Chapter | 12 Optimal Design and Scheduling of Batch Processes 120. OBJECTIVES scheduling batch process. I Toc sae, Then, reactor separa ne examined, where tde-offs ext the rection conversion, ast wires ith rection time, the cast of separation wh 5 “After sting tis chapter, the reader should 1. Re knowledgeable about process unis executed in batch mode and approaches 3. Reale woschedle recipes forthe preston of singe chemical pos, ‘4 Understand ow to chalet plant fo the rodution of mulpe product 121 INTRODUCTION sis forthe man acals, phamaceutalsChapter 12 Optimal Design and Scheduling of Batch Processes 129 OBJECTIVES “This chapter inrouces strategies for designing ond schedaling batch process vit single equipment items focsing on methods fr achieving the optimal batch i tion tne the cost of separation, w Time. Subsequety, methods of 32 pea items ae cold ri, produce in pr nits executed in atch mode an approaches 2 Know how to determine the optimal reaction tine fora batch reactor proces. 3. Reableto 1 Understand how ies fo th proton ofa single chemical product ch plants forthe potion of mip prod 121 INFRODUCTION Continuous poceses are dominant in the chemical res industries fo the man omit che lass, pelcum produ, he mansfactre of Specialy ch macs proeses ist ay ited schematically in 1a. Inth Eine te it tvo pe with ing Then, when the processing ised, the products are removed bach, abegins ve an ch de suing oa pte ole ise of Figue nd the poses J om - gue 12. In atch produc removal he sheila et th process efor: rose Ing bec nd ops 2's Conant a he pro ved conte th prceting cn ar own n Figure 1214 i el, febach and cred tno pcre are cicontinots poss he calng in deigng 2 cfc, ob profat removal pros it in de cing onthe of he esse a th pecs ine Ths coped frie ate 0 presi wher te fw rand concearaon of foc sue re fow ee fro arama fenton of ne only infere he perfonnanc of he proces NO thatthe determination of opin opating pots teed te soln oS pt inl coo poten, Ts njsiaa inthe net eto ach col eclvoniomos prcet se ied fen when peoduon ms ae sl reece tne ae lg, and ott dana is erate, capecily wien he dona fora chemical interspersed withthe demend for neo mre ther rd and he at ts weed nl he ning o orders re uncer. Even bene demand ona thd he prodacon resect nrg 9 joy conan proceso, batch and386 Chee 12 Opin Design nd Schlag of Bach Process semicominous processes ie often designed to provide a elabe, hough isin, tothe proton of chemicals, For exams, i the emulsion polymerization of resins Tote actos a installed on to aod crying out hee highly exotbermie reac ndbuous sed tank reactors, Note however, dt while operation at low-con Steady sate fea es profitable tan batch r semicontinuous prooesing, operation per loop unstable tad) sae offen more profitable. Rater tan install contol {0 sable the operation, may companies prefer to operat in bch or senicont mode, Sil, design feats often opt fr batch and semicontinuous process Wh ‘hamicals are hiatus of tonic or when safety aspects are of great cones, Hecause te desig for coinuous and bach processes ae uscally very fee choice of processing mode s made commonly dusing process yess, nthe takin tion scp 2 dcused in Section 3A, Athi stage, th decision to reject coatnuous pe ings Bised upon ss of thumb, ber thins dtalled comparison ofthe alte “hrongh pocessitmlatio, as Snes in Chapter 4 sd te optimization method sented inthis chapter, more slgoithic methods sre avalable fr selecting fom ano “aris bic ad contnunas process. ‘Unaly. for the prdaton of anal quaties of high-priced chemicals, such 5 imamate of pmaceutial, foods, electronic mater, and specially emis, fadbath, and btclprodctreoval process are prefered This is fin the case processing, for example, when drugs ar syatesaed in a