of torrefaction Generally, torrefaction process carried out at low temperature, thus, almost amount of hemicelluloses was decomposed while partial amount of celluloses was decomposed. The kinetics model can be described as follows. One-step global model A one-step global model describes the de-composition of biomass and the conversion to charcoal and volatile. This model is similar to one-step model of pyrolysis. The one-step global model is a simple model and suitable for applying with low temperature torrefaction.

Chen and Kuo,22 defi ne the torrefaction at low temperature as light torrefaction process. The reaction mechanism of one-step global model is shown in Figure 9.

Figure 9 One-step global model29

Two-step consecutive-reaction model

This model4, 29 assumes that the products of fi rst step decomposition reaction of biomass are volatile

(I) and intermediate. The second step decomposition of intermediate results in volatile (II) and charcoal. The re-actant of the fi rst step can be defi ned as hemicelluloses while the intermediate of the second step can be defi ned as residual substance inside biomass.4 The reaction

mechanism of two-step consecutive-reaction model is shown in Figure 10.

Figure 10 Two-step consecutive-reaction model4

Two-step reaction in series model The decomposition reaction for this model can be divided into two groups, the fast and medium groups. The fast reaction group is defi ned as group of hemicelluloses decomposition. The medium reaction group is referred as the decomposition of cellulose and lignin. The reaction mechanism of two-step reaction in series model is shown in Figure 11.

Figure 11 Two-step reaction in series model30

Experimental setup The experimental setup of torrefaction for determination of kinetic parameters is similar to that of pyrolysis. The TGA/DTG technique is still used in thermal degradation analysis of biomass torrefaction.1, 4, 5, 22 Moreover, Pe´trissans et al.31 creates special equipment for investigating kinetic parameters. This setup measures the weight loss in percentage during operation of torrefaction process as shown in Fig. 12. From the fi gure, the reactor size of 25 x 11 x 2.5 cm3 is fi lled with Poplar of 269–280 g. List experimental design and setup of previous works were summarized as shown in Table 3.

Figure 12 Torrefaction reactor3

Parameters affecting torrefaction kinetics Peng et al.31 considered the effect of particle size on torrefaction of by TGA and tubular fi xed bed reactor.

It was found that increasing of particle size results in decreasing of reaction constant rate.30 The effect of heating

rate on kinetic parameters of torrefaction process was


not found in the literature review. The data of torrefac-tion kinetics parameter conducted by previous work was summarized as shown in Table 4.

Conclusion This paper reviewed the weight loss kinetics studies of biomass in pyrolysis and torrefaction proc-

ess. For a long time, the pyrolysis is known as biomass properties improvement process while the torrefaction is a new technology. The kinetic models of both processes are useful for determine to suitable operating conditions in industrial process. Nowadays, the information of biomass pyrolysis is suffi cient but information of various biomasses in torrefaction process is needed.

Table 1 List of experimental design and setup of pyrolysis process

Technique Atmospheres Feedstock References

Isothermal

Nitrogen

Xylan Straw , corn stalksBeech oil-shaleextracted oil palm fi bers

(Lanzetta and Blasi, 1997)13

(Blasi and Lanzetta, 1998)15

(Branca and Blasi, 2003)16

(Williams et al., 2000)19

(Guo and Lua, 2000)17

Airpine sawdustSweetener

(Bilbao et al., 1997)25

(Conceio et al., 2005)26

Non-isothermal

Nitrogen

walnut shell corn cob olive residue and sugar cane bagasse oil-palm solid wastes oil-shale cotton stalk, sugarcane bagasse, shea meal extracted oil palm fi bers

(Yuan and Liu, 2007)14

(Yu et al., 2008)15

(Ounas et al., 2011)18

(Luangkiattikhun et al., 2008)23

(Williams et al., 2000)19

(Munir et al., 2009)27

(Guo and Lua, 2000)17

AirRice husk cotton stalk, sugarcane bagasse, shea meal pine sawdust

(Mansaray and Ghaly, 1999)24

(Munir et al., 2009)27

(Bilbao et al., 1997)25

Table 2 Summary of pyrolysis kinetic parameters

Feedstock Technique Conditions Reaction mechanisms Kinetic parameter

oil-palm shell 0.3-0.5,

0.5 -1, 1-2

and 2-2.8 mm11

non-isothermal TG-DTG

nitrogen 50 cm3/min


two-step consecutive-reaction model

One-step;A=9.74 X103 s-1

E=54.8 kJ/molTwo-step;A1 = 6.85 X107 s-1

E1=110.3 kJ/molA2=9.82 X1012 s-1

E1=168.4 kJ/mol


Feedstock Technique Conditions Reaction mechanisms Kinetic parameter

corn cob 0.5, 1, and

2 mm10

non-isothermal TG-DTG

373-1073 K10, 20, and 30

K/minnitrogen 4.0

ml/min


two-stepconsecutive-reaction model

One-step;A1=6.4 X10-3 s-1

E1=27.64 kJ/molTwo-step;A1 = 1.49 X1016 s-1

E1=210.71 kJ/molA2=3.76 X1036 s-1

E2=459.8 kJ/mol

Spruce, Pine, Fir12

non-isothermal TGA

Nitrogen 100 ml/min

One-step global models Spruce; 262.4kJ/molPine; 197.14 kJ/molFir; 139.94 kJ/mol

Xylan13 isothermal TGA

473-613 K40-70 K/ s

Two-step consecutive model A1=3.62 X105 s-1

E1=18.3 kcal/molA2 = 3.83 x102 s-1

E2=13.1 kcal/mol

walnut shell 0.154 mm14

non-isothermal TGA

Nitrogen 20 ml/min

5,10, 20, 30 and

40 K /min

Two-step consecutive model A1=5.3X107 s-1

E1=120.158 kJ/molA2 = 6.19x1012 s-1

E2=154.414 kJ/mol

Straw,Corn stalks

100 μm15

isothermal TGA

400–648 K.Two-step consecutive model

Straw;

A1=2.43 X104 s1

E1=15.44 kcal/molA2 = 5.43 x101 s-1

E2=11.3 kcal/mol

Beech particle sizes below

80 μm16

isothermal quartz reactor

528–708 KSemi-global reaction mechanism

ln A1=10.2 s-1

E1=76.2 kJ/molln A2 = 22.5 s-1

E2=142.8 kJ/mol

ln A3= 2.3 s-1

E2=43.8 kJ/mol

Table 2 Conclusion of the kinetic parameters in pyrolysis process (Cont.)


Table 3 List of experimental design and setup of torrefaction process

Technique Atmospheres Feedstock References

isothermalNitrogen

bamboo, willow, coconut shell and woodspruce and beechbeech, willow , larch, strawhemicelluloses, cellulose and lignin

(Chen and Kuo, 2010)22

(Vincent et al., 2010)29

(Prins et al., 2006)4

(Chen and Kuo, 2011)5

Air Eucalyptus grandis wood (Rousset et al., 2012)32

Non-isothermal Air Eucalyptus (Arias et al., 2008)1

Table 4 Summary of the torrefaction kinetic parameters

Feedstock Technique Conditions Reaction mechanism Kinetic parameters

spruce and beech29

isothermal TGA

473-613 K40-70 K/s


two-step consecutive modelTwo step (spruce);k1 = 2.47 X104 s-1

E1=76 kJ/molA2 = 1.1x1010 s-1

E2=151.7 kJ/mol

beech and willow

0.7–2.0 mm4

isothermal TGA

225–300 0Cnitrogen

20 ml /min.10 0C/min

two-step reaction in series model Willow;A1=2.48X104 s-1

E1=76 kJ/molA2 = 1.1x1010 s-1

E2=151.7 kJ/mol

Pine 125 μm30

isothermal TGA,

bench- scale

fi xed bed

T=553 Knitrogen

50 ml/min

50 K/min

two-step reaction in series model k1 =1.48 x 10-3s-1

k2 =4 x 10-5s-1

Acknowledgement The authors would like to thank Energy Policy and Planning Offi ce, Ministry of Energy, Thailand for sup-

porting research fund.

References1. Arias B, Pevida C, Fermoso J, Plaza M G, Rubiera

F, Pis JJ. Infl uence of torrefaction on the grindability and reactivity of woody biomass. Fuel Processing Technology 2008;89:169-75.

2. Chen, W-H and Kuo P-C. Torrefaction and co-torrefaction characterization of hemicellulose,cellulose and lignin as well as torrefaction of some basic constituents in biomass. Energy 2011;36: 803-11.

3. Medic D, Darr M, Shah A, Potter B, Zimmerman J.Effects of torrefaction process parameters on biomass feedstock upgrading. Fuel 2012;91:147-54.


4. Prins MJ, Ptasinski KJ, Janssen FJJG. Torrefaction of wood Part 1. Weight loss kinetics. Journal of Analytical and Applied Pyrolysis 2006;77:28-34.

5. Chen W-H, Kuo P-C. Isothermal torrefaction kinetics of hemicellulose, cellulose, lignin and xylan using thermogravimetric analysis. Energy 2011.

6. Basu P. Biomass gasifi cation and pyrolysis. Oxford: Elsevier; 2010.

7. Chew JJ, Doshi V. Recent advances in biomass pretreatment – torrefaction fundamentals and tech-nology. Renewable and Sustainable Energy Reviews 2011;15:4212–22.

8. Yang H, Yan R, Chen H, Zheng C, Lee DH, Liang DT. In-depth investigation of biomass pyrolysis based on three major components: hemicellulose, cellulose and Lignin. Energy & Fuels 2006;20:388-93.

9. Yang H, Yan R, Chen H, Lee DH, Zheng C. Char-acteristics of hemicellulose, cellulose and lignin pyrolysis. Fuel 2007;86:1781-8.

10. Yu F, Ruan R, Steele P. Consecutive reaction model for the pyrolysis of corn. American Society of Agri-cultural and Biological Engineers 2008;51(3):1023-8.

11. Guo J, Lua AC. Kinetic study on pyrolytic process of oil-palm solid waste using two-step consecutive reac-tion model. Biomass and Bioenergy 2001;20:223-33.

12. Harun NY, Afzal MT. Thermal decomposition kinet-ics of forest residue. Journal of Applied Sciences 2010;10(12):1122-7.

13. Blasi CD, M. Lanzetta M. Intrinsic kinetics of isother-mal xylan degradation in inert atmosphere. Journal of Analytical and Applied Pyroiysis 1997;40-41:287-303.

14. Yuan HR, Liu RH. Study on pyrolysis kinetics of walnut. Journal of Thermal Analysis and Calorimetry 2007;89(3):983-986.

15. Lanzetta M, Blasi CD. Pyrolysis kinetics of wheat and corn straw. Journal of Analytical and Applied

Pyrolysis 1998;44:181-92.16. Branca C, Blasi CD. Kinetics of the isothermal deg-

radation of wood in the temperature range 528–708

K. Journal of Analytical and Applied Pyrolysis 2003;67:207-19.

17. Guo J, Lua AC. Kinetic study on pyrolysis of extracted oil palm fi ber: Isothermal and non-isothermal condi-tions. Journal of Thermal Analysis and Calorimetry 2000;59:763-74.

18. Ounas A, Aboulkas A, El harfi K, Bacaoui A, Yaacoubi A. Pyrolysis of olive residue and sugar cane bagasse: non-isothermal thermogravimetric kinetic analysis. Bioresource Technology 2011;102:11234-8.

19. Williams PT, Ahmad N. Investigation of oil-shale py-rolysis processing conditions using thermogravimetric analysis. Applied Energy 2000;66(2):113-33.

20. Plante AF, Fernández JM, Leifeld J. Application of thermal analysis techniques in soil science. Geo-derma 2009;153:1-10.

21. Mani T, Murugan P, Abedi J, Mahinpey N. Pyrolysis of wheat straw in a thermogravimetric analyzer: Effect of particle size and heating rate on devolatilization and estimation of global kinetics. Chemical Engineering Research and Design 2010;88:952-8.

22. Chen W-H, Kuo P-C. A study on torrefaction of variousbiomass materials and its impact on lignocellulosic structure simulated by a thermogravimetry. Energy 2010;35:2580-6.

23. Luangkiattikhun P, Tangsathitkulchai C, Tangsa-thitkulchai M. Non-isothermal thermogravimetric analysis of oil-palm solid wastes. Bioresource Tech-nology 2008;99:986-97.

24. Mansaray KG, Ghaly AE. Determination of kinetic parameters of rice husks in oxygen using ther-mogravimetric analysis. Biomass and Bioenergy 1999;17:19-31.

25. Bilbao R, Mastral JF, Aldea ME, Ceamanos J. Kinetic

study for the thermal decomposition of cellulose and pine sawdust in an air atmosphere. Journal Analytical and Applied Pyrolysis 1997;39:53-64.

26. Conceio MM, Fernandes Jr. VJ, Sinfrônio FSM, Santos JCO, Silva MCD, Fonseca VM, Souza AG. Evaluation of isothermal kinetic of sweetener. Journal of Thermal Analysis and Calorimetry 2005;79:263-6.


27. Munir S, Daood SS, Nimmo W, Cunliffe AM, Gibbs BM. Thermal analysis and devolatilization kinetics of cotton stalk, sugar cane bagasse and shea meal under nitrogen and air atmospheres. Bioresource Technology 2009;100:1413-8.

28. Chen W-H, Hsu H-C, Lu K-M, Lee W-J, Lin T-C. Thermal pretreatment of wood (Lauan) block by tor-refaction and its infl uence on the properties of the biomass. Energy 2011;36:3012-21.

29. Repellin V, Govin A, Rolland M, Rene´Guyonnet. Modelling anhydrous weight loss of wood chips dur-ing torrefaction in a pilot kiln. Biomass and Bioenergy 2010;34:602-9.

30. Peng JH, Bi HT, Sokhansanj S, Lim JC. A study of particle size effect on biomass torrefaction and densifi cation. Energy Fuels 2012;26:3826-39.

31. Pe´trissans A, Ge´rardin P, Younsi R, Pe´trissans M, Chaouch M. Experimental and numerical analysis of wood thermodegradation: mass loss kinetics. Journal of Thermal Analysis and Calorimetry 2011.

32. Rousset P, Macedo L, Commandré J-M, Moreira A. Biomass torrefaction under different oxygen con-centrations and its effect on the composition of the solid by-product. Journal of Analytical and Applied Pyrolysis 2012;96:86-91.

Original

Properties of Torrefi ed Biomass and Torrefi ed Pellet

Dussadeeporn Baonongbua1, Nitipong Soponpongpiapat1, Supachai Wasananon 1


AbstractThis paper reviews the torrefaction biomass properties, which have to be measured both laboratory and pilot scales. In laboratory scale, the products from torrefaction process are called torrefaction biomass. The properties usually determined in laboratory scale are heating value, grindability, and energy yield. In pilot scale, the torrefi ed biomass is formed into pellet and called torrefi ed pellet. The important properties of torrefi ed pellet are durability, hydrophobicity, and bulk density. The standards procedure used for testing the torrefi ed biomass properties and torrefi ed pellet are described in this review. Finally, the effect of torrefaction temperature on the properties of torrefi ed biomass and torrefi ed pellet is explained.

Keywords: Biomass, Grindability, Properties, Proximate analysis, Torrefaction, and Ultimate analysis

IntroductionRaw biomass has many drawbacks, when consideredas fuel, such as low heating value, high tenancy, high moisture content, low bulk density, poor grindability, high transport cost. Moreover, the raw biomass generatesa lot of smoke when it is combusted due to its high oxygen content. The poor properties of raw biomass can be improved by torrefaction process. The propertiesof torrefi ed biomass are low moisture content, high energy density, and good grindability. In addition, torrefied biomass with pelletization, called torrefied pellet, has a higher bulk density, durability and hydro-phobic than normal torrefi ed biomass and biomass pellet. This paper reviews the property of torrefi ed biomass and torrifi ed pellet.

Torrefaction process Torrefaction process is thermal decomposition

of raw biomass, in inert atmosphere, at low temperature of 200 - 300oC. When the biomass is heated, the com-position of biomass is decomposed. The decomposition mechanism can be described by two step reactions. First, hemicelluloses degraded; consequently, the light volatiles

(mono- and polysaccharide fractions and dehydrosugars)

and permanent gas (carbon dioxide and carbon monoxide) are revealed. Second, partial of cellulose and lignin decomposed. The level of cellulose and lignin decomposition depends on the torrefaction temperature. The products of these decomposition are condensable gas including acetic acid, water, formic, methanol, lactic acid, furfural and hydroxyl acetone which are formed and condensed as liquid product.1 Dehydration and decarboxylation reaction causes the increase of energy density, brittle property (good grindability) and

low moisture content. These properties are attraction properties for fuel in combustion system.

Analysis of torrefi ed products The products of torrefi ed process can be divided

into 3 groups: solid, liquid, and gas product. The importantproperties of solid product, torrefi ed biomass, are heating value, chemical composition, and grindability. Heating value is determined by the direct combustion or estimation by mathematical model. Chemical composition is analyzed

by proximate and ultimate analysis. After pelletization of solid product; bulk density, durability and hydrophobicity are also tested. The liquid products are diluted by 2-buta-

nol because all liquid products do not dissolve in water.1

1 Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, Silpakorn University(Sanam Chandra Palace),

Nakhon Pathom 73000, Thailand E-mail:[email protected]

Baonongbua et al. J Sci Technol MSU74

Later, diluted liquid products are analyzed by HPLC to identify liquid product components. The gas product composition is identifi ed by Micro GC.1,2,3

Properties of torrefi ed biomass In this section, we will discuss three important properties of torrefi ed biomass such as heating value, energy yield, and grinability. The details of discussion are as follows. Heating value Heating value is the most important properties, when the biomass fuel is applied for direct combustion and co- combustion.

Figure 1 data plotting of wood, torrefi ed wood (TW), charcoal,

coal, and peat on Van Krevelen diagram8

Normally, the heating value may be reported on two bases, higher heating value and gross calorifi c

value as well as low heating value, net calorifi c value.4 The heating value is determined by direct combustion in adiabatic bomb calorimeter, which is measurement conversion of enthopy between reactant and product,according to the ASTM – E 711,5 NBR 8633/84.6 Estimation

of heating value can also be determined based on data of proximate analysis or ultimate analysis. Proximate analysis is conducted to determine fraction of ash , fi xed

carbon, moisture and volatile matter by TGA/DSC Sys-tem (Metler Toledo),2 according to ABNT NBR 8112/86,6

ASTM D 5142-04 CEN/TS 14775:2004 (moisture), CEN/TS 14774-1-3:2004 (volatiles) and CEN/TS 15148:2005 (ash).7 Ultimate analysis is conducted to determine ele-

ment composition: C, H, O, N, S, and Cl by elemental analyzer, according to ASTM D 5373-08. For biomass from plants, the quantity of Cl is small compared with C, H, O, N, and S, thus Cl can be neglected.2,4 Sheng et al.4 The correlations for prediction of heating value using data from proximate and ultimate analysis are shown in Table 1. Erol et al.5 The element composition of some biomass is shown in Table 2. In addition, the data from ultimate analysis can also be used to identify the raw and torrifi ed biomass with classifi cation of coal. It can be done by Van Krevelendiagram. Before this diagram was used to classify biomass,Van Krevelen diagram was plotted in order to identify kerogen with type of petroleum.9 Van Krevelen diagram was established with O/C ratio as horizontal axis and H/C ratio as vertical axis. As can be seen in Figure 1, dry wood has the highest O/C and H/C ratio, while coal and charcoal has the lowest. Moreover, torrefi ed wood with temperature of 275, 280, and 285oC has O/C and H/C ratio similar to peat. Thus, the co-fi ring torrefi ed wood with peat instead of peat fi ring may be conducted. Chen et al.10studied the effect of temperature on higher heating value of torrefi ed biomass: bamboo, banyan, and willow. Figure 2 shows the relationship of the higher heating value versus the temperature. It can be seen from the fi gure that when temperature increases, the higher heating value increases. Energy yield Energy yield is defi ned as shown in Eq. (1).2,11,12,13,14

It represents the energy which is still contained in biomass after it has undergone torrefaction process.

It was found that the energy decreased with the increase of torrefaction temperature as can be seen in Figure 3.

