f MERD, Vol.5; No. 1, 2008 ‘THE DESIGN OF PALM KERNEL / SHELL SEPARATORS USING HYDROCYCLONE CONCEPTS C.£.0. Kamala ‘Chemical Engineering Department, Federal University of Technolog, Ower NIGERIA, ABSTRACT In this design of palm kernel 1 shall separator using the hydrocyclone Concepes. design layout and. details are gan as well as the costing. The layout ts step wisely cast, the detall design calculations and costing made Selecively The design computational resulls endfor specifeations of kernal land shall hydrocyelones and bats material balances and ole parameters, are tabulatly presented Keywords: Design, Hydrocytone, Diameter, Kernel, Shel, Separator, Shuey, Bath NOMENCLATURE’ Percentage solid by volume of cyclone feed = Speci gravity of solids Specie grail of quid (normaly 10) Vo tow rae ofthe sury Init shry velocty 0s. farshal and swt equipment cost index usc) Year (eg, 1979, 2004 etc 10 WwerRooUCTION The palm frut processing invowves stelzing of the ft bunches, siping, erude ol extraction, of Claciieaten. and. kemel recovery. in the last process. kernel recovery. nts and fers are cried ‘and preumaticaly separated, nuts are died, broken nd shells are saparaiad from kernels. The kernels ara died, Bagged and stored for sale to companies that reed them for their raw materials (Harley, 2001). Nigeria was the largest world producer of palm komels olowed by Brazil and Zaire on the verge. ‘Ste exported 50% of the wold markets palm kamels. This export value rose steady from 527,000 tans in 1948 to 416,000 tons in’ 1982 (OiSy, 2003), contutary factor to his wa the rushing of palm Kemels to extract palm Kernel af tused inthe rumarous saap factones in Niger Any particular ol patm wil always produce fut having sither dura, tenera or pisfera Tha nut of dda forme fat tat fae relatively thick shall (2 {0 40 fmm thick) If gut waneversely no ring of Mibers is need in the masocarp close fo the shell The cu Of fenera forms fut usualy of thinner shell han the Gura (08 fo 4mm). The pisilera forms fuls thal hhave nee Kernel nor shall tis composed aims fleshy of bearing mesocarp malarial (Muthuraen 2012}, A hyeocyclone Is a eyindcal vassol wih ceniea! Bator having 3h apertie of enical ize ‘There san nla! pipe which siuaied jst below the lower of the two horizontal pistes and which enters the cylinder tangentaly. An ext ppe is stucted rear Ihe top of the cjinder The sclion of the hrydracylone IS somewhat similar to thal on an ai Cyclone. that, by imparing 2 creular molon ta the fd by means ‘of the tangential entry, heavy particles are thrown by centnfugal force tothe wall of fhe cyinder, and, afer acing a heical pain ind thelr way out rough the Bottom af the cyclone. The layout af a typical hyerecyclone plant is shown in Ft ‘The tid used in tha hydrocyclone ss water and the mise of shell and kerole are placed in 2 bath Of waler 1 ulvch is connected @ pump so that water ‘charged wih shell and Keres can be pumped mo the hyckocyelones. By choosing the divensions of the hydrecyetone and the pumpng pressure ‘correc. the result Of Mis & that mast ofthe shel pieces (oelng denser than kernels} pass downwards nd out teugh the conical boom with 3 small proportion of the water ow The lrgor part ofthewater together with most ofthe kemels alter taking aft in an initel downward eeu ‘movement Sradualy move towards the center of the cylinder and slarl_ moving. upwards and leave the hydrocyclone via the vero tube and ext pipe Although’ it is manly the lower densily of the kernels that causes them to pass info the upward ‘moving steam near th