Document byemRzv2LdoeMX0Zo8YOqR8GD

inter.office memo TENNEGO C H E ?/l i G A L 8, INC /H C- -inal?' To Mr. C. G. Thompson frcjm pr< A. Lobo Subject SCALEUP OF PILOT PLANT RESULTS - VCM STRIPPING OF SLURRY AT Piscataway AT Piscataway date September 25, 1974 y to \ Mr. J. Fisher Copy to Dr. R. T. Gottesman Mr. R. S. Miller Mr. W. Miringoff We have all been concerned about our inability to duplicate laboratory and pilot plant results on stripping of VCM from PVC slurry in the Burlington plant. The attached article points out the difficulty in translation of pilot results to the plant and reinforces our thinking that power input maybe too low in the. plant strippers to afford good mass transfer. I understand that tests are underway in the pilot plant to determine the effect of horsepower on stripping efficiency. /jg Attachment P. A. Lobo l COLORITE 017164 Scaleup of Mixers Gas-dispersion mixers are widely used in fermentation processes. Here are the problems encountered in scaling up laboratory data. GEORGE H. LEAMY, Philadelphia Mixer Corp. CHEMICAL ENGINEERING/OCTOBER 15, 1973 Dispersing gas into a liquid is an important industrial mixing application'. The role of the mixer is to develop sufficient liquid flow to disperse the gas throughout the system, and to provide sufficient shear for optimum gasinterfacial area to give the greatest mass-transfer coeffi cients. With the advent of new pharmaceuticals produced by fermentation (during and after World War II), the growth in gas-dispersion mixers became a prime factor in the industrial mixer field. Many of the early mixers were hit-or-miss devices: little thought was given to gas load ing factors, optimum impeller-tip speeds, and total horsepower requirements. After the initial flurry of activity died down, a period of evaluation and optimum-mixer selection followed. Most of these early evaluations were carried on inde pendently by a number of pharmaceutical companies-- this resulted in widely varying mixer selections being made by different users for what, many times, were iden tical applications. During the last 5 to 10 years, a compilation of data from both external sources and laboratory work has been made available. These data, combined with practical information gathered from full-size mixers operating in the field, have led to meaningful mixer selections for gas dispersion. The usual approach to production-size fermentation tanks has been to scale up from a laboratory test vessel. This has often proved to be a difficult step--not because correct scaleup is unpredictable, but because of a misun derstanding of the factors important to the proper transi tion from laboratory mixer tests to full-size production vessels. Analysis of Scaleup Factors Investigation into the scaleup of a gas-dispersion sys tem must take a number of related variables into account. For a gas-dispersion system, shear stress, pumping rate and horsepower per unit volume are the variables most important on scaleup to full size. Z cc AW'D5 Q <x kND3 Y1 Z/M (l) (2) (3) The value of k in the case of mixer horsepower (Z) must 115 COLORITE 017165 gas-dispersion mixers ... Nomenclatur C Impeller off-bottom distance, ft. D Impeller dia., ft. h Tank height,-ft. k Mixer constant varying with impeller configura tion and overall liquid system. For gas-dispersion operations, values generally range from 2.5 to 6.5. kia Overall liquid phase mass-transfer coefficient, lb. mole/(hr.) (sq. ft.) (lb. mole/cu.ft.) N Impeller speed, rev./min. Nf Impeller power number, kt\/1>\V\ dimension less Nr, Reynolds number, dimensionless P Impeller power in ungassed condition, hp. Pe Impeller power in gassed condition, hp. Impeller horsepower Q Mixer pumping rate, cu.ft./min. T Tank dia,, ft. Z Mixer power, hp. y Mixer shear rate, sec.