Properties of Torrefi ed Biomass and Torrefi ed Pellet 75Vol 32, No1, January-February 2013

Figure 2 relationships between higher heating value and torrefac-

tion temperature of bamboo banyan and willow 10

Figure 3 relationships between energy yield and torrefaction

temperature11

Grindability The reduction of particle size results in the increase of the effi ciency and stability of combustion,

and the decrease of the unburnt carbon amount in ash7 Grindability refl ects the amount of energy used in biomass size reduction. In co-combustion biomass with coal and pulverized coal in boiler, the grindability property plays an important role. Improvement of grindability was con-

ducted to decrease tenancy of raw biomass in grinding process, resulting in the decrease of power consump-tion. The grindability for coals was tested by following

the Hardgrove Grindability Index (HGI) 7,15 The higher value of HGI means the lower energy consumption in

grinding or pulverizing process. It is the most signifi cant property in commercially aspect because HGI index is used to identify torrefi ed biomass with classifi cation of coal. The grindability test followed British Standard BS 1016–112:1995,7 ASTM D409-51(1961).11,15 The Hargrove Grindability Machine is shown in Figure 4. According to ASTM D409-51 (1961), 44 torrefi ed sample of 50 g was placed into grindability machine. The compressive force of 290 N was applied at the top of the machine and the grinding time was 3 minute. Once the grinding completed, the sample was sieved through 75 micron sieve and the weight of sample passing through sieve was recorded. In order to calculate the HGI value, the calibration chart had to be established. Standard reference sample with known HGI was ground. The weight of the original sample (50 g) minus the weight of the material retained on the 75 micron sieve was noted. Plotting on linear scalecoordinated with the calculate weight of material passing 75 micron versus the Hardgrave Grindability Index of the standard reference sample. A straight line was fi tted to these four points by method of least squares. The typical calibration chart was shown in Figure 5.

Figure 4 Hardgrove Grindability Machine44

The form of relationship between the weight of material passing sieve and HGI value can be written as follows:


Where, MH represents the weight (g) of ground product

passing the 75 micron sieve, m represents slope of calibration chart, and b represents intersection point on vertical axis. The estimation of HGI for torrefi ed biomass, which is defi ned as HGI equivalent (HGI

equiv), can be cal-

culated by Eq.(3).

Table 3 shows the correlation of HGIequiv

, grinding equipment, and standard used in previous work. It is noted the constant value in these correlation is specifi c to each grinding machine. On the other hand, establishment of calibration chart for a new grinding machine is needed.

Figure 5 the typical calibration chart 7

In order to confi rm that the HGI equivalent can be used to compare with HGI of standard reference coal, Bridgeman et al.7 compared the particle size distribution of torrefi ed biomass with standard reference coal, and the results are shown in Figure 6. From the fi gure, the size distribution of miscanthus D is similar to coal with HGI value of 92 and 66. Thus, it can be predicted that

the HGI equivalent of miscanthus is in range of 66-92.

Table 4 HGI value of willow and miscanthus at various torrefaction temperature and residence time7

t (min)

d (mm)

Willow

Miscanthus

m (%)

HGIequiv m (%)

HGIequiv

- - 0.5 0 0.1 0

10 Small 4.7 24 5.1 26

60 Small 1 0 1.2 1

10 Large 2.6 10 2.8 11

60 Large 9.1a 51a 13.4a 79a

Note: aSingle result, not duplicated

Figure 6 particle size distribution of miscanthus a nd coal7

Table 4 shows that the HGI equivalent of miscanthus D is 79. It is evident that HGI equivalent, determined by above correlation, can be compared with HGI of coal. In addition, it can be seen from Table 4 that the increase of torrefaction temperature results in the

increase of HGI value.

Properties of torrefi ed pellet The torrefaction and pelletization processes are combined in commercial scale. The fi nal product is called

torrefi ed pellet8. The important properties of torrefi ed pellet are heating value, bulk density, dura-bility and hydropho-

bicity. As determination procedure of heating value of torrefi ed pellet is similar to torrefi ed biomass discussed above, this section explains bulk density, durability and hydrophobicity.


Durability There are two methods to determine of durability: hardness and tumbling methods. The details of each method are as follows: Hardness method This method measures durability of torrefi ed pellet by hardness testing called Meyer hardness (H

M).

The torrefi ed biomass pellet was placed on a fl at plate and pressed by 6.35 mm cylinder as shown in Figure 7. Then, the force destroying a pellet was recorded.26,29 Equation for calculation of the Meyer hardness is as follow.

Where, D represents the diameter of probe (m), h represents indentation depth (m), F is the force that destroys a pellet (N).

Figure 7 the test set up of Meyer hardness29

Tumbling method Torrefi ed pellet durability was determined by

tumbling method according to the EN15210-1. The test set up of tumbling method is shown in Figure 8. Torrefi ed pellet without dust of 500 g was fi lled into chamber and exposed to 500 rotations in time interval of 10 minutes. The equation of durability value can be written as fol-

lows.30

Where, bm represents the mass of torrefi ed

pellet before testing (kg), and fm represents the mass of dust free torrefi ed pellet after testing (kg).

Figure 8 the test set up of tumbling method46

Figure 9 the effect of torrefaction temperature on durability30

Figure 9 shows the effect of torrefaction tem-

perature on durability. It can be seen from the fi gure that the raw biomass as reference parameter has the

highest durability, while the durability of torrefi ed biomass decreases with the increase of torrefaction temperature.

Bulk density The bulk density was defi ned as quotient of tor-refi ed pellet bulk mass and its volume. The bulk density could be determined by fi lling torrefi ed pellet into the standard box of 305 mm in length, 305 mm in width,

and 305 mm in height as shown in Figure 10 (ASTM E-873).45,46 The torrefi ed pellet was released from the height of 610 mm, then the box was jolted with 150 mm in height for 5 times to ensure that the box was full of torrefi ed pellet Finally, it was weighed and calculated for


bulk density value. Table 5 showed the bulk density of torrefi ed pellet of previous researches.

Figure 11 the relationship between moisture uptakes and expo-

sure time31

Hydrophobic Torrefaction process resulted in hydrophobic

characteristic of product. The hydrophobicity occured when OH groups were destroyed causing the biomass to lose the capacity to form hydrogen bond. Non-polar

unsaturated structures were formed resulting from above chemical rearrangement reaction. Lignin was decomposed and become plastic coating on biomass.26 These mecha-nisms made low uptake moisture of the torrefi ed biomass. The hydrophobicity of torrefi ed pellets could be evaluated by immersing torrefi ed pellet in water for a specifi c time, e.g., 15 hours.8 The hydrophobicity was evaluated by means of the amount of moisture uptake after testing. Sokhansanj et al.26, Li et al.32 compared amount of mois-ture uptake between torrefi ed pellets with raw biomass. They found that the amount of moisture uptake decreased with the increase of torrefaction temperature as can be seen in fi gure 11.

Conclusion This paper reviews the properties of torrefi ed biomass and torrefi ed pellet such as heating value, energy yield, grinability, durability, bulk density, and hydrophobicity. The test method and effect of torrefaction temperature on these properties are also reviewed.

Figure 10 the test set up for bulk density testing46


Table 1 Correlation for prediction of heating value

Number Name of author Correlation (HHV,MJ/kg) R2

1234567

8

9101112131415161718

19202122

Based on proximate analy-sisJimennezand Gonzalez33 Changdong Sheng et al.4

Demirbas34

Demirbas34

Cordero et al.35

Changdong Sheng et al.4

M.Erol et al.5

Chun-Yang Yin16

Based on Ultimate analysisTillman36

Changdong Sheng et al. 36

Boie37

Graboski, Bain38

Channiwala, Parikh43

Friedl et al. 17

Demirbas34

Jenkins39

Changdong Sheng et al.4 Chun-Yang Yin16

Based on chemical compo-sitionShafi zadeh and Degroota,40

Jimennez and Gonzaleza,33

Tillmana,36

Demirbasa,34

HHV =-10.81408+0.3133(VM+FC)HHV =19.914-0.2324AshHHV=0.196*FC + 14.119HHV=0.312*FC + 0.15332*VMHHV=0.3543*FC + 0.1708*VMHHV=-3.0368+0.2218VM+0.2601FCNHV= -116 – 1.33Ash-0.005VM+1.92(VM+Ash) -0.0227(VM+Ash) -0.0122 (VM)2 + 0.0299 (Ash)2 +6133 (OM)-1- 0.82(Ash)-1

HHV=0.1905VM + 0.2521FC

HHV=0.4373C-1.6701HHV=0.3259C+3.4597HHV=0.3516C+1.16225H-0.1109O+0.0628N+0.10456SHHV=0.328C+1.4306H-0.0237N+0.0929S-(1- Ash/100)(40.11H/C) + 0.3466HHV=0.3491C+1.1783H+0.1005S-0.1034O-0.0151N-0.0211AshHHV=3.55C2 – 232C-2230H+51.2C x H + 131N+20600HHV=0.335C+1.423H-0.0154O-0.145NHHV=-0.763+0.301C+0.525H+0.064OHHV= - 1.3675 +0.3137C+0.7009H+0.0318OHHV=0.2949C + 0.8250H

HHV=0.1739Ce + 0.2663L+0.3219EHHV=[1-Ash/(100-Ash)](0.1739Ce + 0.2663L+0.3219)HHV= 0.1739Ce+0.2663(1-Ce’)HHV=0.0889L+16.8218

0.5330.625-0.647-0.3060.2470.6170.898

0.9953

0.6660.7580.7200.6470.7330.9430.0810.7920.8340.9976

-0.503-0.451-1.068-0.875

aCe represent weight percent of cellulose and hemicelluloses , L represent lignin and extractives on dry basis biomass, and Ce’ represent

cellulose ( cellulose and hemicelluloses) on dry extractable-free basis


Table 2 Element composition of some biomass

BiomassUltimate analysis Proximate analysis HHV

(MJ/kg)C H O N VM FC A

Bagasse (sugar cane)18

Bamboo6,10

Cotton stalk12

Empty fruit bunches(oil palm)19 Kernel shell (oil palm)19 Lucerne14,41

Mesocarp (oil palm)19 Rape stalk20

Reed canary grass18

Rice straw19

Rubber seed kernel Straw21

Straw pellet8,42

Wheat straw18,41

Straw18

Corn stover42

Rice husk2

sugar cane residue13

Banyan13

Beech10,12

Birch8,22,42

Eucalyptus22

Larch11

Lauan wood8,42

Leucaena23

Loblolly pine14

Logging residue chip24

Pine wood chip25

Sawdust25

Willow23,26

Wood briquette8,10,18,42

Wood pellet27

Mallee41

Three pine pellet28

Jeffrey Pine29 and White Fir3

44.8043.8446.4345.5346.6850.1046.9246.9648.6039.0043.2147.5047.3044.3044.3050.4544.0449.1946.2047.2045.5049.0048.8048.7750.1050.2547.2947.2148.5447.2049.3748.5060.351.2149.02

5.806.056.185.465.867.405.896.136.805.085.975.800.636.805.804.656.556.946.086.006.206.106.106.777.405.976.206.646.216.106.590.055.36.025.93

49.1046.5342.6243.4042.0141.8042.6641.9537.3040.9650.2542.4045.5042.4048.2043.9440.4546.5345.2048.2044.6044.9044.3641.8043.3445.1945.7643.9544.8040.2444.9034.1442.6441.26

-

0.250.070.800.451.010.701.120.370.301.030.550.406.400.800.405.350.240.430.080.400.100.200.100.100.70

-0.420.171.500.34

-6.600.180.120.11

67.3173.5676.92

--

87.30-

76.3582.5068.83

-79.0080.8076.4079.00

-80.4593.6077.5784.20

--

82.8075.0886.10

-82.1785.9880.4987.60

-83.0559.50

--

-19.9419.19

--

12.2-

19.3012.1017.46

--

19.2017.30

--

8.603.6518.0315.50

---

17.2213.10

-16.0713.7619.42

10.7019.2016.9535.30

1.533.512.70

--

0.5-

3.165.508.59

-7.104.606.307.10

-10.952.51.110.30

--

0.10-

0.80-

1.770.270.241.702.800.300.70

--

15.5018.70

-17.0219.7820.4019.6118.7519.5017.12

-17.4017.8018.9017.4020.0017.4017.8020.2918.3016.4419.4019.5020.4120.3019.5518.7918.46

-19.0020.0218.5822.0018.6020.32

Table 3 The correlation for HGI equivalent of torrefi ed biomass.


Table 5 Summarized properties of various solid fuels

Properties Initial Particle size (mm) Particle Density

(kg/m3)

Bulk Density (kg/m3) Moisture Content (%)

Wood pellets28

Wood chips28

Torrefi ed pellets28

---

1103-1901N/A

1280-1361

498-649209-273610-668

8-1041-513.9-4.1

Torrefi ed pellets29 0.23 1419 157 1.96-3.06

of Three Pine 0.67 1431 197 2.71-3.58

0.81 1381 165 3.92-4.27

Acknowledgement The authors would like to thank Energy Policy and Planning Offi ce, Ministry of Energy, Thailand for sup-porting research fund.

References1. Prins MJ, Ptasinski KJ, Janssen FJJG. Torrefaction

of wood: Part 2. Analysis of products. Journal of Analytical and Applied Pyrolysis 2006;10:113-6.

2. Medic D, Darr M, Shah A, Potter B, Zimmerman J.Effects of torrefaction process parameters on biomassfeedstock upgrading. Fuel 2012:147-54.

3. Hoekman SK, Broch A , Robbins C. Hydrothermal carbonization (HTC) of lignocellulosic biomass.Energy &Fuels 2011;25:1802-10.

4. Sheng C, Azevedo JLT. Estimating the higher heating value of biomass fuels from basic analysis data.

Biomass and Bioenergy 2005;28(5):499-07.5. Erol M, Haykiri-Acma H, Kucukbayrak S. Calorifi c

value estimation of biomass from their proximate analyses data. Renewable Energy 2010;35(1):170-3.

6. Rousset P, A guitar C, Labbe N, Commandre JM. Enhancing the combustible properties of bamboo by torrefaction. Bioresource Technology

2011;107(17):8225-31.7. Bridgeman TG, Jone JM, Williams A, Waldron DJ.

An investigation of the grindability of two torrefi ed

energy crops. Fuel 2010;89(12):3911-8. 8. Bergman PCA, Kiel JHA, Boersma AR, Ptasinski KJ,

Prins MJ, Janssen FJJG. Torrefaction for entrained-fl ow gasifi cation of biomass. Energy research Centre of the Netherlands (ECN). 2005: p.2-50.

9. Weidendorfer D, Amrein JM, Ruppen D , Erni M, Zographos G, Lehmann S, Jeske T. Origin of hydro-carbons and petroleum: What can geochemistry tell us .Eidgenossische tecchische hochschule zurich swiss federal institute of technology zurich.2010: p.2-21.

10. Chen W-H, Cheng W-Y, Lu K-M, Huang Y-P. An evaluation on improvement of pulverized biomass property for solid fuel through torrefaction. Applied Energy 2011;88:3636-44.

11. Arias B, Pevida C, Fermoso J, Plaza MG, Rubiera F, Pis JJ. Infl uence of torrefaction on the grindability and reactivity of woody biomass. Fuel Processing Technology 2008;89:p.169-75.

12. Wang G, Luo Y, Deng J, Kuang J, Zhang Y. Pretreat-ment of biomass by torrefaction. Chinese Science Bulletin 2011;56:1442-8.

13. Wang MJ, Huang YF, Chiueh PT, Kuan WH, Lo SL. Microwave-induced torrefaction of rice husk and sugar cane residues. Energy 2012;37(1):177-84.

14. Wannapeera J, Fungtammasan B, Worasuwannarak

N. Effects on temperature and holding time during torrefaction on the pyrolysis behaviors of woody biomass. Journal of Analytical and Applied Pyrolysis 2011;92:99-105.

15. Shang L, Ahrenfeldt J, Holm JK, Sanadi AR, Barsberg S,

Thomsen T et al. Changes of chemical and mechani-cal behavior of torrefi ed wheat straw. Biomass and Bioenergy 2012;40(0):63-70.

16. Yin C-Y. Prediction of higher heating values of bio-mass from proximate and ultimate analyses. Fuel 2011;90(3):1128-32.


17. Friedl A, Padouvas E, Rotter H, Varmuza K. Predic-tion of heating values of biomass fuel from elemental composition. Analytica Chimica Acta 2005;544:191-8.

18. Bridgeman TG, Jones JM, Shield I , Williams PT. Torrefaction of reed canary grass, wheat straw and willow to enhance solid fuel qualities combustion properties. Fuel 2008;87:844-56.

19. Uemura Y, Omar WN, Tsutsui T, Yusup SB. Tor-refaction of oil palm wastes. Fuel 2011;90:2585-91.

20. Deng J, Wang G-j, Kuang J-h, Zhang Y-l, Luo Y-h. Pretreatment of agricultural residues for co-gasifi ca-tion via torrefaction. Journal of Analytical and Applied Pyrolysis.2009;86:331-7.

21. Harun NY, Afzal MT, Azizan MT. TGA analysis of rubber seed kernel. International Journal of Engineer-ing 2010;3:639-52.

22. Couhert C, Salvador S, Commandro JM. Impactof torrefaction on syngas production From wood Fuel 2009;88:2286-90.

23. Chen W-H, Hsu H-C, Lu K-M Lee, W-J, Lin T-C. Thermal pretreatment of wood(Lauan) block by tor-refaction and its infl uence on the properties of the biomass. Energy2011;36:p.3012-21.

24. Acharjee TCW, Coronella CJ, Vsque VR. Thermalpretreatment of lignocellulosic biomass. EnvironmentalProgress and Sustainable Energy 2009;28:435-40.

25. Phanphanich M, Mani S. Impact of torrefaction on the grindability and fuel characteristics of forest biomass. Bioresource Technology 2010;102:1246-53.

26. Li H, Liu X, Legros R, Bi XT, Lim CJ, Sokhansanj S. Pelletization of torrefi ed sawdust and properties of

torrefi ed pellets. Applied Energy 2012;93(0):680-5.27. Felfl i FF, Luengo CA, Sucrez JA, Beatn PA. Wood

briquette torrefaction. Energy for Sustainable Devel-opment 2005;9:19-22.

28. Abdullah H, Wu H. Biochar as a Fuel: 1. Properties

and Grindability of Biochars Produced from the Pyrolysis of Mallee Wood under Slow-Heating Condi-tions. Energy & Fuels 2009;23(8):4174-81.

29. Peng JH, Bi HT, Sokhansanj S, Lim JC. A Study of Particle Size Effect on Biomass Torrefaction and Densifi cation. Energy & Fuels 2012;26(6):3826-39.

30. Shang L, Nielsen NPK, Dahl J, Stelte W, Ahrenfeldt J, Holm JK, Thomsen T, Henriksen UB. Quality effects caused by torrefaction of pellets made from Scots pine. Fuel Processing Technology 2012;101(0):23-28.

31. Tumuluru JS, Wright CT, Hess RJ, D. Boardman R, Sokhansanj S.Review on Biomass Torrefaction Process and Product Properties. S-1041 Symposium on Thermochemical Conversion, Oklahoma State University, Stillwater, August 2nd, 2011.

32. Sokhansanj S, Peng JH, Bi XT, Lim CJ, Wang L, Lam PS, Hoi JP, Melin S, Tumuluru JS, Wright CT, Optimum torrefaction and pelletization of biomass feedstock Symposium. on Thermal and Catalytic Sciences for Biofuels and Biobased Products Iowa state University, 2010.

33. Jimennez L, Gonzales F. Study of the physical and chemical properties of lignocellulosic residues with a view to the production of fuels. Fuel 1991;70:947-50.

34. Demirbas A. Calculation of higher heating values of biomass fuels. Fuel 1997;76(5):431-4.

35. Cordedo T, Marquez F, Rodriguez-Mirasol J, Rod-riguez JJ. Predicting heating values of lignocellulosics and carbonaceous materials from proximate analysis.Fuel 2001;80(11):1567-71.

36. Tillman DA.Wood as an energy resource. New York: Academic Press, 1978.

37. Annamalai K, Sweeten JM, Ramalingam SC. Estimation of gross heating values of biomass fuels. Transactions of ASAE1987;301205-8.

38. Graboski M, Bain R.Properties of biomass relevant to gasfi cation. In: Reed TB, editor. Biomass gasifi cation:

principles and technology. New Jersey, USA: Noyes Data Coperation.1981.

39. Jenkins BM, Ebeling JM.Correlations of physical and chemical properties of terrestrial biomass with conversion Symposium energy from biomass and waste IX IGT,

1985.40. Shajizadeh F, Degroot WG. Thermal uses and

properties of carbohydrates and lignins. New York:

Academic Press1976.41. Ferro DT, Vigouroux V, Grimm A , ZanziR. Torrefac-

tion of agricultural and forest residues. Cubasolar, 2004.


42. Prins MJ.Thermodynamic analysis of biomassgasifi -cation and torrefaction Technische Universiteit Eind-hoven, 2005.

43. Channiwala SA, Parikh PP. A unifi ed correlation for estimating HHV of solid, liquid, and gaseous fuels. Fuel 2002;81(8):1051-63.

