水果分選機分選部分設(shè)計,水果,分選,部分,設(shè)計 Research PaperDesign of machine to size java apple fruitwith minimal damageKrawee Treeamnuka, Siwalak Pathaveerata,*, Anupun Terdwongworakula,Chanida BupatabaDepartment of Agricultural Engineering, Faculty of Engineering at Kamphaengsaen, and The Centre of Excellence for Agricultural and FoodMachinery Kasetsart University, Nakornpathom 73140, ThailandbDepartment of Postharvest and Processing Engineering, Faculty of Science and Technology, Rajamangala University of Technology Tawan-ok, PHTIC, Chonburi, Thailanda r t i c l e i n f oArticle history:Received 29 March 2010Received in revised form28 July 2010Accepted 3 August 2010Published online 15 September 2010Java apple fruit features attractive skin, sweet and crispy flesh, and a high susceptibility tomechanical damage. The java apple produced in Thailand has gained popularity in localand foreign markets, is available almost all year round and is sold at good prices. Fruitexport drives postharvest mechanisation, of which fundamentally important machinessuch as a mechanical sizer for java apple are still not available. This research was aimed atdesigning, constructing, testing, and evaluating an efficient sizing machine for java apple.Design concepts featured a) a sizing parameters which were determined by the diameter ofthe fruit and b) a sizing mechanism which causes minimum damage. The sizing machinecomprised a feeding unit and a diverging belt sizing unit that are powered by two 187 W220 V 50 Hz electric motors, gear reducer and pulleys. Performance tests indicated thatvelocity and inclination angle of the sizing belt; feeding belt velocity and the fruit orien-tation significantly affects the sizing performance at p 0.05. The optimum conditions forcontinuous mechanical sizing depended on the variety. The optimum sizing performancewas characterised by a contamination or error ratio of 10.8e16.5%, and a throughputcapacity of 149.7e195.1 kg h?1with no significantly noticeable damage to the sized fruits.Manual sizing of the exported java apple featured an error ratio of 27.9%, a damagepercentage of 13.3%, and a capacity of 107.2 kg h?1. Therefore, the java apple sizingmachine can be operated without adding more mechanical damages to the sized fruit. 2010 IAgrE. Published by Elsevier Ltd. All rights reserved.1.IntroductionJava apple fruit is popular in Thailand and abroad due to itsbeautiful skin colour as well as sweet and crispy flesh. Thefruit is also of high nutritional value. In the year 2005, exportsof the flesh java apple from Thailand reached as much as 44.6million US $, which made it the third ranking export fruit afterdurian and longan.Amongst postharvest operations of the fresh fruit, sizing isconsiderably labour intensive. Sizing is necessary, becausethe sized fruit a) has higher value than fruit that is soldunsorted b) it attracts buyers and c) it facilitates packaging* Corresponding author. Fax: +66 34351896.E-mail address: fengslpku.ac.th (S. Pathaveerat).Available at journal homepage: engineering 107 (2010) 140e1481537-5110/$ e see front matter 2010 IAgrE. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.biosystemseng.2010.08.001designs(Jarimopas,Toomsaengtong,Singh,Singh,&Sothornvit, 2007; Peleg, 1985). Manual sizing is beset by slowsizing rates, considerable errors and mechanical damage. Forexample, the manual sizing of fresh mangosteen (Jarimopas,Toomsaengtong, & Inprasit, 2007) and durian (Jarimopas,Siriratchatapong, Sukharom, Sihavong, & Goto, 1992) wasfound to cause error ratios of 34 and 43% respectively. Sizingmachines developed to work with tropical fruit are usuallycomposed of a perforated conveyor sizer, a belt, a board sizer,and a diverging belt. Jarimopas, Kongwatananon, Rangdang,and Yamashita (1988) tested a commercial tangerine sizingmachine with a mechanism based on the perforated conveyorsizer. Their machine consisted of five perforated cylinders ofvaryingdiametersandsized6gradesofthefruitatacapacityof1.5 tonne