axl of he cyande, the size ‘aod shape of the paricies have some elect For tus Yeason. small uncracked nuts, though tolermesiale in density between shel and Keinels ‘are often caught by the upward water ow and pass ‘ut wih he kernels (Muthrajah, 2002), The postion ‘of the overtow tube is made ‘aduslale and tne purpose of this isto contol the unty of the upward owing kemel steam: Tt the fovertow Iube & set low dou, same shell vl iso ss upwards nin the keene, but by operating, wth ‘high overfow lube, setting the kernel may ofen be ‘made suffeenty pure 10 bag ater Wey have Separates from the wale by means of a screen ad ied (Bradley, 2001) @ when tis is done the shell traction pessing ‘downwards wil certain coniain too” Bight Proportion of Kemal to permit h to be discarded. without further treatment, Its, Wherefore, passed elo @ secand bat of waler and pumped by another pump 10. second" hydratycone, the "shall Iyrcylone, fom ine frst, the kernel hydrecycione Fr convenience the wo water baths ae adjacent wan @ perforated partion between them. Ths slows the water levels t equole, bul ti unpotant to Prevent the yerforatons Becoming so worn thet emesis can pass through inio the shel! fraction (Mutharajan. 3002). The aim ofthis work's la design a. palm heme! ‘and shell ‘separator that wil ‘He working, on Continucus operation separating a reasonable amount of Kemel trom the shell pee hour for a ‘compary in Rivers Site of Nigra. 20 DESIGN METHODOLOGY 24 Assumptions Made in the Design Panicle size In the appication of the hydrocyclone moses. iis expected that Ihe particles shoud be spherical, tha because the funcional relationships used inthe —cpciona ‘modes are corecied rom the pate diameter fand shape of 3 sphere” Bul wilh parties of hon-sphercal shape, the equiprientchometer fs » a“ Nae Woo. 4, 2008 ‘equal to the diameter of the citele having the Same atea as the projected area ofthe particle and is, oreore. 2 function of eriantation 96 wel. asthe shape of the parte (Trawinsk 1986). |p this design, emel-shell ig taken. as the functional sold particle applied fer ealaaton ‘and desig of the-oyeiona, Ths is because he- application of hydrayclone for the separation oF ‘emel fom shetis possitic beeauise the desi of heme! lo weler 18 nagigie ccimpared the density of shel to wales. sr aso, the proportion of kernel inthe misureof the sy is very smal (Frawinske 1980), 2 3) Feed quits taken tobe water at 20° 4) Cyclone geometry — standard cyclone as Aescrived tha introduction, 5) Feed concentationess than 1% sokd’ by volume, 22 The Problem Statements ‘The prablom is 10 design a hydraeylone plant thal yt produce and separate kernal atthe rte of 43min or S.282MTPH With the information rowded, the food flow rote is equal to Sam. ‘wiich is composed af erecied mixture of kemel 10%, shell 20%, and water 70% by volume of sluty. ‘The separation is expected ta underga twa stage processes with the fat feed infow ino kemel hhyleooycione, where, Oa2Smne of shell partes ull be fost inte the overtow ard 0-4 of kernal Wal be taken slong wth shell and Teva water trough the undertow lo the she ath At the sh bath (shel + kemel), more water troduced with the requted proparicn of water ta rake up the-Tou fate of wate ta 23925. This then makes Ine fnlie ow rate ofthe mixure Séméhe ta be pumped to the stand hyrocyeione (shat hydrocyclone) At the shell hyeeacydione, O2mint of shel is expecios lo be lest thagh the overiow whle the nla kemelinvodueed into the shall hydrocylone is eapectea to be brought cut trough the ovcrtow. ‘Snr of water Is expected