44. 1988 Annual book of ASTM standards. U.S.A.:Amirican society for testing and materials;1988 ASTM 409-51(1961) Standard Test Method for Grindability of Coal by the Hardgrove Mahine Method.

45. 1988 Annual book of ASTM standards. U.S.A.:Amirican society for testing and materials;1988 ASTM E873 - 82(2006) Standard Test Method for Bulk Density of Densifi ed Particulate Biomass Fuels.

46. LIFE (Laboratory of Innovation Fuel and Energy), Faculty of Engineering and Industrial Technology, Silpakorn University, Nakhon pathom, 2012.

Original

Torrefaction Reactors

Worason Junsatien,1 Nitipong Soponpongpipat1, Sivapong Phetsong1


AbstractThis paper reviews the torrefaction reactor technology, which is used for improving raw biomass properties. The operating principle of torrefaction reactor and classifi cation of reactors i.e. laboratory, pilot, and commercial scale will be described in this paper.

Keywords: Fixed bed reactor, Fluidized bed reactor, Microwave reactor, Moving bed reactor, Torbed reactor, Torrefaction reactor.

IntroductionNowadays biomass shows considerable promise for an alternative energy source. However, using raw biomass destined for combustion is challenged by severaldisadvantages and one of the drawbacks of biomass combustion is that it causes combustion instability resulted from high moisture content and large combustion chamber due to low energy density of this raw biomass. In order to improve the fuel properties of raw biomass, the torrefaction technology was conducted.1 – 4 The fi rst laboratory investigation of torrefaction process was began in France in the 1930s. At that time, the effect of temperature, heating rate and residence time on the process were investigated. In 1980s, Bourgois and Doat conducted their experiment testing on two types of wood and torrefaction temperature was varied at two values. Their work contributed to the pilot torrefaction plant in 1987s.5 The torrefi ed biomass has been proven to have high energy density, hydrophobic characteristic, and

durability of biodegradation.6, 7 Torrefaction process in

improving raw biomass properties, thus, has been gaining more attention in many parts of the world especially in Europe and North America.8 This paper reviews the oper-ating principle of torrefaction process and type of reactor in laboratory, pilot and commercial scale.

Operating principle of torrefaction process Torrefaction process is a thermal degradation of raw biomass. The temperature for torrefaction process occurs in the range of 200 – 3000C. The sub-procedure for torrefaction process can be divided into 4 procedures: 1) heating, 2) drying, 3) intermediate heating, torrefac-tion, and cooling, as can be seen in Figure 1. First, the raw biomass is heated up in heating procedure in order to increase the raw biomass temperature before going in to drying procedure. The temperatures of heating rate less than 500C per minute is recommended in heating procedure. When the raw biomass reaches the required temperature, generally around 90 - 1050C, drying of raw biomass start. The free water inside biomass is evaporated and the moisture content of raw biomass is decreased. It is necessary that removal water from raw biomass is performed since the higher of water or steam in reactor result in obstruction of raw biomass thermal degradation in torrefaction procedure.1 Once the drying process is completed, raw biomass was heated up to tor-

refaction temperature. The so-called intermediate heating, is required. In torrefaction procedure, the temperature of raw biomass is kept constant in the temperature range of

200 – 3000C depending on type of raw biomass. During thermal degradation period, hemicelluloses in raw bio-

1 Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, Silpakorn University (Sanam Chandra Palace),

Nakhon Pathom 73000, Thailand. Email:[email protected],

Torrefaction Reactors 85Vol 32, No1, January-February 2013

mass generally decompose fi rstly at the temperatures range of 200 – 2500C and later do lignin and partial of cellulose at 270 – 3000C.1, 2, 10 The change in composition of raw biomass from thermal degradation is shown in Figure 2. When torrefaction procedure is accomplished, the torrefi ed biomass increases in a high temperature and reacts with oxygen easily. Thus, the cooling of torrefi ed biomass is necessary in order to prevent hot torrefi ed biomass from spontaneous combustion.1

Figure 1 The sub-procedures of the torrefaction process9 Clas-

sifi cation and operating principle of torrefaction reactor

The torrefaction reactor can be classifi ed into 3 scales i.e. laboratory, pilot, and commercial scale. The production capacity of laboratory scale reactor is less than 20 kg/hr, while the capacity of pilot and commercial scale is 20-600 kg/hr and more than 600 kg/hr, respectively.1, 4,

10

Laboratory scale torrefaction reactor The laboratory scale torrefaction reactor was

established for the fi rst time in 1930s to investigate theparameters that affect torrefaction process.5 The information

from laboratory investigation was used for developing a pilot scale torrefaction reactor. Laboratory scale torrefaction reactor can be divided into 3 types as follows.

Figure 2 Change in composition of raw biomass from thermal

degradation11

Fixed bed torrefaction reactor The fi xed bed torrefaction reactor is the simplest reactor. The fi xed amount of raw biomass was fi lled inside the reactor, and was heated up by heat conduction from the electrical heater around the outside surface of reactor as shown in Figure 3. Jun et al.,12 who studied the torre-fi ed cotton stalk and wheat straw from fi xed bed reactor, found that the biomass yields after torrefaction had higher energy density and improved grindability characteristics compared with raw biomass. In addition, torrefi ed biomass also showed hydrophobic characteristics.12 Medic et al.14 used corn stover as raw biomass in the experiment. It

was found that the energy density of torrefi ed biomass increased by 2–19%, while the weight of torrefi ed biomass

decreased by 45%.14 Rousset et al.15 studied the enhancing of the combustible properties of torrefi ed bamboo and compare its elemental characteristic with lignite and

coal.15 Uemura et al.16 conducted investigation of oil palm waste.16 Bridgeman et al.17 investigated the grindability of torrefi ed energy crops: willow and miscanthus.17

Junsatien et al. J Sci Technol MSU86

Figure 3 Fixed bed torrefaction reactor13

Microwave torrefaction reactor Microwave torrefaction reactor uses the high frequency electromagnetic waves, so called microwave, for vibrating the water molecules inside biomass resulting in increase of biomass temperature. Wang et al.18 con-structed and tested microwave reactor. They suggest that the power level for torrefaction of rice husk and sugarcane residues is between 250 and 300 Watts. They also found that the caloric value of rice husk increase by 26% and 57% of sugarcane residues.18 The microwave-induced torrefaction reactor is shown in Figure 4.

Figure 4 Microwave-induced torrefaction reactor18

Fluidized bed torrefaction reactor The principles of the fl uidized bed reactor are as follows. The raw biomass is placed on the grate, and

the hot inert gas fl ows from the bottom through the raw biomass bed. At a suitable inert gas velocity, the raw biomass fl oats and behaves like a fl uid. This results in

uniformly thermal and temperature distribution throughout the raw biomass bed. The torrefaction of raw biomass

bed occur through. The fl uidized bed torrefaction reactor is shown in Figure 5. Li et al.19 studied the torrefaction of sawdust in a fl uidized bed reactor with nitrogen as inert gas. The nitrogen velocity was at 0.26, 0.29, and 0.32 m/s, respectively. The temperature of nitrogen was measured at 240, 250, 260, 270, 280, 290, and 300 oC. The residence time was 15, 20, 30, and 60 minutes. The study showed that when the severity of torrefaction was increased heating value of torrefi ed sawdust increase, while the energy yield decreases.19

Figure 5 Fluidized bed torrefaction reactor19

Pilot scale torrefaction reactor Based on the laboratory reactor results, pilot scale torrefaction reactor was developed. The data and parameter relation gathered from laboratory reactor will be verifi ed by pilot scale torrefaction reactor. The pilot scale torrefaction reactor can be divided into 3 types as follows. Fixed bed torrefaction reactor The operation principle of fi xed bed torrefaction reactor in pilot scale is similar to laboratory scale except

the heat source. The pilot scale derives a heat from raw biomass combustion as heat source, while laboratoryscale does from an electrical heater. The processdiagram of pilot scale reactor is shown in Figure 6. Like a laboratory scale, a pilot scale reactor cannot be operated

continuously. Energy research Centre of the Netherlands (ECN) designed and constructed this reactor in 2005s. The model is ECN bath reactor which has a production

capacity of 20 l.20 The ECN bath reactor model is shown

in Figure 7.


Figure 6 Process diagram of pilot scale reactor20

Figure 7 ECN bath reactor20

Fluidized bed torrefaction reactor The pilot scale of fl uidized bed torrefaction reactor

was developed by Topell Energy in 2010s. The capacity of this reactor was 60,000 tons/year. This reactor has a short reaction time and higher heat transfer effi ciency. Although

this technology is readily to scale up, there are a problem about particle size limitation and attrition inside reactor.21

Moving bed torrefaction reactor Moving bed torrefaction reactor uses mechanical

mechanisms for moving raw biomass from entrance to exit of reactor. It can be divided into 3 types as follows. Rotary kiln torrefaction reactor

The concept of rotary kiln torrefaction reactoris similar to commercial pyrolysis reactor. The raw bio-

mass was fed by screw feeder in stationary reactor, while the heating element in rotary drum rotates around the reactor. The production capacity of pilot scale was 50 kg/hr. The preferable moisture content of raw biomass for this reactor was 10%-15%/wt. There are two methods for heating up rotary kiln reactor: direct heating and indirect heating. For direct heating, superheat steam was used as heating media, while the hot oil was used as heating media for indirect heating. The residence time of raw biomass depend on rotating speed of rotary kiln. The low rotating speed results in long residence time which consequently led to carbonization rather than torrefac-tion. This reactor has a limitation in scaling up because rotary kiln cannot be use with vary size of raw biomass. There are only one values of the raw biomass size that correspond to reactor length, which contributing the high quality of torrefaction biomass.21, 22 The rotary kiln tor-refaction reactor was shown in Figure 8.

Figure 8 Rotary kiln torrefaction reactor23

Screw conveyor torrefaction reactor The screw conveyor torrefaction reactor

was shown in Figure 9.The raw biomass was conveyed

through reactor by screw. The heat source for reactor

is achieved by the fl ue gas from combustion. The raw biomass size for this reactor should smaller than 10 mm.

In addition, raw biomass with very low bulk density and high moisture content is not suitable for this reactor. Because of the quality of torrefi ed biomass depend on the screw diameter, thus, it is diffi cult to scale up this reactor.22 Energy research Centre of the Netherlands (ECN) designed and constructed this reactor in 2008s. It had product capacity of 60-100 kg/hr with operating temperature of 220-280 oC.24


Figure 9 Screw conveyor torrefaction reactor4

Multiple hearth torrefaction reactor The multiple hearth torrefaction reactor was shown in Figure 10. The biomass was fed from the upper side of reactor into the fi rst hearth. The raw biomass was heated up by contact with hot surface of hearth. Later, the raw biomass was swept into the second hearth and next hearth by rabble teeth which installed on rabble arm.22

Figure 10 Multiple hearth torrefaction reactor25

The commercial scale torrefaction reactor The commercial scale torrefaction reactor can be classifi ed into 6 types as follows.

The fi xed-bed torrefaction reactor This reactor was developed for a long time because of its simplicity. The operating principle of this reactor is similar to pilot scale reactor. However there is a disadvantage for this reactor i.e. irregularly in temperature

distribution throughout the reactor. Integrofuels company

was established this reactor in the 2010s. The production capacity of reactor was 48,000 tons/year. It was found that the loss weight of biomass was 20-30% compare with initial weight, while retaining 90% of its energy.26

Torbed torrefaction reactor Figure 11 show the torbed torrefaction reactor. By injection of high velocity gas about 50-80 m/s through stationary angled blade, toroidal fl ow pattern occurred. The gas particle lifts and move biomass bed horizontally at the same time results in shallow biomass bed moving around vertical axis at center of reactor. The heat and mass transfer occur in this bed easily which consequently led to short residence time and homogenous torrefi ed product.22, 27 Topell tested torbed torrefaction reactor at Poland in 2007s. In 2010s, they designed and constructed this reactor with production capacity of 60,000 tons/year, and operating temperature of 280-3200C. The residence time of this process was 90 seconds.1, 4

Figure 11 Torbed torrefaction reactor27

Oscillating belt conveyor reactor The oscillating belt conveyor reactor is shown in Figure 12.The raw biomass is fed on belt conveyor into reaction. The belt conveyor is oscillated resulting in uniformly thermal distribution and torrefaction of raw


biomass on conveyor. Flue gas residue fro m torrefac-tion process was used for drying of raw biomass before being fed into reactor.22, 27 This reactor was constructed in 2010s by Agritech Producer Columbia, namely Torre -Tech ® 5.0. Its operating temperature was in range of 300 - 400°C, reaction time was 30 min, and capacity was 50,000 tons/year. The product of this reactor retained 80% of raw biomass energy.28

Figure 12 Oscillating belt conveyor reactor26

Rotary drum torrefaction reactor The rotary drum torrefaction reactor was shown in Figure 13. The raw biomass passes through the reactorby lifting fl ights inside inner shell of rotary drum. The heat-

ing gas fl ows in gap between inner and outer shell. It is noted that inner and outer shells rotate together when this

reactor was operated.21 This reactor was established in 2011s by Bio Energy Development North AB (SWE). Its capacity was 25,000 - 30,000 tons/year and Atmosclear

in 2010s with capacity of 50,000 ton/years.4

Figure 13 Rotary drum torrefaction reactor4

The microwave torrefaction reactor The commercial scale of microwave torrefaction reactor was developed by Rotawave Ltd. in 2011s with production capacity of 110,000 tons/year.4 The plant layout of this reactor was shown in Figure 14.

Figure 14 Plant layout of microwave torrefaction reactor4

The screw torrefaction reactor

The commercial scale of screw torrefaction reactor is developed by BioLake B.V. in 2010s with production capacity of 5,000-10,000 tons/year, and FoxCoal B. V.

in 2012s with production capacity of 35,000 tons/year.4 The plant layout of this reactor was shown in Figure 15.


Figure 15 Plant layout of screw torrefaction reactor29

Conclusion Torrefaction is technology for improving biomass properties. The torrefaction reactor can be divided into three scales: laboratory scale (including fi xed- bed reactor, fl uidized bed reactor and microwave reactor), pilot scale (including fi xed- bed reactor, fl uidized bed reactor and moving bed reactor), and commercial scale (including fi xed - bed reactor, torbed reactor, oscillating belt con-veyor, rotary drum reactor, microwave reactor and screw reactor). The operation principle and production capacity of each reactor were described in this reviews. Although there are many types of reactor, the further research in reactor design for minimum energy used is still need.

Acknowledgement The authors would like to thank Energy Policy

and Planning Offi ce, Ministry of Energy, Thailand for supporting research fund.

References1. Tumuluru JS, Sokhansanj S, Wright CT, Boardman

RD, Hess JR. Review on biomass torrefaction process and product properties and design of moving bed torrefaction system model development. ASABE

Annual International Meeting. Gault House Louisville, Kentucky: 1110459; August 2011.

2. Tumuluru JS, Sokhansanj S, Wright CT, Board-

man RD, Hess JR. A technical review on biomass processing: densifi cation, preprocessing, modeling,

and optimization. 2010 ASABE Annual International Meeting: 1009401; June 2010.

3. Tumuluru JS,Sokhansanj S, Wright CT, Boardman RD, Hess JR.A review on biomass torrefaction proc-ess and product properties for energy applications. Industrial Biotechnology 2011;18.

4. Kleinschmidt K. Overview of international develop-ments in torrefaction.Central European Biomass Conference 2011.

5. Bourgois J, Doat J, Bioenergy 1984; 3:153-9.6. Bergman PCA, Kiel JHA.Torrefaction for biomass

upgrading.In Netherlands ErCot, editor.Published at 14th European Biomass Conference & Exhibition. Paris, France; 2005.

7. Arcate. Global markets and technologies for torrefi ed wood in 2002. Wood Energy 2002; 5.

8. Mitchell D and Elder T. Torrefaction? What’s that?. Council on Forest Engineering Annual Meeting. Auburn, Alabama: Fueling the Future; 2010, p. 7.

9. Bergman PCA, Boersma AR. Torrefaction for entrained-flow gasification of biomass. Energy research Centre of the Netherlands 2004.

10. Verhoeff F, Arnuelos AAI, Boersma AR, Pels JR, Lensselink J, KielJHA, Schukken H. Torrefaction tech-nology for the production of solid bioenergycarriersfrom biomass and waste. Energy research Centre of the Netherlands 2011; 75.

11. Bergman PCA, Boersma AR, Zwart RWH, Kiel JHA.Torrefaction for biomassco-fi ring in existing coal-fi red power stations. Report ECN-C--05-013, ECN,Petten (2005a).

12. Jun WG, Hao LY, Jian D, Hong KJ, Liang ZY.

Pretreatment of biomass by torrefaction. Chinese Science Bull May 2011;7.

13. Chen WH, Hsu HC, Lu KM, Lee WJ, Lin TC. Thermal

pretreatment of wood (Lauan) block by torrefaction and its infl uence on the properties of the biomass. Energy 2011;36:3012-21.

14. Medic D, Darr M, Shah A, Potter B, and ZimmermanJ. Effects of torrefaction process parameters on

biomass feedstock upgrading. Fuel 2012.


15. Rousset P, Aguiar C, Labbé N, Commandré JM. Enhancing the combustible properties of bamboo by torrefaction. Bioresource Technology 2011;7.

16. Uemura Y, Omar WN, Tsutsui T, Yusup SB. Torrefaction of oil palm wastes. Fuel 2011;7.

17. Bridgeman TG, Jones JM, Williams A, Waldron DJ. An investigation of the grindability of two torrefi ed energy crops. Fuel 2010;8.

18. Wang MJ, Huang YF, Chiueh PT, Kuan WH, Lo SL. Microwave-induced torrefaction of rice husk and sugarcane residues. Energy 2011;8.

19. Li h,Liu X, Legros R, Bi XT, Lim CT, Sokhansanj S. Torrefaction of sawdust in a fl uidized bed reactor. Bioresource Technology 2012;6.

20. Kiel J. ECN’s torrefaction - based BO2 - technology -

from pilot to demo. Presented at IEA Bioenergy work shop torrefaction.Graz, Austria; 2011.

21. Chew JJ, Doshi V. Recent advances in biomass pretreatment - Torrefaction fundamentals and tech-nology. Renewable and Sustainable Energy Reviews 2011;11.

22. Beekes W, Cremers M. Torrefaction Cracks the Biomass Challenge. Renewable Energy World Maga-zine 20012.

23. http://www.ievb.tu-clausthal.de24. KielJHA, Verhoeff F, Gerhause H,Meuleman B.

BO2 - Technology for biomass upgrading into solid

fuel - pilot - scale testing and Market Implementa-tion. In: Netherlands ErCot, editor. Presented at 16th European Biomass Conference & Exhibition.Valencia, Spain: Energy research Centre of the Netherlands; June, 2008, p. 7.

25. http://www.impexequip.com/equip_calcination.html26. Fuels IE. Torrefied wood.Roxboro, NorthCaro-

lina2008.(http://www.integrofuels.com/production.html)

27. Ontario Power generation. Torrefaction treatment process. October 27, 2010.

28. James J.Using torrefi ed wood as a coal replacement, for superior pellets and cellulosic ethanol production.Agri-Tech Producers, LLC. 2008.

29. http://www.biolake.nl/documents/leafl et_bi-olake.pdf

Original

1 Assoc. Prof., English Department, Faculty of Arts, Silpakorn University (Sanam Chandra Palace), Nakhon Pathom 73000, Thailand.

Email: [email protected],

Components of the Discussion Section in Biomedical Engineering Research Articles

and their Linguistic Characterization

Budsaba Kanoksilapatham1


AbstractThis study aims to develop the academic writing skills of biomedical engineers by analyzing a dataset of 37 discussion sections in biomedical engineering research articles written in English. Specifi c objectives of this study are twofold: 1) to identify the components that make up this section classifi ed as ‘moves’ and 2) to describe how each component of move is expressed. First, the analysis reveals that this section consists of three principal moves, and each move possibly entails a number of sub-units called ‘steps’. The analysis also demonstrates that these units of texts are organized in a particular pattern forming a typical sequence. Second, the examination to scrutinize how these textual units are written shows that a number of linguistic features are conventionally used to serve a particular purpose. This study bears pedagogical implications preparing scholars who are not competent in English to be able to disseminate their scientifi c achievements regionally and internationally.