h?1and a sizing efficiency of 87.5%. The high sizingefficiency was perhaps due to the non-deformable aperture.Although commercially applicable, single layer feeding has tobe manually controlled and this constitutes an undesirablefeature. The sizing of a small number of tangerines is persis-tently impeded by fruits caught on the inclined plane betweenthe cylinders, so that the operator has to manually move thefruit to the next rotating cylinder to continue sizing. This isa cause of exasperation to the operators and it slows down thesizing process. The rotary sizing machine (GREEFA, Gelder-malsen, Netherlands), which is the developed model of thebelt and board sizer, was used to size mangosteen (Jarimopas,Toomsaengtong, & Inprasit, 2007; Jarimopas, Toomsaengtong,Singh, et al., 2007). The mangosteen sizer could sort 3 grades(small, medium, large) at a continuous capacity of 1026 kg h?1and an error rate of 23%. The Rotary Greefa concept has alsobeen applied to size other sphere-shaped fruit and vegetables,e.g. guava (Jarimopas, Rachanukroa, & Chen, 2002). Theadvantageoftheapparatusisitseaseoftransportation,suchason a pick-up truck. However, its performance is limited tosingle layer feeding and it is characterised by errors due to theinconsistent aperture. Jarimopas et al. (1988) developeda diverging belt sizer to sort mangosteen. The diverging belt iscomprised of two 75 mm wide belts which are inclined at anangle of 25?to the horizontal surface. The machine gradedmangosteen at a capacity of 1100 kg h?1at an efficiency rate of80%.Thelowinclinationofthesizingbeltrequiredmangosteento travel considerable distances before calibration of the fruitdiameter and the descent from sizing belt posed a majordisadvantage of the machine. This resulted in a long machine,whichmadetransportationdifficult.Furthermore,itsaperturewas inconsistent because of the non-uniformity of the sizingbelt.For the frustum-shaped fruit, such as pears and carrots,Peleg(1985)suggestedusingthedivergingbeltsizerforefficientsorting. A proper sizing machine ought to feature fast sizingrates, acceptable error ratios with insignificant mechanicaldamage. Mohsenin (1996) stated that mechanical damage wasthe by-product of the use of agricultural machinery. Possibly,theaforementionedmachines causedmechanicaldamagebutthe related evaluations were not reported. Bupata, Jarimopas,& Changtong (2007) reported that the diameter of java applefruit varies proportionally with the fruit weight. The variationof the java apple fruit diameter respective to size and cultivarwas presented by Sarakan, Jarimopas, and Changtong (2007).Based on the compression rupture force of java apple skin, itsFig. 1 e Design concept of the java apple fruit sizingmachine showing the inclination angle a and front view ofthe sizing configuration (horizontal plane is perpendicularto the paper).NomenclatureDmDiameter of the java apple fruit head, mmbDiverging angle of the sizing belt,?aInclination angle of the belt,?X12Diameter of separation of aperture between javaapple grade 1 and 2, mmX23Diameter of separation of aperture between javaapple grade 2 and 3, mmm1average diameter of java apple fruit grade 1, mmm2average diameter of java apple fruit grade 2, mms1standard deviation of java apple fruit grade 1s2standard deviation of java apple fruit grade 2EwSizing efficiencyQThroughput capacity, kg h?1CRMean contamination ratio or errorGiOutflow rate of grade i, kg h?1KiFraction of cost as related to grade iNiNumber of grade i fruit of the total assortment atthe beginning of sizingNijNumber of grade j fruit in the grade i receiverNgiNumber of grade i fruit in grade j receiverNtiTotal fruit in grade i receiver, kgPiFraction of grade i fruit of the total mixture at thebeginning of sizingPgiFraction of correct fruit in grade i receivertFeeding time, hWiWeighting functionwiTotal weight of fruit in grade i receiver, kgwtTotal fruit weight fed into sizing system, kgbiosystems engineering 107 (2010) 140e148141skin was found more susceptible to damage than apple(Jarimopas, Toomsaengtong, & Inprasit, 2007; Jarimopas,Toomsaengtong, Singh, et al., 2007). Abrasion and bruisingwere potentially found on packaged java apples under simu-latedtransportvibrationwhilstgreaterabrasionwasobserved.