series of chemical reactions Faving sal il nd eqiing ica separations to recover small amount ofp This sloth ease for tang! facilites in tel, which prepare foods in batches, ‘any uni operations in he manfactie of semiconductors. As discussed in Chapter {Dthese processes wily lvelve wrap, tht, Sequenceof rss, tobe cated fers of equipmeat nthe later sections of thi chaptr, vations on ath p rs for optimizing the schedules. Schedules are discord, wells el 122 DESIGN OF BATCH PROCESS UNITS Batch Processing ‘When designing » proces wit to operate in batch mode, its usally desired 10d sor which is wsely expe a the vole the butch me a the size miso proc tnt maxinize an objective ike the amcuat prog. To accompli a dynamic model ofthe proces nite formulated andthe degrees af feedom aust Tented in the examples tt fll Ae wl be seen, thea many ways to fom inal coiel problem, Fo simpy the dieusion, modes are presented and so “tious appt, to sce how thy afc the objectives. Emphasis snot placed oa methods of aptinzation For conventional batch processing, with no maria rans © oF fom the bach, Iman is fen improned by along te operating variables such as temperate lon speed. Trough tse alten recor coaversion is improved thereby redo Tc time o achieve the desired conversion. An example x resented ext ht sho {oschicve hisobjective by optimizing he temperature drag batch prosesing 121. Evothermic Bach Reactorte 2 te "ie eh sc, i asi his ed fie on sto ts ng te he sourrioy nth pe opting empties the i bh in. Tenses Based pe te ceveopment by Den 1903) The eer he osu rene ae hy itrttion E12.) wih as) Fed-Bateh Processing Femeatton process for he poston af ug are ually arid ot in fe atch ee tors lathes actor, iis dsible oid the est pole for feeding substaeo the tmcning rsh as isan the ext xarp. igre 122 ToopEXAMPLE 122he ot oes 122. Digna Rah ices Unt 30 a3 7 éy iy i ige 124 Opal poles fer pein ace ‘Cust wo Biegh 198)maimite he poason of pencil Pos), hse rs he ch tine sng varon ess (ogi aunt cpl) ca he son gue 124 [Avon a he pina, he a ie 349, te pte of #705 of pena. he fare of nl 2879 when sh comtt 10 hh, Ta ol" il sigs cc. ‘olan wich fn credo he “opi ete pole ing ete collation mf ‘ic ements denne the lc ens a ese que pga (SP. Batch-Product Removal ‘When ditions are carried ot in batch node, the sls cargo with the fd mistre su the hot i tune on in the reo. The hight species connate i the dla, which is condensed and recovered in baih-product removal mode. Aste igh specie 6recovered its accompanied by increasing faction ofthe esver species wales {sagplied to maintain a igh conenation of ight species. Formaltiomponeatspaations, to ply operation itis oe saifctory oa meee wice while ecovenng each species. When the pay of pei Is brng colloid inthe product secular, dope below i session, he con the rut secular ped in ite pdt eer and the rex ates ‘Stated fre the fra increased the ffculky ofthe separation Btw Tight ad heavy key components acess This istated inthe next example EXAMPLE 123 Bach Disitation sissy ngs [ty bh lain eprint 12 Asm 2. Using etc ooo Shwe of hint connonty he prot 2raegy ot the bh ts of sonwney sted co ths ih 123 DESIGN OF RE. Tots! Step Total Sep Mato! Water "Geol" hm) th) th) \CTOR-SEPARATOR PROCESSES In ths section, an approach to solving the optimal contol problem is inroduce fr rec teraeperatr peceses. The approach inveves the simallaneons determination f the batch times and size factor for both ofthe process unis, Furthermore the interplay butwee the to nits involses te-of etveen them that ce ajith optimization It shold be ‘noted tht ple modes, than in aamalprcie, ar sed here w demons the concept ‘nd theft example, provide an antic slton tat is obtain with elatve ese [Comer te itch reaee-sepmatecombinton in Fire 