Keywords: Biomedical engineering, academic writing skills development, discussion section, linguistic characterization

IntroductionIt is known that the ability to write like a professional is a central requirement for a successful career.1 Given that professional scientifi c and engineering fi elds are dominated by English,2 the policy ‘publish or perish’ in an international journal has been adopted by several universities and academic institutions. At this juncture, being competent in scientific writing and discovery dissemination embodies a knowledge of the genre and its associated textual features including lexico-grammatical features, organization, communicative functions, disci-plinary conventions (how to report numbers and units,

format tables), and content.3-6

Typical academic genres in which scholars and practitioners are involved include research articles (RAs), research proposals, and conference abstracts, Apparently,

RA genre has become one of the primary means by which new scientifi c discoveries and claims are disseminated in various disciplines.7

A number of studies focus on different sections of RAs. Since Swales8-9 developed his ground-breaking framework for analyzing RA introductions, the introduction section has been a particular focus of attention.10-11 The framework has served as a foundation for research into the other conventional sections of empirical RAs, namely, Methods,12 Results,13 and Discussion.14 Moreover, the framework has stimulated a range of discipline-specifi c and cross disciplinary studies of different sections of RAs.15-16 A much smaller number of research studies were conducted on full-length articles.17 All these studies have

demonstrated the power of genre or move analysis as an effective tool to elucidate how academic written discourse is constructed and consequently facilitates the learning and/or instruction of academic writing. A review of genre-based studies reveals that,

among the four conventional sections of RAs, whereas the Introduction section has received relatively substantial interest, few studies focus on the discussion section17

Components of the Discussion Section in Biomedical Engineering

Research Articles and their Linguistic Characterization93Vol 32, No1, January-February 2013

Despite some commonalities shared across disciplines, these studies congruently reveal that disciplinary variation is discernible. Biomedical engineering, as a discipline, plays a pivotal role in our life, focusing on the use of the principlesand techniques of engineering to solve problems in biology and medicine. However, very little appears to be known about RA writing in biomedical engineering. Therefore, the present study aimed to develop the aca-demic writing skills of biomedical engineers, by examining biomedical engineering RAs with reference to Swales’ schematic framework. In fact, this study, as a part of a larger project that focuses on the four conventional sec-tions of full length RAs, focuses on the discussion section of biomedical engineering RAs. Specifi cally, the objectives of this study are twofold: 1) to identify the components of this section or so-called move or text structure, and 2) to linguistically characterize individual components or moves inherent in this section. It is anticipated that this study will be able to cater to the demand of academic and professional uses of English, preparing them for engineering communication in the genre of RAs. The text structure outline, combined with the linguistic description, implicates that developing reading and writing skills of this particular genre can be accomplished with explicit instruction.

Experimental Analytical framework of move analysis

Genre or move analysis created by Swales18-19 is recognized as a textual analysis approach that identifi es

textual organization originally in the academic English discourse of Introductions, and subsequently extended to other sections of RAs. In brief, this textual analysis approach considers that each genre has a pattern, consisting of sub-units called ‘moves’. Each move, in turn,

possibly consists of a number of ‘steps’ or sub-units of a move. The textual segmentation was manually conducted based on the communicative function of a text segment

and the linguistic features associated with the function.Journal article dataset selection and compilation

Journal and journal RA selection were conducted to assure that the dataset analyzed represents the top fi ve quality journals in the fi eld of biomedical engineering. Based on the impact factors of journals in the discipline of biomedical engineering in 2005 (the most recent data available at the onset time of dataset compilation), fi ve jour-nals were selected. From each journal, twelve RAs were randomly included in this dataset. In short, the dataset consists of 60 biomedical engineering RAs published in 2006. Despite the growing frequency of a combined results and discussion section in many academic disci-plines, for pedagogical purposes, this study was designed to analyze the stand-alone discussion section, because it was important for students who are unfamiliar with the genre to be able to understand and to distinguish the distinct functions of the discussion section as revealed by their conventional moves. Given that the focus of this paper is on the discussion section, only RAs from the original dataset that have a distinct and stand-alone section of discussion were analyzed. From the original 60 RA dataset, the fi nal dataset to be analyzed by genre or move analysis in this study consists of 37 discussion sections.

Dataset analysis The 37 discussion dataset was analyzed by Swales’ move analysis. That is, discussion texts were hand-tagged or segmented into textual units called ‘moves’ and their sub-units called ‘steps’, based on their communicative functions and associated linguistic

features. Finally, based on the moves and steps identifi ed, text organization of the discussion section in biomedical engineering RAs is outlined, together with their frequency

of occurrence. The organization presented, however, represents only one common channel of disseminating biomedical engineering knowledge in the RA genre.

Results and Discussion The analysis of the discussion section reveals three major moves. They are Move 1: Preview the present

Kanoksilapatham et al. J Sci Technol MSU94

study, Move 2: Consolidate results, and Move 3: State limitations and possible future research. These three moves are presented in the order likely to be found in the dataset. In order to characterize the linguistic features associated with these moves, in the second objective of this study, representative examples displaying exact verbatim of these moves in the dataset were illustrated, with two modifi cations: 1) replacing citations with (R), and 2) replacing people’s names with XXX. Finally, linguistic features that are associated with a particular move are highlighted. Move 1: Preview the present study

Move 1: Preview the present study usuallyreminds readers of the purpose and the major methodo-logical features of the study, as shown.

(1) It is well known that dynamic properties of closed-loop force controlled robotic devices get worse if the coupling between the robotic device and the object, to which the force is applied, gets stiffer (R). (2) Accuracy of the models generated with the scanner method in our study was determined by com-paring data derived from the virtual images with reference values obtained by manual measurements on the dummy. Instance (1) demonstrates that Move 1 can accomplish its function, for instance, by providing the background of the research topic. As shown, the con-struction of it is + past participle + that clause and the use of reference (R) are used to express the established background information. In addition, the use of present tense verbs (get, is) indicates the contemporary nature of the statement.

However, (2) represents a variation of Move 1 that opts to provide a brief description of the methods (determined and comparing) employed by the present study. In this regard, it is noted that the verb was in past tense (determined) to highlight the completion of the

activity mentioned. These two instances of Move 1 successfully assure readers that the present study was contextualized,

based on background information, and carefully planned and executed. In short, this move provides readers with a snapshot or an overview of the study.

Move 2: Consolidate results Not all discussion sections in biomedical engi-neering begin with Move 1. In the absence of Move 1, the discussion sections are likely to begin with Move 2: Consolidate results. In this move, major results are initially and frequently highlighted and subsequently strengthened by one step or a combination of some of the seven steps to serve a range of functions, as detailed below. Move 2, Step 1: Report results In order to strengthen the results generated by the study, it is crucial that a fi nding be primarily introduced to be subsequently discussed. (3) None of the investigated materials showed any signifi cant acute or chronic infl ammation at or after 7 days post-implantation… (4) Table II and Figs. 8 and 9 show that contrast normalization improves the discrimination ability of individual features. As shown in (3) and (4), the verb show is prominent, varying in tense, either past or present, stating a fi nding of the present study. It is noted that in this discipline of biomedical engineering, the tense choice of this particular verb seems to be determined by its subject. That is, when the subject is Table or Figure, the verb show is in the present tense. However, the verb show in past tense is used when the fi nding is reported. The presence

of this move/step, the centerpiece of the discussionsection, provides a basis for the other steps of this move

to consolidate fi ndings, frame arguments, stake claims, etc.

Move 2, Step 2: Interpret results After a result is presented, one of the most common attempts or strategies to highlight the result is

by means of interpretation. Thus, the section usually proceeds with an interpretation of results. (5) This suggests that the added sensory infor-mation about body sway enhanced the natural, direction specifi c, automatic postural control strategies rather than

superimposing a generalized stiffening. (6) Incubation of Promogran in a buffer contain-ing purifi ed neutrophil elastase indicated that the protein

component of the dressing material is susceptible to degradation.



In (5) and (6), the verbs suggest and indicate (be they in present or past tense) congruently express the researchers’ interpretation of the fi ndings. Move 2, Step 3: Explain results It is also common that the presentation of a fi nding needs to be accompanied by possible explanations.Scientifi c results do not speak for themselves. (7) These differences are mainly attributable to the limitations of the CDPIV system to measure velocity larger than 20 cm/s. (8) This may be because of interdependencies between the occurrence of MAs in an image. The phrases highlighted (attributable to, because of) contribute to the function of offering explanations for the fi ndings. The explanation appears to involve a certain level of speculation, and thus the modal may is concur-rently used in (8). Move 2, Step 4: Compare results A study can be consolidated and contextualized by comparing the results of the present study with those of previous studies. (9) The low localization errors and the relatively high magnitude changes are in good agreement with the results of XXX (R). (10) Fiber strain predictions from the normal FE model were in good agreement with the predictions of the cylindrical numerical model developed by XXX (R). Both (9) and (10) display the use of the phrase in good agreement with to highlight the compatibility and support of the results of the present study with those generated from previous studies (indicated by R or

citations). By comparing fi ndings, the present study is connected to a larger body of evidence. Move 2, Step 5: Conclude results Multiple findings produced by a study in biomedical engineering can lead to a conclusion or a brief

summary of the work, as shown. (11) Overall, this study demonstrated a multi-modal method for assessing biocompatibility in preclinical

testing of cardiovascular devices. (12) On the whole, our study showed no es-sential correlation between the precision of the models

created with the scanner method and the orientation of the scanner laser beam. Both (11) and (12) use connectors like overall and on the whole to conclude the fi ndings. In addition, the use of this/our study indicates the authority of the researchers in stating such a claim or conclusion. Finally, the use of reporting verbs in past tense (demonstrated, showed) to conclude the fi ndings suggests that the authors do not seem to claim a generalization of the conclusion(s) presented. Move 2, Step 6: Exemplify results To substantiate a claim, statement, or argument, examples can be quite helpful in convincing readers. (13) For example, rats would turn their heads left or right, duck, or paw at the food aperture upon tone onset during many of the experiments. (14) For example, in images containing MAs at various stages of development and hence having a range of sizes, it is not important to detect the small faint MAs as long as the largest MA can be detected. In these two instances, the phrase for example clearly serves the purpose. Move 2, Step 7: Claim values of the results Even though it is apparent that our life quality enhancement is in the hands of biomedical engineers, they still feel the need to explicitly justify the value of their fi ndings. (15) Automation as a prescreening phrase of image grading will make screening programs more cost effective by reducing the manual grading workload. (16) The present study provides significant

evidence demonstrating that a number of ionic derivatives of a 60 mol% S-SEBS polymer offer some signifi cant advantages as wound dressing materials.

As shown, biomedical engineers stake their claims about the contribution, application, and implication of the fi ndings by assuring readers that their discovery is cost effective and advantageous.

In summary, of all of the seven steps of Move 2, Step 1 seems to be the center of the discussion. Based

on this move/step, a number of reactions are proposed:

Kanoksilapatham et al. J Sci Technol MSU96

interpretations made, explanations offered, comparisons conducted, conclusions framed, examples provided, and values of the fi ndings claimed. Move 3: State limitations and possible future research avenues In this move, biomedical engineers opt to express opinions regarding the limitations of the study (17). This move allows the researcher to caution readers that the study was not perfect, and thus fi nding interpretation and implication should be conducted with care. In some discussion sections, the authors offer possible avenues or directions for future research (18). (17) The scope of the proposed model is limited to objectively assess technical factors of surgical ability. (18) Other approaches are currently emerging that seem to hold promise for further progress, such as halography, which features millisecond exposure times and an excellent depth of fi eld. (17) and (18) illustrate how biomedical engineers move the fi eld forward. By explicitly stating limitations of the present study and suggesting directions for future research avenues, scholars in the fi eld are aware of some potential weaknesses inherent in the present study and can mobilize concerted efforts to enhance the quality of future research.

Pattern of move structure or text organization The moves and steps presented above seem

to be in congruence with what previous studies on this section of a range of disciplines14,17 have revealed. That is, the section consists of three major moves; some moves with a number of steps. Moreover, as far as patterning is concerned, how the three moves and their constituent

steps interact with each other are multiple and varied. However, in this article, to facilitate scholars in the fi eld who are involved in the task of writing RAs, the pattern

presented here refl ects one of the many possible but common ways to organize discussion texts.

Move 1 (Preview the present study), if found, is likely to begin the section, followed by Move 2 (Con-

solidate results). Similarly, Move 3 (State limitations and possible future research avenues), if available, is likely to end the section. If all are found, these three moves tend to occur in a 1-2-3 sequence or pattern. In addition, cyclical patterning of Moves 1 and 2 is frequently found in this dataset. As previous studies indicated14,17 although the moves and steps are commonly found across disciplines, each discipline seems to have its own preference. It is observed that some moves occurred more frequently than the others. That is, some moves are considered relatively optional, whereas others are relatively conventional. In order to indicate the status of each move (be it optional or conventional), the frequency of occurrence was used as reference, with an arbitrary cut-off of 60%. It was found that Moves 1 and 2 areinterestingly equally frequent, with 94.59% of occurrence, being found in 35 out of a total of 37 discussion sections. Move 3 is slightly less frequent, found in only 29 sections or 78.38%. At this juncture, it can be concluded that these three moves are quite conventional, with more than 60% of frequency occurrence. An examination of the frequency of occurrence is also intriguing. At the step level, the frequencies of the seven steps of Move 2 vary substantially. Step 1: Report results was found in all except two discussion sections (94.59% or in 35 out of 37 sections). The other six steps, in the order of more frequent to less frequent, are Step 2: Interpret results (89.19% or in 33 sections), Step 3: Explain results (75.68% or in 28 sections), Step 4: Compare results (67.57% or 25 sections), Step 5: Conclude results

(43.24% or in 16 sections), Step 7: Claim values of the results (40.54% or in15 sections), and Step 6: Exemplify results (16.22% or in 6 sections). The data above demonstrated the specifi c nature of this section of biomedical engineering. That is, biomedi-

cal engineers prefer the strategies of reporting results the most (Step 1, about 95%), followed by interpreting sections (Step 2) and explaining results (Step 3), with more than

70% of frequency occurrence. The least preferred step of Move 2 is exemplifying results.



Conclusion It is known that developing academic writing is a great challenge. To address this challenge, this study has presented a detailed analysis of the move structure found in 37 empirical discussion sections. Although the discussion section is not claimed to be a default option for organizing RAs, this section is considered to be one of the standard sections for a number of disciplines.8,9,17

Moreover, this section is distinct, making connections between the new knowledge or understanding reported in the result section and previous studies in the fi eld. As shown, in order to succeed in writing this section, authors need to have not only novel, original, exciting results but also the skills and expertise to develop, evaluate, communicate, and argue persuasively to support the claim or evidence offered in order to maintain credibility.

It should be noted that the pattern outlined and presented in this article is not meant to be a recipe. Rather, it should be treated as one of the many possible ways to be successful at conveying a message by writing the section following the academic norms. It is anticipated that the components identifi ed in this section and the lin-guistic characterization of individual components will also allow particularly novice readers to follow the fl ow of the biomedical engineers’ thoughts. For scientists in

peripheral countries, this study will help them to become fully integrated members of the worldwide network of science.

Acknowledgement The compilation and the analysis of the biomedi-cal engineering research article dataset was supported by the Thailand Research Fund, Grant No. RSA5080005.

References1. Hyland K, Salager-Meyer F. Scientifi c writing. ARIST

2008:297-339.2. Ammon U. Language planning for international

scientifi c communication: an overview of questions and potential solutions. CILP 2006;7(1):1-30.

3. Feak C, Swales J. Academic writing for graduate students: essential tasks and skills. Ann Arbor: University of Michigan; 2011.

4. Hyland K. Disciplinary discourse: social interactions in academic writing. Ann Arbor: University of Michigan; 2004.

5. Swales J, Feak C. Abstracts and the writing of abstracts. Ann Arbor: University of Michigan; 2009.

6. Swales J, Feak, C. Navigating academia: writing supporting genres. Ann Arbor: University of Michigan; 2011.

7. John A, Swales J. Literacy and disciplinary practices: opening and closing perspectives. EAP 2002;1:13-28.

8. Swales J. Genre analysis: English in academic and research settings. Cambridge: Cambridge University Press; 1990.

9. Swales J. Research genres: explorations and applications. Cambridge: Cambridge University Press; 2004.

10. Kanoksilapatham B. Language of civil engineering introductions: textual structure and linguistic charac-terizations. Asian ESP 2011;7, (2):55-84.

11. Ozturk I. The textual organization of research article introductions in applied linguistics: Variability within a single discipline. ESP 2008; 26(1):25-38.

12. Lim JMH. Methods sections of management research articles. ESP 2006;25(3):282-309.

13. Brett P. A genre analysis of the results section of sociology articles. ESP 1994;13(1):47-59.

14. Basturkmen H. A genre-based investigation of discussion sections of research articles in dentistry

and disciplinary variation. EAP 2012;11(2):134-44.15. Bruce I. Results sections in sociology and or-

ganic chemistry articles: a genre analysis. ESP 2009;28(2):105-24.

16. Lim JMH. Commenting on research results in applied

linguistics and education: a comparative genre-based investigation. EAP 2010;9(4):280-94.

17. Kanoksilapatham B. Rhetorical structure of biochem-

istry research articles. ESP 2005;24(3):269-92.

Original

A Comparison of Learning Achievement Toward cost accounting I Between The Group has Solution of Homework and The Group not has Solution of Homework

Manop Seeluang


AbstractThe purpose of this research was to compare learning achievement for E-Learning towards cost accounting I sub-ject of students between the group has solution of homework and The Group not has Solution of homework. The sample consisted of 100 undergraduate students in Silpakorn University, who were studying cost accounting I in the second semester 2011. The subject was devided into an experimental group (the group has solution of homework) of 50 student and a control group(The Group not has Solution of homework) of 50 students. The instruments used for this research ware a lesson plan, the learning achievement test and the formative tests. The statistical methods such as the mean(), the standard deviation(SD) and the t-test were used to analyze data. The results of the research are summarized as follow, the experimental group was more statistically signifi cant in learning achievement than the control group at 0.05 level.

Keywords: Information Technology

Introduction Cost Accounting is subject that is quite diffi cult. This Subject is calculated more than other subjects and Students should be trained to do a lot of homeworkTo achieve the skills to calculate costs. From the Teaching experience, some students had copied Homework from their friends, some copied were not Correct some student did not do homework because They could not do it. As a result, these students did not pass the examination. I think that in each week if students received homework andanswer sheet to practice at home, students may under-stand better than before. This method may be better than

the method that gave home work to students to do and when student came in class and listen the answers This is the reason that I need to research a comparison the academic achievement in subject Cost Accounting between students who have not homework answer sheet

and students who have answer sheet.

Objective 1. To know the academic achievement in

subject Cost Accounting between students who have not homework answer sheet and students who have homework answer sheet.

2. To know the attitude in subject Cost Account-ing of students who have not homework answer sheet

and students who have answer sheet.

Benefi ts are expected to receive The results are used as guidelines for teaching

in another method.

Hypothesis 1. The academic achievement of students who

have homework answer sheet is higher than students who have not homework answer sheet. 2. Attitudes of students who have homework answer sheet is higher than students who have not an-swer sheet.

Population The total population is students in 3rd year of Engineering Course of Business of the University of Fine Arts. Campus Sanam Jon Semester 1 , 2554, 100 students were divided into two groups, 1.Group no have

homework answer sheet 50 students (on Tuesday) and 2.Group have answer sheet 50 students (on Friday).

A Comparison of Learning Achievement Toward cost accounting I Between

The Group has Solution of Homework and The Group not has Solution of Homework.99Vol 32, No1, January-February 2013

Related Researchs Nori Boonyasilapa (2006:Abstract) conducted research. Development of the English language. The travel agency in Nakon Pathom for students grade fi ve . The study found that students' attitude was good. Tthe fi ndings showed. Students' attitudes towards learning a series of model. The content of the learning, activities and the implementation in all areas. Edward (1997:Abstract) studying innovative teaching strategies in class, grade 1 and grade 2, with a total of 175 people who returned to teach at home and in class. The variety of materials. The students are motivated to learn. Mac Donald (1973: Abstract) conducted a study Development and evaluation of a series of instructional multimedia. To determine the effect of the success of teaching methods and attitudes. Using a set of classes. The multimedia learning how to self teach simple. The experimental group students' attitudes towards teaching methods that are used in high school than the control group. Statistically signifi cant at the 0.05 level.

Tools used in this study 1. A plan of teaching subject Cost Accounting. 2. Homework answer sheet and tests. 3. A test of students' attitudes.

Research1. Researcher were conducted for 16 weeks, three

periods per week.2. Students who have not homework answer sheet, in

each week they will receive homework to practice at

home and have to submit on Friday. Students who have answer sheet they will receive homework to practice at home and have to submit on Monday.

3. In next week let students ask question and problem which they found in homework and then learn next lesson.

4. To compare the test scores of the two groups.

The experimental schedule (Group: No have answer sheet).