(Jarimopas, Toomsaengtong, & Inprasit, 2007; Jarimopas,Toomsaengtong, Singh, et al., 2007).Commercially java apple yields high prices, and is gain-ing popularity in foreign markets where it is availablealmost all year round. Fruit production is rather uniformbutshortages and costs of labour are becoming crucial.Therefore, the demands for sizing machines have increased.A specific sizing machine for java apple fruit is not avail-able.Thedevelopmentofasimple,compact,ruggedmechanical sizer to work with the highly susceptible fruit isa challenge. Thus, this research is aimed at designing,constructing, testing, and evaluating a sizing machine forfresh java apple fruit with acceptable error rates andminimal damage.2.Materials and methods2.1.Design and operation2.1.1.DesignThe design concept comprised four criteria: a) the sizingparameter was to be determined by the diameter of the javaapple fruit head (Dmin Fig. 1), b) the sizing mechanism had toincurre minimum damage, c) the sizing machine had toprovide mechanisms to consistently meter the fruit diameterand to uniformly release the measured fruit into receivingcontainer of the corresponding size, and d) the machineshouldbesimple,compact,ruggedandtransportable.A divergingbelt designwas selectedbecause(i) the sizingbeltsprovide two contact points that measure the java applediameter whilst the fruit progresses along the belts. (ii) thesizing belts travel simultaneously diverging during move-ment;and asa result,themeasuredfruitdescends onthebeltsand left the belts at their base (aperture) when the apertureloses contact with the measured fruit; the fruit descends tothe receiving container of the corresponding size. (iii) theshape of java apple fruit fairly resembles that of pear. Fig. 2showed the base view of the diverging belt sizing machine;the solid line shows the rim of the aperture.Preliminary drop tests of 15 fresh intact java apple fruit (ofuniform size) from 150, 200 and 250 mm height onto a hardsurface cushioned by 20 mm foam rubber revealed that fruitdamaged occurs at a minimal plunge height of 200 mm. For anassumed uniformly accelerated motion, this corresponds to0.2 s drop time (Meriam, 1975). The travel velocity of the sizingbelt of 200 mm s?1and the combined time of fruit descent andfruit removal from of the receiving tray of 0.5 s, dictatesa feasible placement of the fruits at 200 ? 0.5 100 mmintervals. The diverging belt sizing machine was designed asto allow 3 fruits of the same grade to travel in sequence alongthe belt (the small grade permits 4 fruits) (Fig. 2). Three grades(large, medium, and small) are required. The active length wasto be 900 mm plus 100 mm of additional space for the fruitrebound (if any) on each side, giving a total length of 1100 mmis designates. The diverging angle of the belt is b in Fig. 1. Asper Bupata et al. (2007), b was equal to 0.29, 0.55 and 0.34?forthe varieties of java apple Toonklao, Tubtimjan and Tong-samsri respectively. The inclination angle of the belt is shownas a in Fig. 1. The diverging belts need to be inclined to thehorizontal plane, because the inclination causes 2 pointcontact for every fruit during movement regardless of size.Thus, every fruit is measured at its head diameter. Should a beFig. 2 e Aperture profile of the sizing belt (BL belt length, WL sizing length of grade L, WM sizing length of grade M,WS sizing length of grade S, kt interval of fruits (200 mm).Table 1 e Statistics of aperture and b angles of the sizing belts.VarietyMaximum Diameter for the fruit size (mm)X12(mm)X23(mm)b (?)SMLToonklao52.92 ? 1.5757.70 ? 0.8964.12 ? 1.3759.96 ? 4.1653.01 ? 7.300.29Tubtimjan45.30 ? 2.9358.08 ? 1.5365.80 ? 1.3662.87 ? 9.4094.86 ? 32.730.55Tongsamsri51.15 ? 2.2058.47 ? 1.5363.07 ? 