126, ini resend by Rad ad ‘Wauoe (968 the xc, he tera, nevensie act, A'~='By cid is The ‘82.00, ht te opering cso the reactor are proportional 1 the tat ine (st, Co ~ a,22 Designo Rear Seino Pres 393 ‘son adaption cots increase de 8 mot cl separon, mare dn countering te decrease {he itemediate ble. Fux specifi reactor volume, V, and cok vokume, Vitis dried 8 ns) ‘Scns he Bch ines or he entra cle tn aie cn rocing te eae fount of edt Bl, ns ingle aye, Fling Barer and Evan 98, tea sph enn hte ils roe ae il idee fe hel. ‘et cutie ae wet ncling Gn fe fd, Pd, yl or A, Py a ‘Str forte Ppa C, Pte tlm of he ec (Sh) terme ae So ilistin soln 7 th con flaming eater, Ca, ad lan a Cy ‘Site ity oye ol ata, 2, These seta st gen a Tale 121 EXAMPLE 125 noes ns a am ea394 Chile 12 Optima Dignan Sheng of Buch Process P= 1 Sil P= 02 Stoo, Pe 04 2S 0 = 208 7 100 Se y= too ich, Ce 0 Stat = bat he Heo ere the ation tine (aso opertig ine); te a ine foe ito cl b= C¥APG + Rate + Pes One bth + B= [oul + Galt lipid gre te ot of wiles per ach. This ames tat he iis spied| 224 DESIGN oF sua ey Batch Cyete Times 24 Design of Sg PadstPcesing Sagres 398 _nizai proceeds tis oeessary to ait the folowing inequality canst: ° Poe lizo pogrmming(SQP) in GAM ac Seton IW) Te remo gerne = 4 ope 4 ‘ire speci, Tax = 24 he. These ies woespon ote nema cet, = $10,240. By varying SING E PRODUCT PROCESSING SEQUENCES ‘aving examined small optimal convo problems for bth process unis in Sstion 122 an foe race separate sequences n Section 123, sould be cea ta the dterriation of ‘optimal bch ne, given butch sizes expressed as bach volumes pe unit rs of produ {in be demanding compattonly. Since mst process in pace, have espe wth me Inevus tsk and comparable naber of processing unis (th PA process n Capers Sand), tis normally pot practic to optize the uch anes forte india pressing ‘us when preparing ehedue of aks and equipment tes forthe mance of prod tc Consequently, hen poping a sched of tak an eget tem, common to ‘Speciy batch times for tack ob peformodin specif italy wi ch sees, nd tooptmize cycle tines fors speci ip nome cases, sing the sof proc and els the ese are designe as wel: thar vessel sizes are determined to minimize the Schedules ate determined forth batch groeises that involve ony singe prot In the ex section, the metiodalony i extended for ml-pict bch pewsnes. atch proces design begins with the specication o a recipe of aks o proce pod ‘ect. Incontinuos processing cach ask iearied ot ina speci gue ter, wi one to-one corespondence betwcen them, shown on fowsheet tht remains fixed inte, Simlsy,m batch processes, th take are sesigned to sient en, but ove spi terval oft, which vary with atch ize, whichis oer determined bythe valable quip. reat size. For example, inthe (PA proces in Sections 34 and 5, given the rte of PA ‘roduton (50 pa PAV), where pe ae picograms = 10") and eel concer tion tetween 0.225 1nd 3% 10" ell) the vallabiity of «5.0804 cult termine the Id hatch ime ad the batch sine (2.24 kg of PA, pce in 4000 Lo ‘medium, yielding 1.6 hg of aa produc) for the caltvatr, As dissed io Seton 3.4, proces syess involves the cretion ofa sequence owt of operation, which can be refered wo asa veipe of operations ce toss, Daring the ask integration se, tas ae en combined vo be caved out in single equipment ten for example, eating snd ec Son in a pyrolysis fimace. Alo, ring this sep, the deision to ne comnuous ur bch rocesing is mde. AC ths pin, the avaliable euipnet sites often Sterne the Dich ‘Svs and tines ‘Whe scheduling snd designing batch poceses, several formalisms ae widely weds 8 ‘siened by Reklain (1995) n thi eston, and those ta follow, pros ofthe pesettion tue derived from his atl,In batch processes itis common for tsk to consist of «sequence