Mid-term Exam

No. Date Time

1 7 June2011 09.00 -12.00

2 14 June2011 09.00 -12.00

3 21 June2011 09.00 -12.00

4 28 June2011 09.00 -12.00

5 5 July 2011 09.00 -12.00

6 12 July 2011 09.00 -12.00

7 19 July 2011 09.00 -12.00

Final Exam

8 2 August 2011 09.00 -12.00

9 9 August 2011 09.00 -12.00

10 16 August 2011 09.00 -12.00

11 23 August 2011 09.00 -12.00

12 30 August 2011 09.00 -12.00

13 6 September 2011 09.00 -12.00

14 13 September 2011 09.00 -12.00

15 20 September 2011 09.00 -12.00

16 27 September 2011 09.00 -12.00

The experimental schedule (Group: Have answer sheet)

Mid-term Exam

No. Date Time

1 9 June 2011 13.00 -16.00

2 16 June 2011 13.00 -16.00

3 23 June 2011 13.00 -16.00

4 30 June 2011 13.00 -16.00

5 7 July 2011 13.00 -16.00

6 14 July 2011 13.00 -16.00

7 21 July 2011 13.00 -16.00

Final Exam

8 4 August 2011 13.00 -16.00

9 11 August 2011 13.00 -16.00

10 18 August 2011 13.00 -16.00

11 25 August 2011 13.00 -16.00

12 1 September 2011 13.00 -16.00

13 8 September 2011 13.00 -16.00

Manop Seeluang J Sci Technol MSU100

14 15 September 2011 13.00 -16.00

15 22 September 2011 13.00 -16.00

16 29 September 2011 13.00 -16.00

The statistics which using to analyze the data The statistics include the mean (x) standard deviation (SD) and the statistics are not independent (t-test Dependent). II. DATA ANALYSIS Table 1 Compares the academic achievement of subject Cost Accounting of students who have not homework answer sheet and students who have answer sheet.

Experiment Group

n x S.D. t Sig

No haveAnswer sheet

50 24.05 2.10

15.256 0.000

HaveAnswer Sheet

50 31.22 2.92

Table 1 shows that students who have homework answer sheet have higher academic achievement than students who have not answer sheet .The statistically signifi cant is at the 0.05 level, which is based on as-sumptions.

Table 2 Compares the attitude of 2 groups of students in subject Cost Accounting

Experiment Group

n x S.D. t Sig

No haveAnswer sheet

50 3.43 4.32

5.125 0.000

HaveAnswer Sheet

50 3.97 8.64

Table 2 shows that students who have homework answer sheet, their attitude are higher than students who have not answer sheet. The statistically signifi cant is at the 0.05 level, which is based on assumptions.

III. CONCLUSIONS 1. Students who have homework answer sheet have higher academic achievement than students who have not answer sheet. Because of the answer sheet helped students who did not understand well can do homework by themselves and encouraged them to train more homework and repeating and make them under-stand it better and faster. 2. Students who have homework answer sheet , their attitude in subject Cost Accounting is higher than students who have not answer sheet.

IV. ACKNOWLEDGMENT The author would like to thank Department of Mechanical Engineering and Department of Industrial Engineering and Management, Silpakorn University for all support to this work.

V. REFERENCES[1] Mac Donald.” conducted a study Development and

evaluation of a series of instructional multimedia”

Harvard Business Review,/25-34.[2] Nori Boonyasilapa.”conducted research. Develop-

ment of the English language” Dissertation Abstract International 52(September 2006):8-12.

[3] Edward studying innovative teaching strategies in class. Volume 42, Issue 3,Education Research Journal ,1997.doi:10.1016/S0378-7206(01)00101-X

Original

1 Faculty of Engineering and Industrial Technology, Silpakorn University (Sanam Chandra Palace), Nakhon Phathom 73000, Thailand. Email: [email protected].

Development of Temperature Controlling System in drying chamber by mixing air inlet

Sivapong Phetsong1


AbstractThis research was to develop a temperature controlling system in solar chamber, using the fl at-plate solar collector together with the working of blower with damper and heater to control the temperature. The microcontroller was used to process data: the temperature in the chamber and the temperature of the hot air and cold air at inlet ports, and then controlled the damper to optimize the air mixture at a constant speed of blower. This experiment used two identical chambers to compare the results from 3 systems: a chamber without controller, a chamber with PID controller, and a chamber with a basic controller, damper and heater. The results showed that the system with heater and damper could control temperature at 42 ºC. The chamber without controller had unstable temperature between 35 to 55 ºC. In addition, the temperature in the chamber with PID controller varied depending on the weather. However, the system with damper had the lowest power consumption.

Keywords: Drying solar chamber, Temperature controlling system, Microcontroller

IntroductionThere are several types of agricultural products in Thai-land. Fruits are one among the others of agricultural products. Some types of fruits grow dependently on season (e.g., longan) but some grow throughout the year (e.g., banana). In addition, many types of fresh fruits cannot be kept for a long time. Therefore, there

are many people tried to develop techniques to preserve those agricultural products, especially fruits, such as

drying, syruping. We founded that drying is the easiest and cheapest way to preserve products but there are still

issues of an uncertainty in sunlight intensity, the problem with dust, insects, etc.1

These problems lead us to use drying solar chamber, which can prevent the problem discussed above. However, the drying solar chamber in the present

day cannot control temperature as needed.2 In this research, we develop the chamber in which the temperature

can be controlled. The system is integrated by 3 parts: fl at-plate solar collectors, chamber and controller. We also install a heater to heat input air temperature when that solar collectors cannot collect enough power from sunlight. We also apply the mixture of cool air from ambient to cool down the air from solar collector when its temperature is too high. We use the constant speed blower.

Experimental Design Solar Chamber The components of the solar chamber are as

follows. - Solar collectors: a fl at plate V-shaped with a 3x1 square meters, covered with a glass thickness of

5 mm.3,4 - Two chambers covered with galvanized sheet with 0.5x0.70x0.75 cubic meters. They are side-by-

side assembled together but can work independently.

Vichuda Mettanant J Sci Technol MSU102

Figure 1 Solar chamber components

- Damper powered by 6-volt servo motor with 1500-watt heater installed in the front of damper to provide additional heat when necessary. - Blower with 12-cm diameter and use 12V DC power. It was setup to work at a constant speed, 80% of maximum speed. - Controller made up from AVR ATmega168 microcontroller and ds1820 temperature sensor, to facili-tate software development and cable wiring.5

Figure 2 Damper and heater installed position

Figure 3 Microcontroller circuit in the controller

The experiment was designed to use two iden-tical chambers to compare the results from the three systems: the control system with damper and heater, the chamber without a temperature control system but using a constant speed blower, and the temperature control system with PID control6,7 by controlling the temperature at 42 ºC. We use data logger to record the temperatures at various points as shown in Figure 4 and record power consumption of the system every 15 minutes interval.

Figure 4 Thermo-couple installed position

The controller software The controlling software is an important part of

the temperature control system in the chamber in which control all hardware in the system. The overall process is shown in fi gure 5. In our experiments, the software controls the blower to run at 80 percent of maximum speed. Then, it

will read the temperatures from sensors in the chamber and average them. If the temperature in the chamber is not equal to the set-point temperature, the controller will

adjust the angle of the damper to change the mixture proportion of hot-air and cold air. In the case of without enough sunlight, the hot air temperature is not high enough and is lower than the set-point temperature. The software will automatically run

the heater.

Daylight Performance of an Automated Vertical Blinds System 103Vol 32, No1, January-February 2013

Figure 5 Software fl owchart

Results and discussions From the experiment with 42 °C set-point temperature, we have the results as shown in graphs in fi gure 6 to fi gure 8 indicating an average temperature in chamber and time. We also found that the temperature in the chamber with damper and heater approached 42 °C with less than 1 °C error as shown in fi gure 6. This is a result of the system which can automatically adjust the inlet air mixer. The heater can increase temperature as needed. However, there is an approximately 1 °C error left because the system does not have any control condi-tion in the controller. For the system without temperature control

system, the temperature varies between 35-55 °C. The system with PID controller also has the same results. This is because the blower is too small for the system and the

ambient temperature changes continuously depending upon the weather variation.

Figure 6 Temperature in the chamber with damper and heater

Figure 7 Comparing temperature in the chamber with damper and

heater and uncontrolled chamber

By comparing the results between the chamber with controller and damper and the chamber with PID

control system as shown in fi gure 8, we fi nd the differ-ences in these results. The weather during 10:00 to 12:00

can be easily predicted that it rains. The temperature in the chamber with PID controller is always lower than

the set-point temperature. Also, during 13:00 to 15:00, it is sunny. The temperature in the chamber with PID controller is always higher than the set-point temperature

because the air leaving from solar collector is too hot. The blower cannot blow hot air off to make the temperature in-range. Unlike the chamber with damper and heater, it can control temperature better because the heater raises the temperature as needed during 10:00 to 12:00 and the

damper changes an air mixture to increase cold air during 13:00 to 15:00 to make the temperature in the chamber more stable.


Figure 8 Comparing temperature in the chamber with damper and

heater and the chamber with PID controller

When comparing the energy consumption of the chamber with damper but without heater and the chamber with PID controller, we found that they consume the same level of power with only a little bit different. The main reason of this comes from the blower in the chamber with damper which works at a constant speed while the blower in the chamber with PID controller can reduce speed for effi cient power consumption. However, when using the heater with the damper, we found that when the heater is working, it will consume much more power than the chamber with PID controller.

Figure 9 power consumption in chamber

Figure 10 Comparing power consumption every 15 minutes interval

Conclusion The controller with damper and heater can control the temperature in the range of 42°C with the error of less than 1 °C. If we compare the results with the other system, the system without temperature controller has the temperature between 35 to 55 °C. The temperature in the chamber with PID controller has the temperature in the same range as the uncontrolled chamber. However, when comparing the energy consumption, the system with the heater consumes power the most. In contrast to the same system without the heater, it consumes power the least but it may cause a problem when there is an absence of sunlight for a long period. In addition, if we develop the better software to control blower, it will be able to adjust its speed as the one used in the chamber with PID Controller. More energy will be saved.

References1. S. Chanchaay, “Solar Drying Machine”, Departments

of Physics, Faculty of Science, Silpakorn university,

2004.

2. J. Naga Raju, “An energy effi cient drying chamber for FCV tobacco curing process”, Solar & Wind Technol-ogy, Vol. 6, p. 159–163, 1989.

3. Salah A. Eltief et al, “Drying chamber performance

of V-groove forced convective solar dryer”, The

Ninth Arab International Conference on Solar Energy

(AICSE-9), Kingdom of Bahrain, 30 April 2007.


4. Azharul Karim and Hawlader M.N.A., “Performance investigation of fl at plate, v-corrugated and fi nned air collectors”, Energy, vol. 31, pp. 452-470, 2006.

5. E Makarn, “Understand AVR Microcontroller with Arduino”, ETT, Bangkok, 2009.

6. K. Sarawut, S. Phetsong, S. Pullthep, “Development of a temperature controlling system in solar drying chamber by using a microcontroller”, The 4th National Conference on Applied Computer Technology and Information Systems, Thailand, April 10, 2012.

7. K.M Lawtona and S.R Pattersona, “A high-stability air temperature control system”, Precision Engineering, Vol. 24, pp. 174–182, 2000.

Original

Design and Development of an Automatic Swine Feeding Machine Cooperated with Radio Frequency Identifi cation Technique

Saroj Pullteap1


AbstractIn this paper, an automatic swine feeding machine controlled by the radio frequency identifi cation technique has preliminary been designed and developed. The developed system is, generally, consisted of two main parts: input unit and controlling unit, respectively. The former part is operated as the tag detection system and also the weighting system, while the latter has, consequently, been employed for the machine controlling. The system operation is, fi rstly, used to check the electronics tag status that is entered to the input unit. Consequently, the tag information (registered tag) has then been identifi ed and displayed on a LCD device under controlled by a microcontroller unit (MCU). Moreover, the swine’s weight achieved from the weighting system has also been detected and then displayed the weight value on the LCD. This data is, generally, used for controlling the machine to load the swine feeding. By investigating the 100 registered tags and 50 of non-registered tags over 20 times of repeatability, the corrected results of 100% can be exploited. Otherwise, the accuracy and precision of the weight detection system has also been investigated. By exploiting the swine’s weight from the range of ~500 - 10,000 grams over 20 times, a percentage of accuracy and precision can thus be resulted of 99.27% and 98.3% respectively.

Keywords: Electronics tag, radio frequency identifi cation technique, microcontroller, automatic feeding machine

IntroductionNowadays, the world population has been increased. We could say that every 5 minutes more than 1,000 babies are born. Absolutely, the basic requirement of food is important. It might, therefore, be proportioned to the population number. Moreover, the quality of food is, however, a signifi cant factor that has always been con-sidered and improved by the producer due to reliability

and trustworthiness to the consumers or wholesaler etc. Pork is an example of a raw material that is often used for making the favorite food. Thailand is the one of sev-eral countries that exports the pork to the international markets such as Hongkong, China, India, and Japan

etc.1 However, there is several conditions have thus been

considered and cared before exporting i.e. pork’s color, foot and mouth disease, and also parasite etc. To eliminatesuch problem, the swine’s herdsman or the producer has to think over in every procedures of swine’s nurtur-

ing. The using of technology might thus be an answer to solve those problems and also very helpful to the producer. The computer technology, Electrical & Electronic technology and also Mechanical technology are examples

of the technology for reducing those effects. The radio frequency identifi cation (RFID) technique is a type of the

electronic technology that could thus be applied to use for the animal nurturing. Several publication papers of the technique have, consequently, been published in the recently years ago. For example, Trevarthen et al2 uses the RFID technique for investigating and controlling the cow’s weight at the cattle farm in Australia. Moreover,

Garfi nkel3 was described some concepts of the RFID theories and its applications in science technology and engineering fi elds. Several advantages of the technology

can, generally, be exploited in terms of the reduction in production cost, labor number, production errors, and also improve the product quality, etc.

1 Assist. Prof., Faculty of Engineering and Industrial Technology, Silpakorn University (Sanam Chandra Palace), Nakhon Phathom 73000,

Thailand. Email: [email protected].

Design and Development of an Automatic Swine Feeding Machine Cooperated

with Radio Frequency Identifi cation Technique107Vol 32, No1, January-February 2013

In this paper, an automatic system based on RFID technique has preliminary been applied to operate in the swine’s nurturing. The developed system has been composed of two main parts of as followed by hardware system and software system respectively. The fi rst part is, however, divided into four sub-sections; reader/identifi cation system, weighting system, controlling system, and feeding machine, while the latter part is, consequently, operated for two purposes; identifi cation the tag’s information, and indication the weight’s information on a LCD device. The tag information is, generally, detected and data transferred by using a RFID reader. This data is then processed and displayed its information such as tag’s name, weight, and time duration to control the feeding machine on the LCD device by using a microcontroller (MCU). Moreover, the MCU is also used to control a stepping motor that has been composed to the feeding machine for feeding or non-feeding. However, it might imply that the feed quantity is, normally, proportioned to the weight value that is meas-ured by the weighting system. Finally, the aim of this work is obtaining an automatic system for the swine’s feeding machine. It leads to the reduction of labor number in the swine’s farm, capability to monitor the swine’s health, and also improving the pork quality.

Briefs of RFID Technology The radio frequency identifi cation or called as “RFID” is a smart technology, which has been employed for identifying the object appearance. It was developed since 1980s for instead of the barcode system4. Basically, it consists of two parts: reader system and electronics tag,

respectively. The electronic tags is, generally, produced in the forms of a smart card that some electronics data such as ID code, user name, registered status, has, directly, been recorded on the tag5. These data are, normally, demodulated and transferred to the reader system.

Several advantages of the system over the barcodesystem are exploited in terms of non-contact device, simplify to operate and, capability to identify in multiple

tags, high speed processing, small size, etc.6-7. Many applications, currently, use the technology for security

system, product identifying, and animal monitoring etc.8

As mentioned before, the RFID system has been composed with the electronic tag and the reader system. In general, the tag would often be attached to the desired objects such as goods, animal body, and/or security card. This system is, currently, resembled to the barcode system. However, the advantages and disadvantages between the RFID technology and barcode system can thus be summarized in table 1.9

Table 1 Comparison details between barcode versus RFID system

Descriptions Barcode system RFID system

Demodulation tech-

nique

Image encoding

technique

Radio frequency

technique

Distance to de-modulate

< 10 cm. > 10 cm.

Number of target to demodulate

1 barcode / time 60 tags / time

Type of identifi cation Text and Numbers Frequency

Visibility Need direct line of sight to read

No need direct line of sight to

read

In this work, the RFID system has been operated for the swine’s feeding. The structure of RFID system can, basically, be shown in fi gure 1.

Figure 1 Structure of RFID reader system

There are four important signals has been required for the reader system that is followed by +Vcc, Gnd, TXD, and RXD respectively. The power requirement

of the system is, generally, +5 volts, while the transmitted and received signals from the RFID tag are denoted as TXD, and RXD respectively. These signals are, normally, used for data interchanging between the RFID tag and

Saroj Pullteap J Sci Technol MSU108

the reader system. If the tag is, however, registered into the database system, the tag information would be de-modulated by the reader system.

dsPIC30F4011 Micro-controlling Unit In this work, a dsPIC30F4011 microcontroller (MCU) fabricated by the Microchip Inc. has been employed for data transferring from the RFID tag to the reader system10. Moreover, it’s also used for controlling a stepping motor to switch on/off a feeding machine. A brief of the MCU specifi cation can thus be detailed as a 16 bits microprocessor, which has a digital signal processing inside. Consequently, it also consists of the 10 bits of analog to digital convertor (ADC) with a sampling frequency of 500 KS/s has been generated. Furthermore, there are six input/output ports (I/O) within the MCU part. In this work, a microcontroller embedded board “model ET-BASE ds PIC30F4011” has, preliminary, been used as a processing system. A structure of the board can, therefore, be illustrated in fi gure 2.

Figure2 ET-BASE dsPIC30F4011 microcontroller board used as

processing unit

The advantages of the dsPIC30F4011 micro-controller over the other MCU families such as MCS-51,ARM7, or FPGA are exploited in terms of low cost, easily to implement, build-in DSP module, and high speed clock etc.11 However, the specifi cation details and also

the interfacing diagram would then be explained in the next section.

Experimental Setup As mentioned above, the developed system can be classifi ed into three main parts; input, processor and

output sections respectively. The signal from the RFID tag is fi rst scanned and demodulated by the reader system at the input section. This signal is then transferred to the

microcontroller section for checking and monitoring the tag status or information from the database system that is stored in the internal memory at the embedded sys-tem. Moreover, the sensing signals from a strain gauge (included in the weighting system), which its maximum amplitude has been amplifi ed to 5 volts by a signal con-ditioning circuit, are also transferred to the MCU as input signal for demodulating the swine’s weight. These signals; tag and sensing signals, have thus been data processed at the processing part and then send to the output part for controlling a stepping motor and also displaying the processed information on a LCD device. A diagram of the hardware system can be summarized in fi gure 3.

Figure3 A structural diagram of developed system in hardware

section

An overview of the hardware section as men-tioned above can, precisely, be illustrated the interfacing details between the micro-controlling unit and peripheral devices in fi gure 4.

Figure 4 An interfacing diagram between MCU and peripheral

devices of developed system



There are three interfacing pins; tag status, TXD and RXD, of the RFID reader have been used to connect to the microcontroller via the RD3, RF2, and RF1 ports. These ports have thus been operated for transmitting-receiving the serial data, and also transferring the tag information to the MCU section. Furthermore, the sensing signal from the input strain gauge after amplifi ed by the instrumentation amplifi er circuit has next been sent to the MCU via the RB7 port. This signal is then processed and used to control the stepping motor at the RD0 and RD1 ports respectively. In addition, the output signal would also be displayed at the LCD module via the RB0 – RB5 ports. Moreover, an application program from Cprogramming has, therefore, been developed as the controlling system. A fl owchart of the developed program can thus be illustrated in fi gure 5. In preliminary process of the developed system, any tags which are supposed as the swine in the farm have, fi rstly, been registered theirs information, such as ID number, name, sex etc., into the database system. Consequently, the program is then checking the entered tag status that is scanned through the reader system. If the input tag is, however, found some information on the database system after scanning, the processing unit will demodulates the input data and then display the tag information on the LCD device. On the other hand, if the input tag is not registered,the developed program will send a message of “No tag registered” on the LCD device.