*1 p Liquid viscosity, cp. p Liquid density, Ib./cu.ft. be evaluated for both the type of impeller and the overall liquid system. The same will hold true for k in Eq. (2), Since shear rate (y) in Eq. (3) is dependent on the vis cosity and P,, any changes in Pt will cause changes in the shear rate. Therefore, the maximum shear rate occurring at the impeller tip will increase in scaleup to higher values if impeller peripheral speeds are increased. Experimentation1 has shown that gas bubble size is a function of impeller speed. When the impeller operates in a speed range where it begins to approach a Reynolds number of 1,000 or greater, a decrease in bubble size will occur. Further work* indicated that the impeller blade width has no influence on bubble size at constant peripheral speeds. Variations in blade width will, however, give wide variations in the power drawn by these impellers. Increasing the blade width will give higher flowrates from the impeller and a compromise must be reached between acceptable Pt levels and turnover rates in the tank. This turnover rate must be sufficient to bring all of the tank contents to the high-shear zone near the impeller, at a rate greater than the coalescence rate of the dispersed gas. The writer's experience in evaluating data derived from laboratory-scale gas-dispersion tests from many sources indicates that too often the test has been con ducted with too small a vessel; e.g., 1 to 2 gal. Although impeller-tip speeds in these small vessels are kept within acceptable scaleup limits (800 to 2,000 ft./min.), the pumping rate of the impeller in relation to the vessel size will give turnover rates far in excess of any rea sonable scaleup figure. This will produce several undesir able side effects that will lead to incorrect data and conclusions. .Small test vessels usually have a liquid height equal f to the vessel diameter. As mixing intensity is increased \ to the point where NRt approaches 20,000, surface aera; tion of the liquid will occur. At NRe of 100,000 and above, severe induction of surface air will take place. As NRc approaches 10s, all sparge air can be turned off without any noticeable decrease of the gas interfacial area in the liquid. ! This phenomenon is not experienced to any great extent : in production-size fermenters since the height-todiameter ratio is large enough to preclude any appreci- / able surface aeration. Small vessels also experience high turnover rates that often exceed 50 or more turnovers/min. This produces rapid circulation through the high-shear zone close to the impeller at a rate in excess of the dispersed gas coales cence rate. Since Q (or tank turnover) is proportional to the function of ND, the turbine bubble-size formula: D. oc (-- V-^-- (4) illustrates that any increase in pumping rate will give a decrease in bubble size. Thus, any excessive pumping rate in a small-scale vessel that cannot be reasonably scaled up will result in a gas holdup rate and bubble size decrease beyond the point of scaleup. This results in high masstransfer rates that will not be obtained on produc tion-sized tanks. Another factor to consider is the effect of proximity POWER CURVE for standard baffled tanks is applicable to gas-dispersion operations--Fig. 1 116 COLOR!TE 017166 ERSION MIXERS ... >/r= 0.375 C/Cf- 0.75 Z/T= 1.04 L Test tank dimwuioni; ' Inside dia.',35'/i in. ) Straight side, 36 in. -j Depth of dish, 5 in.- ' j 3affles, 4, each 3 in. wide' 1 ~ I___ rr Full-scale tank dimensions;' Insidedia.; 144in. ' .-\zr Depth, 215 irtt ' Baffles, 4,,_each 12_in. wkfe-, ! .. ... - ':7VT-0:375 : C/Q-.^`(tS98 . 'rpaWScf fi r ih ih* SCALEUP by direct tank-volume relationship--Fig. 3 .ion effect on impeller power--Fig. 2 ,ler blade tip to the wall of a small test-vessel, ment of high velocity flow on the test vessel luce an added shear rate that will not relate isels on scaleup. dispersion operations are normally done in tanks, the power curve for standard baffled I) ^Aplicable to this type of operation. ae^P^ shows that as NKe increases, the flow :s change from viscous to turbulent. Since kP, DWp (5) wer number of the impeller will not be 1 NRi > 1CH. Thus, when gas-dispersion labare conducted in a 1 to 2 gal. vessel, it is event surface induction of gas due to high ;s, while at the same time NRt below 10,000 insistent power levels with changes in im, since NRe remains in the transition zone. sumpti n of Impellers s dispersion in fermentation, industry has aded, disk-type flat-blade turbine the most of impeller. It combines reasonably high th good pumping rates. Its most important disk, which forces the gas to