to be ios tough he Lunderfow wit the separated shel. Then tho shel ‘nyéracycone overtow products are: recjced lo te keenel ~ shell mistoro bath for eontauaton of he process (see Fig. 2) (Lynch ela, 2002) The problem now is fo size aha select ne appropiate Sperating cantons ofthe hycosyclone23 Hydrocyelone Design Procedure Step 4: The material balancing of the bo hrarocpclones inflow, overflow and underiow are node. end the speciic graviles. of their sry Caletated ‘Step 2: Evaluaton af the eyelone intemal ameter The slandariged relative dmansion of a cyione guen:im terme ol cyclone internal dlameler (Oc) vata from the folowing relation (35 Yok ow rate of he strry Inet stury velocity oe a Ve win ‘Step 3: To aatermne tha pressure crop forthe ‘Operation of hydrozyelone: th the D, valve, and the nel flow rate of the surg, the corresponding pressure opis determined Fam a graph (Fi. 3). Slep 4: Evaluston of dytbase): tis 1s the micron Soe hala ‘standard cyclone. can achive operalng tind the bate condition and fs calclaled fram the Foowing equation Dasbase) zune .. @ Slep 5: Evaluation of dys (application): this is the meron a2 of ihe soit paricie expected of the ‘yelone 10. schiave operating. under the, applied onditon, and. cakulalad rom the folowing equation Fig), as lapplicaton)=d(oa8e) x Cy XC1%0, (8) y= conection fr feed density = (062-VyS9) + a 5 correction for pressure drop = 327» APS o y= carrecton for specie gravy of soles =(1681G,-G)"* © Step 6: Determination of ne proper apex size with the ‘undertow stury ow fale, the coresponding \olue ofthe apex dlamelar is determined trom araph mig 9 42 @ NERD, Vol 5, No. 4, 2008 ‘step 7: The proper sting of the hysreeylene omponenis in relation withthe cyclone diameters Sone, hus (Mula et al, 2000, Aterburn, 1906). Sfirea of teint nozzle = 0.050." 1 Abtex tinder etameter = 0.350, 1 Entre eylindneal section length = 1.40, | fitigatofthe enanee nozzle = 02773850. ‘with ofthe enttance nozzle = 01825740. Length af he ext duct = 0.1250, 1 The included ange ofthe cjcione = 20" 24 Design of the Cracked Mbxture Water Bath In designing the beth, the volume has to be considered suoh thal is nal designed to cover the Filed space fer carving out other duties. Snge he tperation is @ contauaus process, the cracked tmoture wil be supaled steadily throughout the pevod of the operation. Il is not necessary (© Iroduce the slur immesdioy fo Ihe hyerocyetone ‘A pari of fo 20 minutes has fo elapse belore the Staring of the hydrocyclone plant (Gupta, 2001) Since the flow rate ito the bath is SOm‘h or ‘amine of the ‘slurry, ah espacty that wil bold the slury for the wating porod wil be selected (Wusar eta, 2000) 244 Step 1: Waling period Is selected, in the range of 5 15 20 minutes, Stop 2: ‘The volume af the balh that wil hold he Suny fr Ihe waling period i fed = flow rte x Wating tine Step 3 The shape of the bath for sutable exting ‘nd introdeing ofthe sry is selected Step 42 The appraprate dimension ofthe bah are Speciied and ealcusled {rom existng geometical falsbons (Goleman at al, 1980) Step 6: The appropiate entrance and ext area and poatons ate chosen and calculated Design Pracedute forthe Bath 28 Costing of Hydrocyctone Costing is made on the major equipment of ths plant (hydracylona“comporents) havo. the Frowedge of the Mancial involvement. before fmborking on such project. The formalin fr this aeing Is gollen trom the pot of the purchase cost {ollae| versus capacity (rin) made in 1870, and from. Marshal and Sui’ Equiment Cost Index (SCE (Tart, 4899) from the year before 1976 (0 {79 (Tables 1-6) The curtent cost of he cyclone al the same capscty is obtained from obtained by projecting the orginal cost ae 2 base year usina Fig biter, 1959)~ NJERD, Vol. §, No. 1, 2006 mth of story ale by volume of slury hall oy volume ol shen sel by volume of sry | | | 709% walerby vaume of sry 70.37 20%e shel by volume ot shiry 39 34 10% kemel By volume of sityVolumetne Flowtate (USGPM) Fig. 2: Presure Drop Ve Volumetric fow Rate DSO (Microns) Fig. 4: Dig V5 Cyclone Diameler @ NJERD, Vol. 5, Ho. 4, 2006 wenine ar opine |= cyne ean vietsure es Pt eyctond Hamster (inches) ~*~ 44NuBRD, Vol. 5, No. 1, 2006 Flow Rate (GPM) fa 6: Cost of Colecias and Their Relative 4530. RESULTS AND DISCUSSION The resuls of the specications of the hydrooyclone plant are shown in Tables 1-8. The Sbparaton of shell fom kernal in a Pydeocyclone is mode possible bythe iflrences In their densies, proportion and cenistogalaction, of hycocyone ‘The densty of Kernel io 1-08 x 10g? while thal ot shell is 12 x 10"aglmn® The proportion of kerel fo shel is another determining factor fr separation In every samele of cracked mature 6! Keme! and Shalit is siscovered thal the amount of ker fund tne mature fs almost half Ua of shell This Imakes the amount of Kernel not high enough 10 Gstur in the separation of Kernel fom shel wth wale the greatest portion of he mistre. During the operation care must be taken to see tnat the proportion of water tothe cracked midure should be 26 high a8 70% by volume ofthe shit, Just a5 used! in the materi! balance (Tables 1 - 0} ‘iso the tow of the evacked mature into the bath ‘Should be made steady enough ana should balance tha inoduced water (Pit, 1908). As used inthis deuign tho major consivaint 1s the size of Ihe fycrecycone equipment that wil achieve the ‘paraton mn quesiin Ths is why more stots are is made to desigo the fyarocytone that fowresonsva io the Now fate mind for the speratan The man parameter lar every cyclone design 18 the chameler from waich ne ater parameters ofthe Cyclone components aze graded fo conform to the fandard structure. A baller estimation of the appropriate cyclone diamelar is obtained from ontinuty equation for fowing fd From the material balance shown on Tables + and 2, care was taken in olher nol 10 disrupt Ihe broporion af kemel to shell mire. By consiieing {ne reoyled slream from shell hydrocyclone. the Now ate of the shell was made two ines the ow fate ol the kemel, Ths wil ensure stay fow rate Cl ihe shar. with steady specie gravity vo the Kemal ngdreyclona and, also, tvoughout the entre Fagin in the pan! ow The pressure drop gotten from the eaisation was not accurate for the main ‘operation of Ihe tydrocyeione. plant. This ile ‘vation from really can be ikered lo erors made Siring the correlaton of the formas and since the Skury tends fo increase the capaciy af a cyclone ‘ver that showm in water Ihe deviaon 1 prone to 46 NJERD, Vol. 5, No. 4, 2006 8 ‘occur. To that effec, the use of prassue do that Pinthin the range mentioned inthe theory wil Be proper (ie. 40.10 70kpa) to ensure oplimum fparation (Arterburn, 1996), To ansuce thal the feed pressure isthe operating pressure drop of the plant, the ovedow pressure Ffscharge. is mode to. slay at almosphecc level {Grediey, 2001), The piping of the slury ound the plant requires an optimum velociy of the sty 10 fhe cycone which wil ensure tht the paricles 30 fol etic. around the pe, even, in honzonial ection and that wil minumize wear. For tis design &veloiy of msec i selected Inthe constructon of these hydrocyclones, it