Figure 5 A fl owchart of developed program

The weighting system is next operated for meas-uring the swine’s weight and then displayed the weight’s value on the LCD. Moreover, this value is next transferred to the microcontroller for generating a controlling signal to control the stepping motor. Consequently, the time duration to switch on/off the motor is also depending on the measured value.

Experimental Results and Discussions The structure of developed system can be shown

in fi gure 6.

Figure 6 Structure of swine’s feeding machine

Saroj Pullteap J Sci Technol MSU110

To verify the performance of the developed system, few experiments have been studied. The fi rst experiment is testing the demodulation signals from the RFID reader. By using a hundred registered tags and fi fty of the non-registered tags, which has thus been entered to the reader system over 20 times per each, we found that the automatic feeding machine can correctly identify all tags. This implies that the developed system has an accuracy of 100%. The output signal from the MCU can, consequently, be shown some results in fi gures 7. These results display the examples of three different tag statuses; no tag detected, no registered tag detected, and registered tag detected, respectively. In fi gure 7(a), the output message from the LCD module is displayed when there is no tag detected into the system, while fi gure 7(b) shows that there is a tag has been scanned into the developed system, but it is no data registered. Moreover, the output information (ID number and weight’s value) as shown in fi gure 7(c) are indicated on the LCD, when there is a registered tag has been scanned into the system.

Figure 7 Experimental results for tag detection on LCD device (a)

no tag detected, (b) with tag detected but without regis-

tered tag, and (c) with tag detected and with registered

tag

The second experimental process is investigat-ing the performance of weighting system. By using the weight’s value from ~500 to 10,000 grams, the results over 20 times of repeatability can thus be summarized in table 2. We found that a maximum percentage error has been obtained of 3.64%, while a minimum error is 0.27% respectively. These caused to three important factors; un-stability of the strain gauge to obtain an exact value, conversional error, and gross error respectively. The un-stability error has been generated when the temperature, and also the environmental vibration changed, while the conversion error has, therefore, been exploited due to missed conversion of the ADC part for converting a small input analog signal to digital signal. For example, if the small signifi cant value of input signal changed, it might not be changed to the digital output signal, leads to meas-urement error occurred. The last error effect is, normally, happened when the user has to be undisciplined to the measured system, and missing understand or unknown to the measurement system. Otherwise, an accuracy and precision of the experiment has also been investigated. From the results as summarized in table 2, we found that the weighting system has an accuracy of 99.27%, while a precision of the system is 98.3% respectively.

Table 2 Summary of weight investigation in range of ~500 - 10,000 grams over 20 times of repeatbiliy

Average Standard de-viation

Error bar % Error

509 25.95 18.56 3.64

1,003 15.18 10.86 1.08

2,025 45.24 32.36 1.60

3,016 30.57 21.86 0.72

4,056 61.09 43.70 1.07

5,042 36.14 25.85 0.51

5,952 42.51 30.41 0.51

7,014 56.36 40.32 0.57

7,972 42.99 30.75 0.38

9,025 38.25 27.36 0.30

9,976 37.81 27.05 0.27



In last experiment, the performance of controlling system has been studied. By dividing the weight condition into 2 levels; lower and upper levels, the controlling signal generated from the MCU has been used to switch on/off the stepping motor. However, the time duration to switch is, normally, corresponded to the feeding quantity. Moreover, it can be assumed that the lower level has a weight range of ~500 - 5,000 grams, while the latter are approximately 5,500 - 10,000 grams. In the experiment, we attempt to enter the input weight from ~500 to 10,000 grams into the weighting system and then measure the output signal from the feeding ma-chine. We found that the feeding machine can, correctly, be switched on/off the stepping motor by relating to the input weight. This implies that there is no error occurred. It should, therefore, be emphasized that the controlling system has, excellently, be operated on the desired func-tions/procedures as the developed system needed.

Conclusion An automatic swine’s feeding machine cooper-ated with the RFID technique has been developed in this work. The system was consisted of three main parts; RFID reading system, weighting system and controlling system, respectively. The fi rst part was employed for tag detection. The experimental results were shown that a 100% of the developed system has capability to detect the RFID tags in both statuses; registered and non-registered tags. In second part, the weighting system controlled by the microcontroller has consequently been investigated. By investigating the swine’s weight in the range of ~500 - 10,000 grams over 20 times of repeatability, a minimum

and maximum percentage error were obtained of 0.27% and 3.64% respectively. These lead to an accuracy and a precision of 99.27% and 98.3% achieved. The controlling

system developed by the microcontroller cooperated with C programming was next studied. This part, in generally, was used to generate a controlling signal to control the stepping motor for the feeding machine. The signal was related to the tag information and also swine’s weight.

By investigating the 100 registered tags with the weight’s range of ~500 - 10,000 grams over 20 times of repeat-

ability, a 100% of accuracy was exploited. This implies that the developed system has excellent performance and reliable for applying to operate in the real swine’s farm. Last but not least, the developed system have had three important factors to obtain the measurement error; unstable of the sensing signal due to low quality of the output signal from the strain gauge, gross error by the user, and also conversional error from the ADC part inside the micro-controlling unit, respectively. To improve or eliminate such problem, several factors such as environmental controlling, using of a signal conditioning circuit, comparing the measured data with the reference data from a commercial instrument, and also using of the weighting standard, have signifi cantly been offered for operating to the system.

References1. Centre for Agricultural information. Thailand foreign

agricultural trade statistic 2009. Offi ce of Agricultural Economics; 2009.

2. Trevarthen A, Michael K. The RFID-enabled dairy farm: towards total farm management. Proceeding of the 7th International Conference on Mobile Business; 2008 Jul 7-8; Barcelona, Spain; p. 1-10.

3. Garfi nkel S. RFID: applications, security, and privacy. Addison-Wesley Professional; 2005.

4. Glover B. RFID essentials. O’Reilly Media; 2006.5. Li D, Chen Y. Computer and Computing in Agriculture

V. Springer; 2012.6. Xu M. Managing strategic intelligence: technique and

technologies. IGI Global; 2007. 7. Roper AT et al. Forecasting and management of

technology. John Wiley & Sons Inc; 2011. 8. Hunt VD, Puglia A, Puglia M. RFID: a guide to radio

frequency identifi cation. John Wiley & Sons Inc; 2007.

9. Prakashan N. Emerging trends in information technology. Rachana Enterprises; 2007.

10. Boukas EK. Sunni FM. Mechatronic systems:analysis, design and implementation. Springer; 2011.

11. Wilmshurst T. Design embedded systems with PIC

microcontroller: principles and applications. Elsevier Ltd; 2010.

Original

1 Assoc. Prof., Doctor of Business Administration Program, Ramkhamhaeng University, Bangkok 10240, Thailand. Email: [email protected] Assist. Prof., Department of Statistics, Ramkhamhaeng University, Bangkok 10240, Thailand. Email: [email protected].

The Structural Relationship Model of Factors Affecting Perceived Shared Value for Intention to Stay

Napaporn Kantanapa1, Montree Piriyakul2, Dechaphan Ratsasanasart3


AbstractThis research aims to study the structural relationship of shared value which affected to the intention to stay in staffs from four major banks in Thailand. The correlated factors consisted of perceived shared value, organizational trust, cooperation, job satisfaction, and intention to stay. The subjects were 500 employees who work in four major banks. The researcher collected data by questionnaire and calculated the mean, S.D. and analyzed the structural equation modeling by the statistical software. From studying the structural relationship model, the researcher found that the level of perceived shared value, organizational trust, cooperation, satisfaction, and intention to stay were extensively high. For structural equation analysis, it was found that the outcome of relationship was consistent with hypothesis; the perceived shared value directly affecting to organizational trust at highest level (0.517), followed by organizational trust which affects to cooperation (0.410), job satisfaction which affects to intention to stay (0.297) and cooperation which affects to job satisfaction (0.259)

Keywords: perceived shared value, intention to stay

Introduction

Figure 1 Top three industries of the highest turnover rate

According to the result of survey in 2010-2011 by Hewitt Associates (Thailand) Ltd1, to 189 business fi rms in Thailand and South-east Asian, the average of employee

turn-over rate was 8.0 %. The major sectors were (24.9 %), fi nancial services (12.2 %), and IT services (12.0%). As country-wise, it was found that Vietnam had the

highest turnover rate (14.8 %), followed by Malaysia (13.0 %), Singapore (10.0 %), and the Philippines (9.5 %).

For Thailand, it was ranked as the fi fth country (8.0 %). One remarkable fi nding from this survey was that the management of most companies had continuously implemented a wide variety of interventions in order to decrease the turnover rate as well as sustain employee engagement. Furthermore, one of key successful factors was employee recognition as Figure 1. In 2015, the Association of Southeast Asian Nations (ASEAN) will transform to the ASEAN Community which is comprised of three main pillars, namely ASEAN Political Security Community (ASC), ASEAN Economic

Community (AEC), and ASEAN Socio-cultural Community (ASCC). For the second, in particular, AEC envisages to single market and single production base which focuses on free movement of goods, services, investment, and labor. Also, the ultimate goal and framework towards

agreement on free trade had been established to facilitate the movement.

The Structural Relationship Model of Factors Affecting Perceived

Shared Value for Intention to Stay113Vol 32, No1, January-February 2013

Since the free movement within region had been more easily taken place, the workforce can freely move from one country to another. Furthermore, there will be the loss of skilled intellectual and technical labor or “hu-man capital fl ight” which causes from signifi cant drivers organizational trust, employee engagement, satisfaction in job profi le, reward, career path, colleagues or prospect for the better well-being in other countries. Therefore, most organizations will have spent a large number of money in investing to human resource maintenance e.g. replacement recruitment, training and development to newcomers so that they can perform roles as effectively as the resigned staff, etc. Moreover, they necessarily reserves extra budget to setting up additional motives such as special merit increase, bonus, incentive pay, profi t sharing, or employee stock option programs, etc. so as to retain employee to contribute to company through relentless efforts and effective and effi cient practices. From literature review, the researcher found that the banking and fi nancial services has been built for many years and gained organizational trusts from the clients. Furthermore, branches are located throughout country; therefore, banks become the main sources of savings and investment funds in Thailand and considerably connect to Thailand’s economy. To prevent these issues, the researcher focusedon studying the appropriate structural relationship of perceived shared value which affects to employee satisfaction leading to employee retention in major banks. Remarkably, this will aim to competitive advantage and sustainable growth in Thailand. This research aims to

study following factors which signifi cantly infl uence to employee retention, level of intention to stay of employees in Thailand’s major banks and structural relationship of perceived shared value, organizational trust, cooperation, job satisfaction, and employee’s intention to stay.

Figure 2 McKinsey’s 7-S Model

The organizational theory which was mainly used in this research to study the structural relationship of perceived shared value affecting to employee’s intention to stay was McKinsey’s 7-S Model (7-S Model) by Peter and Waterman2 In this theory, the organizational effec-tiveness was caused by systematically interconnected seven elements strategy (corporate strategy), structure (organizational structure), system (operational system), staff (human resources), skill (competency), style (man-agement style), and shared values (corporate values). All can be divided into two categories: hard skills (strategy, structure, and system) and soft skills (staff, skill, and style). According to McKinsey, hard skills can be prac-tically identifi ed and directly manageable e.g. corporate strategic planning, organizational design, and operational system setting. In contrast, soft skills are more compli-cated to identify and are more considerably affected by

shared value and organizational culture. For organiza-tional development, it requires both two clusters so as to achieve corporate goals and sustain organizational growth. To sum up, all seven factors are signifi cantly

interrelated then need for mutual changes and alignment as McKinsey’s 7-S Model in Figure 2.

Kantanapa et al. J Sci Technol MSU114

Figure 3 Hypotheses and framework Remark: PSV = perceived

shared value, TRUS = organizational trust, COO = co-

operation, JST = job satisfaction and STAY = intention

to stay.

According to research on studying shared values which affect to employee engagement in bankingemployees, indicated that the participants had the high level of employee engagement, shared value, and co-operation, Furthermore, the age factor had signifi cantly affected to employee engagement. Additionally, working position factor had signifi cantly affected to employee engagement in driver of intention to stay. Next, com-pensation and reward system had signifi cantly affected to employee engagement in driver of trust to corporate goals and shared values. Furthermore, organizational culture in driver of team-working had signifi cantly affected to employeeengagement in driver of intention to stay and commitment to contribute to organization. In conclusion, the overall employee engagement was signifi cant. Also, the belief in corporate goals and shared values in driver of commitment to contribute to organization were signifi cant3-4All hypoth-eses are framed as Figure 3.

Research Approach Subjects The subjects in this research were employees who work in four major banks in Thailand Bangkok Bank

PLC., Siam Commercial Bank PLC., Kasikorn Bank PLC., and Krungthai Bank PLC. The researcher followed the data collection and sampling to Comrey and Lee5 which specifi ed the sample size and set a very-good number of subjects at 500 participants.

Data collection method The questionnaire was used and designed by using Likert scale. The data interpretation was done by using Best6 the criteria as follows Mean score = 4.50-5.00 means the needs item was at highly extensive level Mean score = 3.50-4.49 means the needs item was at extensive level. Mean score = 2.50-3.49 means the needs item was at moderate level. Mean score = 1.50-2.49 means the need item was at fair level. Mean score =1.00-1.49 means the needs item was at low level. In calculation, the researcher used mean S.D. and t-statistic .The researcher analyzed fi ndings and de-veloped the Structural Equation Modeling (SEM) by the statistical software.

Table 1 Mean and S.D

construct mean S.D Result

PSV TRUSCOOJST

STAY

3.954.294.333.96

3.81

0.460.420.400.460.95

ExtensiveExtensive ExtensiveExtensiveExtensive



Figure 4 Result of analysis.

According to table 1, the mean of perceived shared value was 0.46, that of organizational trust was 4.29, that of cooperation was 4.33, that of job satisfac-tion was 3.96, and that of intention to stay was 3.81. All items were ranked at extensive level and the mean of cooperation was the highest (4.33) as Figure 4.

Table 2 Result of analysis for affecting.

variable R2 effect antecedents

PSV TRUS COO JST

STAY 0.551 DEIETE

0.0000.0660.066

0.0820.5030.585

0.0000.0770.777

0.297**0.0000.297

JST 0.726 DEIETE

0.0000.6560.656

-0.0570.5470.490

0.259**0.0000.259

N/AN/AN/A

COO 0.401 DEIETE

0.0080.1380.146

0.0510.1920.243

N/AN/AN/A

0.259**0.0000.259

TRUS 0.267 DEIETE

0.517**0.0000.517

N/AN/AN/A

0.410**0.1920.602

-0.0570.6100.553

When considering the infl uence level of anteced-ents of factors both external endogenous variables and internal endogenous variables, it was found that PSV, TRUS, COO, and JST had signifi cant affects to STAY. Furthermore, TRUS had both direct and indirect affects to STAY. Next, it was seen that PSV, TRUS, and COO had signifi cant affects to JST and TRUS had affect at the highest level both directly and indirectly while PSV had only indirect affect and COO had only direct affect. For COO, it was found that PSV, TRUS, and JST had affects to COO. In particular, both PSV and TRUS had at the highest levels since these have both direct and indirect effects whereas COO and JST had only direct effects. In addition, it could be seen that PSV, COO, and JST had effects to TRUS. Especially, both COO and JST had at the highest levels since these have both direct and indirect effects while PSV had only indirect effect.

Table 3 Result of testing to research hypotheses

Hypothesis (p) t-stat Output

H1: Perceived shared value af-fects to organizational trust.

0.517 17.130** Support

H2:Organizational trust affect

to cooperation.

0.410 8.678** Support

H3:

Cooperation affects to job

satisfaction.

0.259 9.643** Support

H4:Job satisfaction affects to

employee’s intention to stay.

0.297 5.767** Support

Table 3 illustrated as follows:

H1: Perceived shared value affects to organiza-tional trust. It was found that the perceived shared value affected to organizational trust. The p was 0.517 and t-stat was 17.130. This supported hypothesis at the 0.01 level of signifi cance.

H2: Organizational trust affects to cooperation. It was found that the organizational trust affected tocooperation. The p was 0.410 and t-stat was 8.678. This

supported hypothesis at the 0.01 level of signifi cance.

Kantanapa et al. J Sci Technol MSU116

H3: Cooperation affects to job satisfaction. It was found that the cooperation affected to job satisfaction. The p was 0.259 and t-stat was 9.643. This supported hypothesis at the 0.01 level of signifi cance. H4: Job satisfaction affects to employee’s inten-tion to stay. It was found that the job satisfaction affects to employee’s intention to stay. The p was 0.297 and t-stat was 5.767. This supported hypothesis at the 0.01 level of signifi cance.

Table 4 Result of analysis for quality of structural relation-ship

Construct Average

Communality

Average

redundancy

R2

Perceived shared value

0.759 0.000 0.000

Organizational trust 0.643 0.199 0.310

Cooperation. 0.734 0.299 0.408

Job satisfaction 0.624 0.463 0.742

Intention to stay 0.798 0.455 0.670

Average 0.711 0.383 0.544

GoF (goodness of fi t)

0.622

For quality of factors, the researcher used R2 in order to consider which level primary factors signifi cantly affect variables and how many average communality were in each. The average communality is used to explain how construct refl ects to average redundancy and whether the independent-variable construct can refl ect to other dependent variable construct. For goodness of fi t (GoF), it is used to represent at which level SEM affects to all constructs. Typically, the GoF should necessarily be high to signifi cantly relative to R2

From table 4, it was seen that the average communality was ranged between 0.6243 – 0.7982. This means that endogenous variables were relatively signifi cant to constructs. Furthermore, average redun-dancy was ranged between 0.1999 – 0.5048 and that of perceived shared values was lowest score. This implies that construct which were independent variables in SEM signifi cantly affected to dependent variables. For R2, it could be seen that the average was between 0.3106 –

0.7420 and this was high but acceptable since this was higher than 0.25. At overall, SEM defi nitely refl ected to constructs as GoF was extensively high (0.622)

Conclusion According to research fi ndings, as for structural relationship of factors which affect to the perceived shared value, all relatively affected as hypotheses at the 0.01 level of signifi cance: the perceived shared value affected to organizational trust; the organizational trust affected to cooperation; cooperation affected to job satisfaction; and job satisfaction affected to intention to stay. This implies that in order to retain the potential employees, the mutual perception towards organizational direction and policies are importantly established via perceived shared values. Moreover, the company necessarily gains reliability from all employees by following methods: sharing information; building positive perception towards organization; reviewing job description; implementing the effective performance management system; developing career development plan and talent management; creat-ing coaching culture; and building the team synergy, etc. Recommendations for further research, this research had limitations in subjects. Therefore, the sub-jects should be extended in terms of sources and type of institutions and more specifi ed e.g. operation staffs vs. managerial staffs. Other factors should be also studied because these might affect to employee’s intention to stay. Also, theories and management models can be used in order to gain the more reliable outputs and results.

References1. The average of employee turn-over rate of survey

in 2010-2011, Aon Hewitt (Thailand), 20112. Peter and Waterman, Structure is not organization,

Business, Horizon, pp.14-26, 1980.3. Kanya Rodpitak, 2008, Organizational Culture

Affecting of Organization Commitment Employee’s

the Government Savings Bank Regional 3 Siriraj Zonal Branch. The Graduate School, Dhonburi Rajabhat University.



4. Nathakorn Torchote, 2011, A Study of Organizational Commitment of Personal in Government Saving Bank Regional Offi ce 12. The23rd National Graduate Research Conference, The graduate School, Kaset-sart University.

5. Comrey, A.L. and Lee,H.B.,1992, A fi rst course In factor analysis, Hillsdale, New Jersey: Erlbaum.

6. Best, John W. Research in Education. 5thed.Eglewood Cliffs : Prentice-hall, Inc., 1981.

Original

1 Department of Mechanical Engineering, Faculty of Engineering and Industrial Technology, Silpakorn University (Sanam Chandra

Palace), Nakhon Phathom 73000, Thailand. Email: [email protected].

The Effect of Equivalent Ratio on the Performances and Emissions of Diesel Engine

Thibordin Sangsawang1, Supachai Wasananon and Aminan Amornpattarakiat


AbstractThe research studies about the effect of equivalent ratio on the performances and emissions of diesel engine. The engine is the 2L-II Toyota diesel engine, which is tested by using diesel and biodiesel fuel (B10). An engine is tested at 2,400, 3,000, 3,500, 4,000 and 4,200 rpm respectively. The study presents that engine power of diesel and biodiesel fuel increased with increase in equivalent ratio. The lowest brake specifi c fuel consumption and the highest fuel conversion effi ciency occur at equivalent ratio close to unity. For measuring an exhaust gas, the highest nitrogen oxide occurs at equivalent ratio close to unity. The hydrocarbon and carbon monoxide is higher with increase of equivalent ratio.