1.8265.33 ? 10.0247.39 ? 13.290.34biosystems engineering 107 (2010) 140e148142too large the fruit falls and does not symmetrically lie on thebelts with respect to the y-axis and the head diameter istherefore not measured. If the a is too small, the fruit shouldermight stay at the rim, and diameter measurement does notoccur. The appropriate inclination angle (a) therefore neededto be determined. Eq. (1) (Peleg, 1985) suggests the estimate ofdiameter of separation (X12and X23) between java apple fruitsof adjacent grades from the mean and the standard deviationof each adjacent grade. The values of X12and X23are given inTable 1.X12?m2s21? m1s22?s21? s22?m2s21?1s22s21? s22?2?m22s21? m21s22? 2s21s22lns1=s2?s21? s22#121whereX12 diameters of separation of aperture between java applegrade 1 and 2m1 average diameter of java apple fruit grade 1m2 average diameter of java apple fruit grade 2s1 standard deviation of java apple fruit grade 1s2 standard deviation of java apple fruit grade 2Fig. 3 shows the developed java apple fruit sizing machinewith the feeding unit and the diverging belt sizing unit. Thefeeding unit, 598 mm wide by 1430 mm long by 520 mm high,was made of steel. Two 25.4 mm diameter cylindrical shaftswere mounted at the top of the frame to facilitate beltadjustments. The feeding belt, Ammeraal type Belt Flexam EM10/2GreenFG(AmmeraalBeltech,Heerhugowaard,Netherlands), was 100 mm wide by 2400 mm long and parallelFig. 3 e Schematic diagram of java apple fruit sizing machine (A feeding unit; B diverging belt sizing unit; 1 receivingtray; 2 steel frame; 3 conical cloth tube; 4 cushioned sizing belt; 5 universal joint; 6 cylindrical shaft; 7 electricmotor and reduction gear; 8 U-shape rubber holding java apple; 9 feeding belt).Fig. 4 e Typical motion of java apple fruit and the conical cloth tube of the sizing machine.biosystems engineering 107 (2010) 140e148143to the floor. The feeding unit was driven by a 0.19 kW 220 V50 Hz electric motor, a 1: 20 gear reducer, a chain and a pulley.A U-shape rubber cushion was attached to the surface ofconveyor belt with 80 mm spacing of the partial containers ofthe fruit for uniform semi-automatic fruit feeding. The sizingmachine, made of steel and 390 mm wide, 1510 mm long and765 mm high, featured a galvanised steel receiving tray600 mm wide 1100 mm long, and 80 mm deep. The receivingtray was cushioned with 20 mm rubber foam with twoadjustable partitions for the desired fruit size separation. Thesizing belts diverged and were inclined to the horizontal planeatadegree.Each ofthetwosizingbelts,AmmeraaltypeNonexEM 8/2 Blue FG were 100 mm wide and 2500 mm long. Theouter surface of the belts was lined with 4 mm latex foam.Each belt was driven at the same velocity by a 220 V 50 Hzelectric motor operating at 1450 rpm with reducing gear anda universal joint to facilitate the adjustment of belt angle.A cloth conical tube was placed between the end of thefeeding belt and the beginning of the sizing belt to control thevertical descent of the java apple fruits, to decelerate fruitdrop onto the sizing belt thereby reducing damage to the fruit.The movement of the fruit on the sizing belt was belowhorizontal thereby minimising vertical sliding. The conicalcloth tube was 120 mm long and made of yarn. It hada 100 mm aperture diameter at the top and 60 mm aperturediameter at the bottom. The feeding machine and the sizingmachine were placed in proximity to the cloth conical tubeand were well aligned.2.1.2.OperationJava apple fruits were manually continuously fed into the U-shaped rubber of the feeding belt by an operator in such a waythat the stem end of the fruit was parallel to the belt move-ment.This providesa uniformly automatic feed of the fruits tothe sizing belt. However, the feeding machine capacity waslimited by the operators competence. The conveyed fruit leftthe feeding belt and dropped vertically into the conical clothtube (Fig. 4). The conical cloth tube not only carried the fruitvertically, but it