wll be propet to selec @ good liner mater, which wit optimize wear, The use of natural gum robber for fhe liner in constuction of the hydrocylones ‘ecommened due ois felatvely lw cst, excelent rear ease of handing, and non-ragiiy Dyakkowsks etal, 1994). The metal for the frousings are to be made with high-carbon sleet because of is good quality Due to the fact thal the feed slum being delivered to the cyelone should be as stesdy 2s Doseble wih regatds to both volumatic flow rate Eh sry densi, a steady pump operabion has to Be used. A sumppump system for = cyclone ‘opteation which should have a sump win 3s much opin a6 possible and @ minimum cross-sectional ‘tea consstent wih the ump, ls selected (Rictema, 7051, Lynch ea, 2002), The bats for holding De slurry por to delivery to the fnyrcyelone » ae fesigned, for convenance, not lo. cover the fmited. space in the erea of operation. So, it was designed to have wating time of less than 10 fminsles before the slury is. delivered to. he Idrocyelone for seperation (Dyskowski st. al 1694), From the cost evaluation of yrocycione, it was discovered thal a hydrocycione equipment is very Gepensive This then means thal a wel-eslablshed Smpeny_ only. can embark upon. such project (Gupta, 2001) When this method of separation is Compared wih other methods, ke cay bat and a Separation mathods. sereening and sticking devices, tne can discover at tis method is quite exoran but since ican be applied in @ continuous operation fan can handle high capacity per hour, these Compensale for ils exorblance.sed et i Table 4: The Rasuts ofthe Kern Hydrooyons {Skea ie Components Values Cys Diameter 36.iem ‘ea ofthe Inlet NEE AEST ‘ortes Finder Diameter Enlire Cyindca Section Leng “The inclided Angle ofthe cone Seaton pes Oumar Teh ofthe enlanas ‘Wi of he entrance. Length of ie Ext Duct ys 388) Giher Selection of Paramators | Values in Kernel Hydrocyelone Flow Rate (Vourete) Boar Veloaly Znsee Pressure Digg (CaTeueS) Topi Pressure Dros (specified) ‘B.3pst aeetRppteatn) ibaa rans 32 asGmisens ateral of Canaclion and Unar Tigh Carbon- Staal or Housings and Natural qu uber far ner aterial Balance ta Kernat Values Hycrocyelone| Tallow Feed” Volumen fow e oT anery ——| Soma Valrvate How fae of Water [Sn Valorie Row ral of shel [TO Volume fow rae of Kemer ‘Specie gravy of he shy 48 Tavern Volunsiie ow lear Siay | S035 Voumnere fw rale of Watec | 28m%e Voiumelefow rate of Shall [0.026 Valmet fw tate of kernal | 49m Speck Granly athe Stony ——| 1072 Underiow Valine Tow rate of Suny] OTST Vous Faw rate of Water | Ti Voluriate Tow rae of Shel] 9 3757 Volumeli how ale of kernel —_[O.tr Speci gravy of St Loar 41 & NJERD, Vol. 5, No: 1, 2006‘Table 2: The Results ofthe Shel Hyarocyclone T Sie af the Shell Tyaroayelana | Values Compone [—Sidione Giamaier ran “area of he inet NES a7 er Vortex Finder amt 10:75am Entire Cyndrcal Lena a28em The lneuded angle of ha Cone Seaton [70 ‘Apex Diameler Soba Oiler Seloction af Parameters mn Shall | Valuss Hydrocyclone| Flow is [vaiymetiey eM Veet Dave. Pressure Diep [Cae] Spe Pressuce Drop (specie) psi ‘Gig (Apeteston 3a 7S i (base) 27 Semirens ‘Maleralof Conaicion and Uiner High Carbon-seal for Housing and. Natural gum ubber for iner 3. Matera Batane in Kemal Hydrosyelone Vales Tintlow Fea ‘Voluneine Naw rate o Sar Sony ‘Volume fw ae of Walar— 230760 Volumes few rae of Shel Seesre Volumetric Now fate of kernel ‘Oimine Specie qrviy ofthe Slur Tases, TEOverfiow Volumes Raw tte of Suny Fae ‘Voiumtie fw rate of Waler 20 S26 alumi Row rae of Shel ‘azar. ‘Volume flow