Keywords: Equivalent ratio, Performances, Emissions, Biodiesel, Diesel

IntroductionThe transportation vehicles are powered by diesel engines have increased and likely to increase several fold in the next decade. This situation generates the increase of the carbon monoxide and carbon dioxide and releases to the atmosphere. It is cause of the Green House Gases and leads to the critical issue of climate change. From the highly demand fossil fuel, while the resources of the fuel are decreasing, so the biodiesel fuel is used as an alternative fuel in diesel engines. Many

researchers1-5 studied the optimal blending ratio between the diesel and biodiesel, and reviewed effect on engine

performance and emissions. Altun et al.2 used sesame oil (B50) in IDI engine. The experimental results showed

that the power and torque of the sesame oil–diesel fuel were closed to the values obtained from diesel fuel and the amounts of exhaust emissions were lower than those of diesel fuel. Similarly, various oils such as Karanja vegetable oil3, Marula oil4, Cottonseed oil, Soybean oil and Sunfl ower oil5. In addition, the studies of the effect of different parameters such as compression ratio, injec-tion timing, injection pressure that affected on engine

performance and emissions as well. Sayin and Gumus6 found that compression ratio, injection timing and injection

pressure signifi cantly effected on the engine performance and exhaust emissions of a DI engine using biodiesel-blended diesel fuels. Raheman and Ghadge7 found that the differences of BTEs between diesel and mahua oil were also not statistically signifi cant at engine settings. Sangsawang et al.8 suggested that 10% of crude palm oil in diesel was the best for a small, single cylinder diesel engine by the performances and emissions. One of the engine parameters that effect to the performance and emissions is equivalent ratio. This parameter can be defi ned as “the ratio of the actual fuel-to-air ratio to the stoichiometric fuel-to-air ratio”. In this studied, the effect of equivalent ratio on the performances and emissions of diesel engine using diesel and biodiesel fuel (B10) are examined.

Experimental apparatus and procedure The experiments were performed on a Toyota Hilux Mighty-X (2L-II), four cylinder, four-stroke, water-

cooled, direct injection diesel engine. The basic specifi -cations of the engine are given in table 1. It is equipped with a water brake dynamometer STUSKA XS-111 to

investigate the engine power. The schematic of theexperimental apparatus is shown in fi gure. 1.

The Effect of Equivalent Ratio on the Performances and Emissions of Diesel Engine 119Vol 32, No1, January-February 2013

Table 1 Technical specifi cations of the test engine.

Model of engine Toyota Hilux Mighty-X

(2L-II)

Type OHC, 4 stroke, water-cooled,

diesel engine

Cylinder number 4

Bore 92.2 mm

Stroke 92 mm

Displacement 2,446 cc

Compression ratio 22.2:1

Maximum power 61 kW at 4,200 rpm

Maximum torque 165 Nm at 2,400 rpm

Figure1 The schematic diagram of the experimental setup

In the tests, diesel fuel and 10% of biodiesel

fuel from used cooking oil (B10) were used. The engine was tested at full throttle condition and engine speeds of

2,400, 3,000, 3,500, 4,000, and 4,200 rpm respectively. The speed and load were recorded by digital indicator. A

Testo 350 XL model gas analyzer with special probe was used to measure carbon monoxide (CO), hydrocarbon (C

xH

y) and nitrogen oxide (NOx).

Results and discussion After the engine reached the stabilized work-ing condition for each test, engine power, brake specifi c fuel consumption, fuel conversion effi ciency and exhaust gases emission were investigated on the engine using diesel and biodiesel fuel. Engine power From fi gure 2a and 2b shows that the engine power is increased with the higher of the equivalent ratio due to the greater of torque from the fuel consumption. Maximum engine power occurred at the equivalent ratio of 1.0 to 1.2 for both fuel and not signifi cantly different can be found from the fi gure 2a and 2b Brake specifi c fuel consumption (BSFC) Brake specifi c fuel consumption is the ratio of the fuel fl ow rate to engine power. From fi gure 3a and 3b in engine speeds of 2,400, 3,000 and 3,500 rpm, the BSFC is lower with the greater of equivalent ratio while decreased with the higher equivalent ratio at the engine speed of 4,000 and 4,200 rpm. The fi gure 3 shows that the best BSFC occur at equivalent ratio close to unity due to the stoichiometric combustion. The diesel and biodiesel mode present the slightly different BSFC. Fuel conversion effi ciency (η

f)

Fuel conversion effi ciency is inversely propor-tional to BSFC. From fi gure 4a and 4b in engine speeds of 2,400, 3,000 and 3,500 rpm, the η

f decreased with

the higher of equivalent ratio due to the BSFC increase cause of incomplete combustion. Therefore, η

f can be

reduced while the BSFC increased in the engine speeds of 4,000 and 4,200 rpm. As mention above, the best η

f

occurs at the equivalent ratio close to unity due to the

stoichiometric combustion.

Sangsawang et al. J Sci Technol MSU120

Hydrocarbon emissions (CxH

y)

The CxH

y is increases with the rich fuel condi-

tions which equivalent ratio is greater than unity due to the incomplete combustion as appeared in the fi gure 5a and 5b. The C

xH

y is not signifi cantly difference between

diesel and biodiesel.

Carbon monoxide emissions (CO) CO would be generated from the incomplete combustion as C

xH

y. The amount of O

2 in the combus-

tion chamber is defi ciency to combust completely. The equivalent ratio is more than unity. As present in fi gure 6a

and 6b, the CO is increase with the higher of equivalent

Figure 2 Engine power & equivalent ratio (a) Diesel fuel, (b)

Biodiesel fuel

Figure 3 Brake specifi c fuel consumption & equivalent ratio (a)

Diesel fuel, (b) Biodiesel fuel

ratio as mentioned. The CO from biodiesel is less than that of diesel because of the oxygen in the composition

of biodiesel and make the more complete combustion than those of diesel.

Nitrogen oxide emissions (NOx) The high temperature and plenty of O

2 in the

combustion process will generate NOx. The change in the NOx emissions is displayed in the fi gure 7a and 7b. The fi gure shows that the highest value of NOx is at equiva-lent ratio close to unity due to the maximum combustion effi ciency. The NOx from the engine using diesel and

biodiesel are not signifi cantly difference on comparison.

The Effect of Equivalent Ratio on the Performances and Emissions of Diesel Engine 121Vol 32, No1, January-February 2013

Figure 4 Fuel conversion effi ciency & equivalent ratio (a) Diesel

fuel, (b) Biodiesel fuel

Figure 6 Carbon monoxide emissions & equivalent ratio (a) Diesel


Figure 5 Hydrocarbon emissions & equivalent ratio (a) Diesel fuel,

(b) Biodiesel fuel

Figure 7 Nitrogen oxides emissions & equivalent ratio (a) Diesel


Sangsawang et al. J Sci Technol MSU122

Conclusions In this study, the effects of equivalent ratio on the engine performance and exhaust emissions of diesel engine using diesel and biodiesel fuel was experimentallyinvestigated. Based on the results of this study, the con-clusions can be drawn as follows: 1. The engine power that using diesel and biodiesel fuel are similar. The engine power increased with the increase in the equivalent ratio for the both fuels.The brake specifi c fuel consumption that using diesel and biodiesel fuel are lowest in the equivalent ratio close to unity. The brake specifi c fuel consumption that using diesel and biodiesel fuel are similar. 2. The fuel conversion effi ciency that using diesel and biodiesel fuel will be highest in the equivalent ratio close to unity. The fuel conversion effi ciency that using diesel and biodiesel fuel are slightly different. 3. The hydrocarbon emissions increased with the increase in the equivalent ratio. The C

xH

y emissions

of diesel fuel are similar to biodiesel in range of the equivalent ratio more than unity and at high load. 4. The lowest carbon monoxide occurs at equivalent ratio close to unity. The CO emissions of diesel fuel are higher than biodiesel in range of the equivalent ratio more than unity and at high load. 5. The highest nitrogen oxide occurs at equiva-lent ratio close to unity. The NOx emissions of diesel and biodiesel fuel are slightly different.

Acknowledgements The authors would like to thank the Department of Mechanical Engineering, Faculty of Engineering and

Industrial Technology, Silpakorn University for the fi nancial support.

References1 Lapuerta M., Armas O., Fernández J. R., Effect of bi-

odiesel fuels on diesel engine emissions, Progress in Energy and Combustion Science 34 (2008) 198–223.

2 Altun S., Bulut H., Öner C., The comparison of engine performance and exhaust emission characteristics of sesame oil–diesel fuel mixture with diesel fuel in a direct injection diesel engine, Renewable Energy 33 (2008) 1791–1795.

3 Misra R.D., Murthy M.S., Comparative Performance Evaluation of Karanja Vegetable oil and Karanja Biodiesel Blends with Diesel in C.I. Engine, IEEE First Conference on Clean Energy and Technology (2011).

4 Gandure J., Ketlogetswe C., Comparative Perform-ance Analysis of Marula Oil and Petrodiesel fuels on a Variable Compression Ratio Engine, IEEE Africon (2011), The Falls Resort and Conference Centre, Livingstone, Zambia, 13 - 15 Sep. 2011.

5 Rakopoulos C.D. et al, Comparative performance and emissions study of a direct injection Diesel en-gine using blends of Diesel fuel with vegetable oils or bio-diesels of various origins, Energy Conversion and Management 47 (2006) 3272–3287.

6 Sayin C., Gumus M., Impact of compression ratio and injection parameters on the performance and emissions of a DI diesel engine Fueled with biodiesel-blended diesel fuel, Applied Thermal Engineering 36 (2011) 3182 – 3188.

7 Raheman H., Ghadge S.V., Performance of diesel engine with biodiesel at varying compression ratio and ignition timing, Fuel 87 (2008) 2659-2666.

8 Sangsawang T., Bijayendrayodhin P., Tiansuwan J., Combustion Phenonemna and Emissions of IDI engine Fuelled with Blended with CPO/Diesel, Applied

Engineering in Agriculture 27 (2011), 13 – 23.

Original

Daylight Performance of an Automated Vertical Blinds System

Vichuda Mettanant1*


AbstractThis paper describes the mathematical calculation of interior illuminance when vertical blinds are used. A computer program was developed and verifi ed with the measurement data from an experiment. A prototype setup of an automated vertical blinds system was used in the experiment. The auto blinds were programmed to be able to adjust their angles according to the position of the sun in order to prevent direct illuminance which can cause visual discomfort. Only useful daylight from the sky can enter a room. The results of interior illuminance from the calculation were in agreement with the measurement data from the experiments. The computer program was used to calculate the interior illuminance level for rooms of which window are in the north, east, south, and west orientations. The results show that, when an automated vertical blinds system was used in rooms with window in the east and west directions, interior illuminance level was higher than when that was used in a room with the south or north windows. Useful daylight illuminance can supplement illuminance from electric light. Calculation results show that using automated vertical blinds could save electricity of lighting system up to 18%.

Keywords: Daylight, Vertical blinds, Automated blinds

IntroductionDaylight is a useful source to illuminate an interior space. However, it needs to be careful when using natural light because direct sunlight can cause glare or visual discomfort. Diffuse sunlight is the part that is desirable. Use of an automated blinds was suggested as one of the solutions.Chaiwiwatworakul et al.1 suggested the algorithm to calculate interior illuminance when venetian blinds are

used. The vertical blinds are highly used in offi ce buildings, and the study of this type of blinds is still limited. This

paper presents the study of interior illuminance level when automatic vertical blinds are used by modifying the algorithm of Chaiwiwatworakul et al.1

Calculation of Daylight Illuminance through Vertical Blinds System The vertical blinds system includes a single glass pane and vertical blinds installed inside a room as shown in Figure 1.

Direct luminance If the blinds angle is fi xed or is not adjusted according to the position of the sun, direct sunlight can enter a room. This can cause visual discomfort. For

example, when the position of the sun is as shown in Figure 1, part of direct sunlight can enter a room. The part of daylight going through the blinds system can be expressed as the difference between the length “w” (in Figure 1) and the blinds separation (S

b).

The length w can be calculated as

where a is the difference between solar azimuth

angle (gs), window azimuth angle (g

w), b is the tilt angle

of blinds slats, and Wb is the width of a blinds slat.

Figure 1 Top view of two adjacent slats and the line of incident

sunlight.


Diffuse Illuminance When an automatic blinds system is used, blinds slats are turned to prevent direct sunlight, and only dif-fuse light is able to pass into a room. Diffuse light from the window system is composed of light from the sky, refl ected light from the ground, and refl ected light from blinds slats. Sky luminance To calculate diffuse light from sky, the lumi-nance distribution of the sky is treated as non-uniform sky. This means the luminance level at different patches of the sky is varied. The sky luminance is calculated based on Atmospheric Sciences Research Center-the Commission Internationale de l’ Eclairage (ASRC-CIE) sky model.2

Ground luminance The luminance of ground L

g can be determined

from global illuminance and refl ectance of ground using Equation 2.

where ρg is the ground refl ectance, and Evg is the global-horizontal illuminance on the ground. Blinds luminance Blinds receive luminance on their surface from direct sunlight, diffuse sky light, and refl ected light from the ground. This is initial light on blinds surfaces, then initial light refl ects from one side to the other side of blinds slat. After inter-refl ection of light between slats, the fi nal luminance is a light source from window into interior

space. Figure 2 shows segments of a blinds system. Segments 1 and 2 (S

1 and S

2) are imaginary lines. Their

lengths are equal to blinds separation. Segment 4 is the

part that receives direct sun light. Its length is calculated as

Segment 3 is an opposite side of Segment 4. The length of Segment 6 is equal to the subtraction of S

4

from the blinds width. The lengths of Segments 3 and 5 are equal to those of Segments 4 and 6, respectively.

Figure 2 Top view of a blinds system showing segments of blinds

surfaces.

Only Segment 4 receives direct sunlight. The beam illuminance (E

vb) that transmits through a glass

pane strikes on a blinds slat and refl ects off that slat. The initial exitance of Segment 4 due to direct light from the sun can be expressed by Equation 4.

4, 4( ) coso b W W vb bM Eτ η ρ η= (4)

where tw(h

w) is the transmittance of the window

as a function of the incident angle on window, rw is the

refl ectance of Segment 4, and hb is an incident angle on

the blinds. The direct exitance due to direct light of other segments is zero.

The initial exitance of the segment due to dif-fuse light from the sky and the ground can be obtained

by Equation 5

( ), cosrgt up

lft low

oi d w w bM L d dγ θ

γ θ

τ η η θ φ= ∫ ∫ (5)

where L is sky or ground luminance level, q

low and q

up are the lower and upper limits of elevation

angle of the segment, and grgt and g

lft are the right and

left limits of azimuth angle of the segment. The fi nal exitance of each segment can be calculated

by solving simultaneously a matrix of linear equation below

i oi i j ijj

M M M Fρ= + ⋅∑ (6)


where Mi is the fi nal exitance of segment S

i, M

oi is the

initial exitance of segment Si, r

i is the refl ectance of the

segment Si, F

ij is the view factor from Segment i to Seg-

ment j. The luminance on the surfaces of the right and the left blinds slat can be expressed by Equations 7and 8.

3 3 5 5

3 5

( )π( )rt

S M S MLS S+

=+

(7)

4 4 6 6

4 6

( )( )lt

S M S MLS Sπ+

=+

(8)

Window luminance As shown in Figure 3, at a particular blinds tilt angle b, a series of blinds slats can be divided into two main regions: right region and left region, and four sub-regions A to D.

View point

β−

D

C

B

A

β

1 1 sintancosB

Bββ

− ⎛ ⎞−⎜ ⎟⎝ ⎠

1 1 sintancosB

Bβ

β− ⎛ ⎞+

− ⎜ ⎟⎝ ⎠

Left Region

Right Region

Figure 3 Top view of a series of blinds slats.

The right region represents the area where the surface of right slats is invisible while left region repre-

sents the area where the surface of left slats is invisible. The proportion of the exterior environment that is visible through the blinds slats from a point in the interior for

the right region and the left region can be expressed by Equations 9 and 10, respectively.

bR

b

S x yVS− −

= (9)

bL

b

S x yVS+ −

= (10)

where Sb is the blinds separation, x and y are

distances as shown in Figures 4 and 5. The black dot in Figures 4 and 5 indicates a point of view inside a room, and the dash line represents the line of eye sight.

Figure 4 Top view of two slats showing visible ratio of the right region.

Figure 5 Top view of two slats showing visible ratio of the left region.

In sub-region A, only the surface of the left blinds can be seen. Therefore, the luminance of the window is equal to the luminance of the left blinds as calculated from Equation 8. In sub-region B, both exterior environment and

surface of the left blinds can be seen. The luminance of the window in sub-region B can be obtained from

For sub-region C, exterior environment and the surface of the right blinds can be seen. Equation 12 shows the luminance of window in sub-region C.


In sub-region D, only the surface of the right blinds can be seen. The luminance of the window is equal to the luminance of the right blinds. It can be calculated from Equation 7.

Experimental Setup The experimental setup is located at latitude 13.49oN and longitude 100.2oE. It is a class room with the dimension of 3.85 m width, 6.36 m depth, and 3.50 m height. The window is oriented in the north-west direction with azimuth angle of 125 degree. The window is 1.88 m length and 1.45 m height. The window is located 0.94 m above the fl oor. There are twelve lamp fi xtures in the experimental room as shown in Figure 6.

Figure 6 The experimental room

The window system consists of a clear-glass pane with 6 mm thickness and the automatic vertical blinds as shown in Figure 7. Transmittance of window

pane is 0.87. The blinds slats are made of aluminum painted in white. Refl ectance of the slats is 0.75.

Figure 7 Installation of automated blinds in the experimental room

The controlling part of the automatic blinds system is composed of a stepping motor and a microcontroller module to drive the motor as shown in Figures 8 and 9. The stepping motor is connected to the shaft of the blinds system (see Figure 8).

Figure 8 Stepping motor for blinds slat motion

The microcontroller module to control the motor and the blinds includes an interface board model AVR Mega 328, a real time clock circuit to input data of time for the microcontroller board, and a limit switch to check the blinds angle.

Figure 9 The microcontroller module for driving the motor

The blinds slats were programmed to tilt automatically according to position of the sun in order

to fully shade direct sunlight. The measurement was conducted by using a lux meter model Testo 545.

Results and Discussion 5.1 Comparison between simulation and

experimental results To verify the simulation program developed, the experiments were conducted for six days in March 2012. Figure 11 shows results on 22nd of March 2012. Global

and diffuse horizontal illuminance data is acquired from department of physics, Faculty of Science, Silpakorn Uni-versity, Nakhon Pathom, Thailand. The station is nearby


the experimental site. Figure 12 shows global and diffuse horizontal illuminance data on March 22nd, 2012 during 9:00-17:00 hr.

Figure 10 Global and diffuse horizontal illuminance on 22nd March

2012

Inside the experimental room, Figure 11 shows the illuminance level on the fl oor at 17%, 50%, and 84% of the room depth when it is measured from the window. The solid lines present results measured from the experi-ment setup while the dash lines present results calculated from the simulation program.

Figure 11 Interior illuminance on 22nd March 2012

Because the window is facing the north-west direction, the automatic vertical blinds adjusted their an-

gles to prevent beam illuminance in the afternoon. As a result, the illuminance levels were low in the afternoon. The illuminance level decreases when the position away from the window increases.

The results from simulation are in the same trend with those from the experiment. The Root Mean Square Error (RMSE) and Mean Bias Difference (MBD) between the measuring and calculating results are showed in Table 1.

Table 1 The Root Mean Square Error (RMSE) and Mean Bias Difference (MBD) between the measuring and calculating results

Position

(% of room depth)

RMSE (%) MBD

(lux)

17 30.76 -9.92

50 20.96 12.74

84 4.54 -1.20

Effects of window orientations To express the effects of using the automatic blinds for a window in other orientations on illuminance levels, illuminance was calculated by using the simulation program. The illuminance levels were calculated for the experimental room with window orientation of north, east, south and west on four reference days including 21st March, 21st June, 21st September, and 21st December. The average illuminance levels on the fl oor at 33% of room depth are 308.68, 306.32, 289.04, and 238.69 lux when the window is facing west, east, south and north, respectively. Electrical saving According to the Commission Internationale De L’Eclairage (CIE) standard, the illuminance level of a

room for a general purpose should not be less than 300 lux. The calculation was evaluated for year 2011. The fl oor was separated into 35 sections. When the interior illuminance level was lower than 300 lux, the lamps above that area were turned on. Each fi xture had two

sets of 36-W-fl uorescent lamps and a ballast of 14 W. The electricity charge was assumed as 2.78 Baht/kWh, the results showed that the electricity saving was 18%

when and the payback period was 6 years.