of tps to Be cut in the same equipment unt. For example, Figure 12.10 shows apical recipe wi thks ad tp Note tat each ep involves a Bch ime, which is determined y the ‘esing rates andthe atch fr: that the amount of th nl product manufactaed Fetch Furthermore, «production fine sx set of expt ems assigned 0 thet reepe to proce «prec, When production line wed to produce a sequence a Cat atc, the eel ime iste time botween te competion of batches. To beter vs the schedule of prodicton, an eqeipment occupation capa, known a8 Gant hi prepared, showing the periods of time ding which each equipment item i uli own in Figure 12.1Ta Noe that the unt having th Ingest batch tine (6). U2, bonteneck unit sits always in operation, Note sso ht the second bch is beg in (o produce the feed t the unit U2, when the lata becomes avalible fer proces fist batch. This diagram, the blcher ae taser fom snitto-usit immediate Called zen sratgy, with no intermedi tore lized), Cleary, the cyl time iste batch me of U2 In he sched in Figure 12.1, the serial proves has snc tsk assigned 1 ceuipment item, Often, to wilize te equipment more efficient, i possible 10 u ‘Suipment item ocr out wo a or tasks, Note that thi mays possible when ‘lcturing specialty chomals hat ae very sesitve to contamination at nthe an tf pharmaceucals, Retin othe schedule a Figure 12a, when the furs sk acid out n UI, this unt ister wed nd Us canbe released for prodcton es inthe batch plan as shown in Fire 12.11, Noe tht to ahve this seed, w ning interme trae, necessary t0 retain the batch within US uni Ube ‘valle, Futrmore, to increare the ecieny ofthe schedule, tat reduce the time its common fo ad ene or more units in paral. When in has, is clear t chine forthe unt i redced to where he number of unis in parallel fr For example, when two U2 units, each half-size, ae installed in pl the effective time fori U2 dco ead the cycle imei reduced to 4, with U the neck unit Aertel, te paral units can be sequenced ovo phase, without the bate tines shaw in Figure 121Te tn both eases, the U2 botleneck is eli nthe eyle tine ie edwed to = CEhits = Za L mS es se Fema aEE he bam EDEL Hs ‘ammi = re rs e ARTOSEEE398 Chaler!2_ Optima Dsgn and Sheng of ah Prose Intermediate Storage Bate Size ‘Clery, without paral operation, te bach cyte ine, C7 the maximum of he ties yd = yeni ‘where M isthe mame f unique equipment unis. With units in pall an in ps jel ine given by cr a Returning tothe example, when tw units U2 ae installed in paral to perform tak 2: man $43) = 4b a Ths far, evo storage options have been luted. No storage ie usd inthe shod Figures 12 1a and 12 Ile, withthe content of ech unt anferred immed othe ‘unt experiencing no delay ae its tah been completed. As tenonod above, is so-called zeros (ZA) sttegy In te sche of Figure 12.11, U3 proves inter tte storage unl Ul Becomes available, Hence, a erowait aly i inpleented, ‘tos intent storage when necestry. This refemed oat an intermedia sora Stategy. The thin stategy invelves united intermediate rage (UIS), afc to the contents ofthe product from unit having lengthy batch dine to be used repeated ‘unt having half the Batch time oles, a llstaed in Figure 12.12 Hee, UI is al times and the eyele time is edoced fom 9 10 3h, Ta padace «specie tou product the Batch size ie reduced by factor of one-thied snc the eye time i i Tes convenient to dine the size Aco, S, fr tsk ja the capacity rogue per w rot Commonly is defined a he volume reget podace aunt mass Of po For example, fr ths thn alvaor inthe tPA proers of Sections 3.4 and 5, 4000 moti yields 2.24 kg of PA which eventually ils 1.6 ke of Saal tPA product. C ‘enya factor i 00D 1.6 kg = 2.50 Lg tPA proc Sie fcr canbe ted foreach task in a ecpe. Normally, eipment vest sizes ae seated