ala of keel Ome ‘Specie Gravy of he Str Doz Te Underfow Volumetric fw eo] Sura Tra Value flowrate of Wale ‘Son ‘Volumels fo rale of Shel | ‘Volume flow ale of keel [ome | Saecic gravy of Slt 115504 40 g NJERD, Vol. §, No. 1, 2006Table 2: Cracked Mtr Dimensions: Values Capsci 5.3330 ‘Shape The lop should Be Gyindeeal andthe tor resum Tegharne Cynaneat Section [5m Biameler of fe Cvingheat Sector [2 06m Capsaty ofthe Cyindeal Secton [5 On Slant Haigh of he Frusium Secson_[ 002m. Banealn Diameter of the Frostum | 008620 Secon Sapseiy atthe Fassium Sedien | 335o0" Benmaler ofthe Top Circe Sam Perimelerof the Bollom Cucle | 03020 Aiea ol he Apex ent ara Matonal af Cnstiucion laslomer Watered Such as Neoprene Table 4: Shell for Seconds) Bah, Results o is Capacty and Dimensions Dimensions Values Gapaci 37m Shape ‘The fon snow be ytindical na the ator Ws TishTaf he Cyineneal Seaton [1.5m De aaieraltne Goindical Seckon [tm | ‘Capacity ofthe Cyindncal Secion | 34mm Sian Haight ofthe Frusium Secton_|7.002m. ‘Beneath Dame's of he Frastum | 0078m ‘Sacton ‘Capacity ofthe Frastum Seaton | 227" Panilor of he Too Gucle 5 3am PEnmeler ofthe Boum Ciel [0.24 Area of he Apex ext ‘0007 Maienal of Const laionier Matera Such as Neoprene “Table 5: Spéctienton of the Kernal and Shell Hiydracyelone Pumps ‘Dimensions —[ SH Pump. KH Punp | Ty Pe ist PC {60K Cansei Sate 0mm Head 25 25m ‘Speed 600 Borsa — [500 = 200m ier Se SDK ‘Table 6: Quotation ofthe Price ofthe Hysracyclone Pant [Remal Ryaracyelone | WE Aa 49.00 ‘Shall yaracycione [6.352580 00, Exchange vale of $i per N130 49 BHIERD, Vol 5, No.1, 2006 Water Baty Results ofits Capacty and Dimensions4.0 CONCLUSIONS The design specications for a high capaci) hydrocyclone. separator for palm Kemals from is hell has been estabished in tis study, Although {hs spaciieatons have not been used to construct 2 hydrocyclone plant s0 thal the efficiency wil bs onfrmed. the information provided wll go a long (tay pling up 8 Workable plan for this operation REFERENCES: Aderum R.A. (1996: The Siting Fyoeysiones, Krebs Engineers, Mento Parks, CA of ‘Bradley 0. (200%): The Hydzocyslone, Pergarion Press, Oxlord Coleman D. &., Thew Ml. and Comey D. R (1880): Hydeecyeones for Oil - Waler Separation, Int Cont on Hyszocyciones. BHRA, Cranfield byakowski T., Hornung G. and Williams R.A (1994); Simulation of Nen-Newtonian Flow in & Hydrocyelone, ChE RD, 7244, 513. Gupta J P. and Grover P. D. (2001): Optimum Design af fyarocycione, Chem. Process, Landon 20(6) 30-38 Hartley C. W. 8. (200%) Group Leited London, The Of Palm, Langan 60 NJERD, Vol. 5, No. 4, 2008 s Lynch A. Ju, Rao T. C. and Prisbrey (2002) Muence of Hydoeycione iameter on Reduced Etfeney Curves, inleratianal Joumal of Mineral Processing, 1.173 ‘Mular A. {and Jul. A, (2000): The Selection of Cyclone. Classifies, Pumps and Pump Boxes for Gening Cirats, Mineral Processing Ptan! Design [AIME, New York Muthurajoh RH, (2002): "Palm Oil Factory Process Handhook”, Research Instiie, Malaysia Olayde $. 0, and Ofatunbosun D. (2003): Trends Gnd Prospects of Nigerian's Agricultural Expod, Tada University Press, Plt L. R. (1988): A Mathematical Mode! of the Hydrosyclone Classi, CIM Bull 6, 114 Rejagopolar S. and. K. Basu (2000): Theory and Design of Cyetones, Chem, Age, India 271) Rletema I, (1961): Performance and Design of Fiydreui: Cyclones, Chem. Eng. Sel. 15 298), Tarr D, 7, (1999): Pracical Applicaton of Liquid Gyclones in Mineral Dressing Problems, AIME Fal Meeting Trawinski H.-F. (1980): Conference, _Cambvigge, Engineering, 178 Hycrooycone BHRA Fluid 4