Conclusion The simulation program was developed to calculate the interior illuminance level when a vertical blinds system was used. The program was verifi ed with the experiments and the results were substantiated. The prototype of automatic vertical blinds was set. The blinds can adjust their angle automatically to fully prevent direct sunlight which cause glare or visual discomfort and allow only useful diffuse light into a room. The calculation also shows that use of the automatic blinds system can reduce electricity consumption of the lighting system.

References1. Chaiwiwatworakul P, Chirarattananon S, Rakkwamsuk P.

Application of automated blinds for daylighting in tropical region. Energy Convers Manage 2009;50: 2927-2943.

2. Chaiwiwatworakul P, Chirarattananon S. Evaluation of Sky Luminance and Radiance Models Using Data of North Bangkok. J. Illum. Eng. Soc. 2004;1,107-126.

คําแนะนําสําหรับผูนิพนธ

วารสารวิทยาศาสตรและเทคโนโลยี มหาวิทยาลัยมหาสารคาม กําหนดพิมพปละ 6 ฉบับ ผูนิพนธทุกทานสามารถสงเรื่องมาพิมพได โดยไมตองเปนสมาชิก และไมจําเปนตองสังกัดมหาวิทยาลัยมหาสารคาม ผลงานที่ไดรับการพิจารณาในวารสารจะตองมีสาระที่นาสนใจ เปนงานที่ทบทวนความรูเดิม หรือองคความรูใหมที่ทันสมัย รวมทั้งขอคิดเห็นทางวิชาการที่เปนประโยชนตอผูอาน และจะตองเปนงานที่ไมเคยถูกนําไปตีพิมพเผยแพรในวารสารอ่ืนมากอนและไมอยูในระหวางพจิารณาลงพมิพในวารสารใด บทความอาจถกูดัดแปลง แกไข เนือ้หา รปูแบบ และสาํนวน ตามทีก่องบรรณาธกิารเหน็สมควร ทั้งนี้เพื่อใหวารสารมีคุณภาพในระดับมาตรฐานสากลและนําไปอางอิงได

การเตรียมตนฉบับ1. ตนฉบับพิมพเปนภาษาไทยหรือภาษาอังกฤษก็ได แตละเร่ืองจะตองมีบทคัดยอทัง้ภาษาไทยและภาษาอังกฤษ

การใชภาษาไทยใหยึดหลักการใชคําศัพทการเขียนทับศัพทภาษาอังกฤษตามหลักของราชบัณฑิตยสถานใหหลีกเลี่ยงการเขียนภาษาอังกฤษปนภาษาไทยในขอความ ยกเวนกรณีจําเปน เชน ศัพททางวิชาการที่ไมมีทางแปล หรอืคําที่ใชแลวทําใหเขาใจงายขึน้ คาํศพัทภาษาองักฤษทีเ่ขยีนเปนภาษาไทยใหใชตวัเลก็ทัง้หมด ยกเวนชือ่เฉพาะ สาํหรบัตนฉบบัภาษาองักฤษ ควรไดรับความตรวจสอบที่ถูกตองดานการใชภาษาจากผูเชี่ยวชาญดานภาษาอังกฤษกอน 2. ขนาดของตนฉบับ ใชกระดาษขนาด A4 (8.5x11 นิ้ว) และพิมพโดยเวนระยะหางจากขอบกระดาษดานละ1 นิ้ว จัดเปน 2 คอลัมภ ระยะหางระหวางบรรทัดในภาษาที่ใช double space ภาษาอังกฤษลวนใหเปน single space 3. ชนิดของขนาดตัวอักษร ทั้งภาษาไทยและภาษาอังกฤษใหใชตวอักษร Browallia New ชื่อเร่ืองใหใชอักษรขนาด 18 pt. ตัวหนา ชื่อผูนิพนธใชอักษรขนาด 16 pt. ตัวปกติ หัวขอหลักใชอักษรขนาด 16 pt. ตัวหนา หัวขอรองใชตัวอักษรขนาด 14 pt. ตัวหนา บทคัดยอและเนื้อเรื่องใชตัวอักษรขนาด 14 pt. ตัวหนา เชิงอรรถหนาแรกที่เปนชื่อตําแหนงทางวิชาการ และที่อยูของผูนิพนธ ใชอักษรขนาด 12 pt. ตัวหนา 4. การพมิพตนฉบบั ผูเสนองานจะตองพมิพสงตนฉบบัในรปูแบบของแฟมขอมลูตอไปนี ้อยางใดอยางหนึง่ ไดแก ".doc" (MS Word) หรือ ".rft" (Rich Text) 5. จํานวนหนา ความยาวของบทความไมควรเกิน 15 หนา รวมตาราง รูป ภาพ และเอกสารอางอิง

6. จํานวนเอกสารอางอิงไมเกิน 20 หนา 7. รูปแบบการเขียนตนฉบับ แบงเปน 2 ประเภท ไดแก ประเภทบทความรายงานผลวิจัยหรือบทความวิจัย

(research article) และบทความจากการทบทวนเอกสารวิจัยที่ผูอื่นทําเอาไว หรือบทความทางวิชาการ หรือบทความทั่วไป หรือบทความปริทัศน (review article)

บทความรายงานผลวิจัย ใหเรียงลําดับหัวขอดังนี้

ชื่อเรื่อง (Title) ควรส้ัน กะทัดรัด และสื่อเปาหมายหลังของงานวิจัย ไมใชคํายอ ความยาวไมเกิน 100 ตัวอักษรชื่อเรื่องใหมีทั้งภาษาไทยและภาษาอังกฤษ ชื่อผูนิพนธ [Author(s)] และที่อยู ใหมีทั้งภาษาไทยและภาษาอังกฤษ และระบุตําแหนงทางวิชาการ หนวยงานหรือสถาบันที่สังกัด และ E-mail address ของผูนิพนธไวเปนเชิงอรรถของหนาแรก เพื่อกองบรรณาธิการสามารถติดตอได

บทคัดยอ (Abstract) เปนการยอเนื้อความงานวิจัยทั้งเรื่องใหสั้น และมีเนื้อหาครบถวนตามเรื่องเดิม ความยาวไมเกิน 250 คํา หรือไมเกิน 10 บรรทัด และไมควรใชคํายอ คําสําคัญ (Keyword) ใหระบุไวทายบทคัดยอของแตละภาษาประมาณ 4-5 คําสั้น ๆ บทนํา (Introduction) เปนสวนเร่ิมตนของเน้ือหา ทีบ่อกความเปนมา เหตผุล และวัตถปุระสงค ทีน่าํไปสูงานวจิยันี ้วรใหขอมูลทางวิชาการที่เกี่ยวของจากการตรวจสอบเอกสารประกอบ

วัสดุอุปกรณและวิธีการศึกษา (Materials and Methods) ใหระบุรายละเอียด วัน เดือน ปที่ทําทดลอง วัสดุ อปุกรณ สิง่ทีน่าํมาศึกษา จาํนวน ลกัษณะเฉพาะของตัวอยางทีศ่กึษา อธบิายวิธกีารศึกษา แผนการทดลองทางสถิต ิวธิกีารเก็บขอมูลการวิเคราะหและการแปรผล ผลการศึกษา (Results) รายงานผลท่ีคนพบ ตามลําดับข้ันตอนของการวิจัย อยางชัดเจนไดใจความ ถาผลใหมซับซอนและมีตัวเลขไมมากควรใชคําบรรยาย แตถามีตัวเลข หรือ ตัวแปลมาก ควรใชตารางหรือแผนภูมิแทน วจิารณและสรปุผล (Discussion and Conclusion) แสดงใหเหน็วาผลการศกึษาตรงกบัวตัถุประสงคและเปรยีบเทียบกับสมมติฐานของการวิจัยที่ตั้งไว หรือแตกตางไปจากผลงานท่ีมีผูรายงานไวกอนหรือไม อยางไร เหตุผลใดจึงเปนเชนนั้น และมีพื้นฐานอางอิงที่เชื่อถือได และใหจบดวยขอเสนอแนะที่นําผลงานวิจัยไปใชประโยชน หรือทิ้งประเด็นคําถามการวิจัย ซึ่งเปนแนวการสําหรับการวิจัยตอไป ตาราง รปู ภาพ แผนภมู ิ(Table, Figures, and Diagrams) ควรคัดเลือกเฉพาะทีจ่าํเปน แทรกไวในเนือ้เรือ่งโดยเรียงลําดับใหสอดคลองกับคําอธิบายในเน้ือเร่ือง และมีคําอธิบายสั้น ๆ เปนภาษาอังกฤษ ที่สื่อความหมายไดสาระครบถวนกรณีที่เปนตาราง คําอธิบายอยูดานบน ถาเปนรูป ภาพ แผนภูมิ คําอธิบายอยูดานลาง กิตติกรรมประกาศ (Adcknowledgements) ระบุสั้น ๆ วางานวิจัยไดรับงานสนับสนุน และความชวยเหลือจากองคกรใดหรือผูใดบาง เอกสารอางอิง (References) ระบุรายการเอกสารท่ีนํามาใชอางอิงใหครบถวนไวทายเร่ือง โดยใช Vancouver Style ดังตัวอยางขางลาง และสามารถดูรายละเอียดและตัวอยางเพิ่มเติมไดที่ www.journal.msu.ac.th

1. การอางอิงหนังสือรูปแบบ: ชื่อผูแตง. ชื่อเร่ือง. พิมพครั้งที่. สถานท่ีเมืองพิมพ: สํานักพิมพ; ปที่พิมพ. p 22-5. (ชื่อชุด; vol 288).ตัวอยาง: Getqen,TE. Health economics: Fundamentals of funds. New York: John Wiley & Son; 1997. P. 12-5 (Annals of New York academy of science; voll 288).ชมพูนุช อองจิต. คลื่นไฟฟาหัวใจทางคลินิก. พิมพครั้งที่ 5 กรุงเทพ: จุฬาลงกรณมหาวิทยาลัย; 2539

2. การอางอิงจากวารสารรูปแบบ: ชื่อผูแตง. ชื่อเร่ืองหรือชื่อบทความ. ชื่อยอวารสาร. ปที่พิมพ เดือนยอ 3 ตัวอักษร วันที่;ปที่ (ฉบับที่): เลขหนา.ตัวอยาง: ก. วารสารไมเรียงหนาตอเนื่องกันตลอดปRussell FK, Coppell AL, Davenport AP. Ln vitro enzymatic processing of radiolabelled big ET-1 in human Kidney

as a food ingredient, Biochem Pharmacol 1998 Mar 1;55(5):697-701พิจารณ เจริญศรี. การปรับความพรอมเทคโนโลยีสารสนเทศและการส่ือสารกอนรุนเขาสูโลกกาวิวัฒนครั้งใหม. นักบริหาร 2547;24(2): 31-6

ข. วารสารเรียงหนาตอเนื่องกันตลอดปRussell FD, Coppell AL Davenport AP. Ln vitro enzymatic processing of radiolabelled big ET-1 in human Kidney as a food ingredient, Biochem Pharmacol 1998;55:697-701

พิจารณ เจริญศรี. การปรับความพรอมเทคโนโลยีสารสนเทศและการส่ือสารกอนรุนเขาสูโลกกาวิวัฒนครั้งใหม. นักบริหาร 2547;24(2): 31-6

3. รายงานจากการประชุมวิชาการรูปแบบ : ชื่อผูแตง. ชื่อเร่ือง. ใน: ชื่อคณะบรรณาธิการ, editors. ชื่อเอกสารรายงานการสัมนา เดือน (ยอ 3 ตัว) วันท่ี; เมือง

ที่สัมมนา, ประเทศ. เมืองที่พิมพ: สํานักพิมพ; ปทีพิมพ. P.1561-5ตัวอยาง: Bengtsson S, solheim BG. Enforcement of data protection, privacy and security and security in medical infromatics. Ln: Lun KC, Degoulet P, Piemme TE, Reinhoff O, editors. MEDINFO 92. Procedings of the 7th World Congress on Medical Informatics; 1992 Sep 6-10; Geneva, Switqerland, Amsterdam: North Holland; 1992. P.1561-5.

พิทักษ พุทธวรชัย, กิตติ บุญเลิศนิรันด ทะนงศักดิ์ มณีวรรณ, พองาม เดชคํารณ, นภา ขันสุภา. การใชเอทธีฟอนกระตุน การสุกของพริก. ใน: เอกสารการประชุมสัมมนาทางวิชาการ สถาบันเทคโนโลยีราชมงคล ครั้งที่ 15. สถาบันวิจัยและพัฒนาสถาบันเทคโนโลยีราชมงคล. กรุงเทพฯ; 2541. หนา 142-9

4. การอางอิงจากพจนานุกรมรูปแบบ: ชื่อพจนานุกรม. พิมพครั้งที่. เมืองหรือสถานที่พิมพ; ปที่พิมพ. หนา.ตัวอยาง: Stedmin's medical dictionary. 26th ed. Baltimore: Williams & Wilkins; 1995. Apraxia; p. 119-20.พจนานุกรม ฉบับราชบัณฑิตยสถาน พ.ศ. 2542. กรุงเทพฯ: นานมีบุคพับลิเคชันส; 2546. หนา 1488

5. การอางอิงจากหนังสือพิมพรูปแบบ: ชื่อผูแตง. ชื่อเร่ือง. ชื่อหนังสือพิมพ ป เดือน วัน; Sect.: sohk 15.ตัวอยาง: Lee G. Hospitalizations tied to ozone pollution: study estimates 50,000 admissions annually. The Wash-ington Post 1996 Jun 21; Sect. A: 3(col.5).พรรณี รุงรัตน สทศ ตั้งทีมพัฒนาขอสอบระดับชาติมั่นใจคุณภาพ. เดลินิวส 12 พฤษภาคม 2548.

6. อางอิงจากหนังสืออิเล็กทรอนิกสรปูแบบ: ชือ่ผูแตง. ชือ่เรือ่ง. ชือ่วารสารอเิลก็ทรอนกิส [หรอื serial online] ปทีพ่มิพเอกสาร ถาจาํเปนระบเุดอืนดวย; Vol no(ฉบบัที)่: [จาํนวนหนาจากการสืบคน]. ไดจาก: URL: http://www.edc/gov/neidoc/EID/eid.htm วนัที ่เดอืน ปทีท่าํการสบืคน (เขียนเต็ม)

ตัวอยาง: More SS. Factors in the emergence of infectious disease, Emerh Infect Dis [serial online] 1995 Jan-Mar; (1): [24 screene]. Available from: RL: http://www.edc/gov/neidoc/EID/eid.htm Accessed 25, 1999.ธีรเกียรติ์ เกิดเจริญ. นาโนเทคโนโลยีความเปนไปไดและทิศทางในอนาคต. วารสารเทคโนโลยีวัสดุ ตุลาคม-ธันวาคม (17): 2542 ไดจาก: http://www.nanotech.sc.mahidol.ac.th/index.html May 13 2005.

Instruction for Authors

Research manuscripts relevant to subject matters outlined in the objectives are accepted from all institu-tions and private parties provided they have not been preprinted elsewhere. The context of the papers may be revised as appropriate to the standard.

Preparation of manuscripts:1. Manuscripts can be written in either Thai of English with the abstract in bort Thai and Ehglish Papers

should be specifi c, clear ,concise, accurate, and consistent. English language manuscripts should be checked by an English language editor prior to submission. 2. Manuscripts should be typed in MS word ".doc" or ".rtf" (Rich Text) on standard size paper, A4 or 8.5x11 inches, and arranged in two columns: single space for English, double space for Thai language. 3. Browallia font type is required with font siqe as follows: Title the article: 18 pt. Bold Name(s) of the authors: 16 pt. Main Heading: 16 pt. Bold Sub-heading: 14 pt. Bold Body of the text: 14 pt Footnotes for authors and their affi liations: 12pt. 4. The number of pages to 15, including references, tables, graphs, or pictures. 5. Types of munuscripts accepted: research articles and review articles. 6. Orgenization of research articles.

Title: denoted in both Thai and English, must ge concise and specifi c to the point, mormally less than 100 characters.Name(s): of the author(s) and their affi liation must ge given in both Thai and English.

Abstract: This section of the paper should follow an informative style, concisely covering all the imortant fi nding in the text. Authors should attempt to restrict the abstract to mare than 250 words.

Keywords: Give at least 4-5 concise words.

The body of the text comprises the following headings:

Introduction: A summary of who is doing what, why where, and when?

Materials and Methods: A discussion of the materials used, and a description clearly detailing how the experiment was undertaken, e.g., experimental desigh, data collection and analysis, and interpretation. Results: Present the output. Li the information in complicated, add tables, graphs, disgrams etc., as necessary.

Discussion and Conclusion: Discuss how the results are relevant to the objectives or former fi ndings, why? Finally state what recommendations coulld be drawn.

Tables, fi gures,diagrams, pictures: should be screened for those important to support the fi ndings, and separated from the text. Captions should be placed above the tables but under the fi gures.

Acknowledgement: the name of the persons, organization, or funding agencies who help support the research are acknowledged in this section.

References: listed and referred to in ancouver style.(http://www.library.uq edu.au/training/citation/vancouv.thml)

7. Authors of review articles should follow the typical format style the includes and introduction, the body of content, conclusion, and references.

Submission of manuscripts Manuscripts can be submitted to the Editorial Board, Department of Research Support and Development, Mahasarakhom University, Khamriang Subdistrict, Kantarawichai District, Maha Sarakham Province, 44150. Tel: 0-43754416 or 0-43754416 ext. 1339. Fax: 0-43754416. The author should submit the original paper and one copy together with a written deskette.

Review of manuscripts:

1. The editorial board will review all manuscripts for format compliance. Manuscripts formatted in correctly will be returned to the author for correction. 2. Following submission of the corrected manuscript, the Peer Review Committee will review and offer comments 3. Manuscripts receiving the approval of the Peer Review Committe may be returned to the author for revision as advised by the Committee. Manuscripts failing to adopt the Committee's sugges

ใบสมัครเปนสมาชิกวารสารวิทยาศาสตรและเทคโนโลยี

มหาวิทยาลัยมหาสารคาม

วันที่............เดือน.......................................พ.ศ..............

ชือ่-นามสกลุ ......................................................................................................................................................................

ทีอ่ยู บานเลขที.่...............หมูที.่............ถนน.....................................................แขวง/ตาํบล...............................................

อําเภอ.....................................................จังหวัด.....................................................รหัสไปรษณีย.....................................

โทรศัพท.............................................โทรสาร...........................................E-mail.............................................................

หนวยงาน.........................................................................................................................................................................

........................................................................................................................................................................................

........................................................................................................................................................................................

ถนน..................................................แขวง/ตําบล...................................................อําเภอ................................................

จังัหวดั.....................................................รหสัไปรษณยี.................................................โทรศพัท......................................

❏ สมัครเปนสมาชิกหนึ่งป คาสมัคร 400 บาท

❏ สมัครเปนสมาชิกสองปติดตอกัน คาสมัคร 800 บาท

ทานสามารถส่ังจายธนาณัติหรือตั๋วแลกเงิน สั่งจาย ปณ.โนนศรีสวัสด์ิ ในนาม:

นายสุนทร เดชชัย งานวารสาร กองสงเสริมการวิจัยและบริการวิชาการ มหาวิทยาลัยมหาสารคาม

ตําบลขามเรียง อําเภอกันทรวิชัย จังหวัดมหาสารคาม 44150

✄

Membership Application Form

Jorunal of Science and Technology Mahasarakham University

Date.................................................................

Name (Last)......................................................................(First)......................................................................................

Mailing Address...............................................................................................................................................................

........................................................................................................................................................................................

Sub-district...................................................................... District..............................................................................

Province (City/state)........................................................ Country.............................................................................

Postal Code..................................................................... E-mail Address.................................................................

Telephone No.................................................................. Fax No.............................................................................

❏ One-Year membership (400 Baht)

❏ Two Year Membership (800 Baht)

Please send yoru personal check or money order to the following address:

Mr. Soonthorn Dejchai, Division of Research Supprt and Development, Boromarachakumaree

Building, Khamriang Sub-district, Kantharawichai District, Maha Sarakham Province 44150

พิมพที่ : โรงพิมพคลังนานาวิทยา โทร.043-328589-91 โทรสาร 043-328592 Email : [email protected]

✄

Editorial - Mahasarakham Universityjournal.msu.ac.th/upload/journal_file/jfile_no1_45563.pdf ·...

Documents

Transcript of Editorial - Mahasarakham Universityjournal.msu.ac.th/upload/journal_file/jfile_no1_45563.pdf ·...