tate Tnch volume by 10 wo 20%. Clay, the batch factor in volume/mass produced is ined hy the rat of processing the hth (K/h) maiptied hy the bath ine (hr Alvided by the density ofthe Bate (kL) an a as of ped produced (ks)125. Designo Mago seing Sogences 389 N OF MULTIPRODUCT PROCES ‘A mukprodes batch pant pues 2s of prods whose ele ttre a the sare, SING SEQUENCES ach 125 DESI 7 at i a maces ep cat (0 ait aa ee ee atl nase ok eng eat a ee aa = fee ces EE oat [freee pepeoeraresr tree aT _ obi ace Los = igure 113 Gat ha fos pups pa.490 Coser 12 Optins Desgn ad Sted of Bah Poceser, Cath alsin aS EET Ta DTIC Si eC as ees Ta FST (aa | = — - : onsieratons, In svg the problem, that etmiing am otal chal, t tier n which tases sete equipment and eure i deere ith speci ings the ts provid, tht ptinize te plat performance (wich can be pected in many way ‘hea the plant des not ens that, when 8 cower re esally not well defined. Otherwise, he specications oe identical In fact th design problem encompasses the skedling probit nba it solution involves deter ins the munber and cape of th quiet tems in adn ote optimal shee. the design problems thas ar Getrmined to ontinie an objective that includes the inves nen costs ofthe equipment, ssh a he anntlied cos, Because the product nde ot aswel known dig the design stage, ts comin to slve the shading problem Je Figo ‘insta oF 3s known fnetios ofthe bath sie. These can be ‘operation of each equipment tm as discssed in Seton 12 Tess cannon to fous te design poblen fora mtproduct batch plant involving proccsing of butch catpelgne ia eis Ge, onetime commonly refered to ax Flowshop per nnlinea peogram (MINCE. Then, be formulation Jn tmes for each tak ae often specifi dy optiniing ‘impli for solution using stategie tht are Beyond the scope ofthis book (Biegler et 1098). Heel av an indo, 1 pial formulation x presented witht simplicaion. a0 teins wih te objective nin c= $ map Je ny isthe amber of out-of-phase units sgn to tsk a ‘gn and Viste se ofthe ant ase to ak (silly in Ts 384 x ont octets) This objective x minknied commonly sabjet o aoquies tha ively tte prod (a he il produc sz, tycall ia hp. and The see favor frist) roungprodut (peal 0 ik). This nequaly ins thn the unit sie exceeds the smalls sae rege wo peeduce all ofthe products. Tn tim ower and app bounds ar pete onthe enupment sie i acordance with ma fseorng imitans:ties 40% Inequalities ae asia ls with he eee time and time hoion: $2 cneu =m where C7; ste yee time for producing pres which can be determined using Eqs. (2:12) and 0213), {sth batch ime for kin predicing product, Qs te anal de Inu for pec (ypcaly in kp, an the productos hous avaiable anual. 126 SUMMARY. Insally, this chapter focuses onthe optimal copilot processing unis, wi emphasis fn rocing the bach time and beh size, Then, the atch tines Tor reactor separ frceses ne optimized with emp on he interactions between the proces nis and the tolcoffs in djing thei batch Gres. Fall, the problem of dtnnining operating ‘Shuts fr snle- and prt batch pan ivlving the possiblity of itemedine Sore ad complex eipes with mmerous ass in numerous proces uit is examined REFERENCES. Seren, LBB “pn Oe ‘Bape, LT, LE. Cionimaan, and AAW. Weserbary,Sytmatic “Sa nt ba, mame Open a So EXERCISES 122 Fee he pein ect i Example 122 ing pts ecu ons [At {os he subtest ntee how the ply ems Mette ede role ‘Sty oven te einen ey use 124 tn Example 124, dive Ep (126 (12.7, Ten a 125 Ina 125, drive Bp (2902.1 Fae 1214 bt whe nk igen pected in Fie 12.13. Compe in eget he of ming, Sho ening oF Us, Dat and US, espestively, deteoine the eye times, and 1. When new pron vse UAA pres at Wo "epufcoon an ake cpr desi estePart Three DETAILED DESIGN, EQUIPMENT SIZING, AND OPTIMIZATION - CONFIGURED PRODUCT DESIGN