also decelerated the fruit to minimise theimpact on the sizing belt. The appropriate alignment andsymmetry of the feeding belt, the cloth conical tube, and thesizing belt improved the vertical orientation of the java appleduring its descent at the start of the sizing process.The sizing belt carried the java apple fruits past 2 contactpoints. The forward motion, divergence of the sizing belts andthe fruit weight caused the fruit to descend towards the lowerend or the aperture of the belts. Therefore, the two contactpoints on the sizing belts and fruit surface were continuously.Such phenomenon could cause skin rubbing on the fruit andmight give rise to abrasion. This was prevented by cushioningon the surfaces of the sizing belts. When the fruit reached theaperture that infinitesimally exceeded the fruit diameter, thefruit lost contact with the belts and fell into the correctreceiving tray.2.2.Performance testSamplepreparationforthedeterminationofoptimumworking conditions included the random selection of newly-harvested, uniform and damage free java apple fruit. Thirtysamplefruitsofeach size(threesizes:small,medium,large)ofeach variety (three varieties: Toonklao, Tubtimjan and Tong-samsri) were collected. Maximum fruit head diameter, lengthand weight of each sample were measured with a verniercalliper and an electronic balance (Sartorius G 6000, SartoriusAG, Goettingen, Germany).2.2.1.Determination of the optimum angle of inclination andthe velocity of sizing beltSamples of java apple fruit were prepared as mentionedabove. Testing conditions included two control factors; theangle of inclination of the sizing belt (3 levels: 75?, 80?and 85?)and the sizing belt velocity (3 levels: 10, 20 and 30 m min?1).Uniform feeding of java apple samples to the sizing belt wasregulated at 15 m min?1. Five replications were carried out fora combination of thecontrol factors.Performance of the sizingmachine was evaluated with Eqs. (2e4). Analysis of variance(ANOVA) and Duncan Multiple Range Test (DMRT) wereapplied to determine the influence of the control factors.EwX?pgiWiGiQPi?(2)Q wtt(3)CRSNijSNi(4)wherePgiNgiNtiNti Ngi NijWiKiPiSKiPiPiNiSNiGiwitFig. 5 e Shape of the head of a java apple fruit.biosystems engineering 107 (2010) 140e148144andwhereEwissizingefficiency,Qisthroughputcapacity,CRismean contamination ratio or error, Giis outflow rate of grade i,Kiisthefractionofcostasrelatedtogradei,Niisthenumberofgrade i fruit of the total assortment at the beginning of sizing,Nijnumber of grade j fruit in the grade i receiver, Ngiis thenumberof gradei fruitin gradejreceiver,Ntiis thetotal fruitingradeireceiver,PiFractionofgradeifruitofthetotalmixtureatthe beginning of sizing, Pgiis the fraction of correct fruit ingrade i receiver, t is feeding time, Wiis a weighting function,wiisthetotalweightoffruitingradeireceiverandwtisthetotalfruit weight fed into sizing system.2.2.2.Determination of the optimum feeding belt velocityThe sample preparation as described above was repeated forthese tests. The determined condition of velocity and incli-nation angle of the sizing belt for a specific variety was notaltered throughout the determination. The control factorswerethefeedingbeltvelocities(3levels:15,20and25 m min?1). Five replications were carried out. The sizingmachine performance was analysed using ANOVA and DMRT.2.2.3.Determination of the optimum fruit orientation and theassociated mechanical damageThe shape of the java apple fruit (Fig. 5) determined themaximum and the minimum dimensions of the fruit head.Both parameters were significantly different (Bupata et al.,2007). The fruit shape influenced the test condition of fruitorientation in regard to the sample placement on the feedingbelt(2possibilities:randomplacementandhorizontalmaximum diameter placement). Each fruit sample was placedwith its stem orientated towards the sizing unit. The deter-mined conditions of velocity and inclination angle of thesizing belt, and feeding belt velocity for a specific variet