flow out to zone at the blades. Although uncommon, ers are sometimes used. Also, adjustable d 8-blade impellers are often specified to n the impeller when the product is varied onditions. i of data obtained from actual field units tests has confirmed to a great extent the bank4. Applied to both large and small cunwliowii in Fig. 2 has stood the lest lot^^B the correlation between gassed and ungassed power as a function of both gas rate and impeller speed and diameter. Although some previous data have been presented concerning the effect of gassing on multiple-impeller installations, nothing definitive has been presented. The writer's experience, both in laboratory tests and actual installations, indicates a modifying factor of 0.70 to 0.85 of the Pa/P factor for a single impeller should be applied to all impellers located above the lowest impeller. This factor assumes a spacing of at least 1% impeller diame ters between impellers. An important point often overlooked in fermentation scaleup is bubble residence-time. Usually, scaleup is based on geometric similarity of the vessel and impellers. If, at the same time the superficial gas velocity is held constant, the bubble reside nee-time increases propor tional to the tank size increase. Oxygen depletion of the air will result and has to be considered in scaleup. The mass-transfer coefficient is proportional to the gas volume used, and it is important on scaleup to retain the same or smaller bubble residence-time. This can be accomplished by using a direct tank-volume relationship. This relationship is illustrated in Fig. 3. Bt References 1. Wesserterp, K. R., Thcsii, Techrusche Hogeschool, Delft, Netherlands. 1962. 2, Uhl, Vincent and Gray, Joseph B., "Mixing Theory & Practice," Academic Press, New York. 196? 3, Moo-Youns, M B, Ph.D. Thesi?, London University, 1961. 4, Calderbark. P, H., Trans. Inst. Chem. Eng. (London)^6, p. 443 (1958). Meet the Author Georg* H, Leamy is District Manager of Soutneas'ern Sales lor Ptvtadelphia Mixer Cor? , Schuylkill Express way, King of Prussia, PA i9*0S, where fus prevous job was that of technical director He was associated with Mixing Equipment Co for a number of veers as a process engi neer and ss *-ea<J ot the sample test ing laboratory He received his 8 S. tn chemical e^ neenng from Clem* son Univers e OCTOBER 15, 1973/CHEMICAL ENGINEERING /1 i 1 1 I i COLORITE 017167 & RfiPH m ( COLOR!TE 017168 Or R Pi PH BZ COLORITE 017169 / ' (5^c. z>.a. ^/- c.p) /My p<y .Coa-j7},v,'~____ ,, _; &LSHO Tifir _ , ~3/-Ls'/:y___ ______ _ .___ / <?Zs^>/_i.Qd&s--'A, _<y . _ . 55 -7^%' f_ _ __r-o a,5'_y, 7______ /"/z?, \o _Z/5 1i'J' // ______ tt 'Vs^'O /?!_ _ /y-5 t.9 . _________ z^i.s _. __________________________________ 'Z Z?SZ?i>5 " /Z/> /r ?./ Mo, \S 2-77a, ' // >/ Mo, . j -2so / V S -y>77 " // * // ' /S.3 _____ /js^,, y &-2H / Z7Sto / 275ccC> / if %7>c>7 " /V// " /V/f " A/'? //.o ___ .5.5j? _ J$aa__ Hnog C,h L _/0. /_ // 5^ _______ it _______ ^------------------ 2yss62 ^ ,?b ________ 36Sfw______________________ ________ ____ Zt7 l r i 277^9 ' X7.52*j? _______________ " 1-r ! J L MP&tO 7/7 t *5/*z/g,y -55 /go +^o s*L it n - /o2 spy -27sa&f,, . j:__> __ __Ox/oz " zy ._ P7.Cf^/iy ..,, . ... __ it _____ _____________ jPjS'b* ___ _77Mpo_9___ Z_.2/. . . . yy&of/y _ ___ Z77p_o_J____ ` 2/, ____X7^^e7____ , 2f )#'*. n _ '______ O^'O/y) 1 t* _____ Z * - .,/.& . _ . 3?o --------------- .,,. , ___ 2?Z ... //.F ----- - . --------- ......... ., .zySapTf .. " M 7 73on - 2-/SO/Z " .2* " 2/ _ fay fayfaT fay Z=>r t ..................______7f________ _____ 2? ft. Z/SeJZ " ^y ca/3, fay ' - J/f - .. 27 76! 7 ... " 5?2 . Z75of Z?7oj s' ....fa/te _ ' *2 a > f/caz> fa) " (V5 fa\ s\,/37_ ___ _ /" it %/ ........- -9t_________ ?,/ it /o COLORITE 017170 SaaA/> /o7/*rcM ' pe**f. Ak '-------- T/flf-------- --fay Po7 --fay - . --ff&jX________ `isPP Z7Sqj7 1750JZ " *3 " 23 4fey fay -ZJS&jL_____ " J3 .... _____fay______ __________"^?3_ fay_______ _ 27^*/ 2.J2./ 2.7 son 3/zS-Z1> fay_ i!lt- 'X150L3']_ {2of&'T/p/'/s____ Vest P/w ,S2c/#siy_ 3S ^yt^Pi-ZOdPAL. 6-3____________ J/9 </ 6-0.... _____________/52 tt .3f* 6.0 /* s>? .z r</ 6.7 6,-0 /3r fa /oz /ST /c? 12.3 t>.s COLORITE 017171 - /zs 2{2//7i (&&?. &8. 1 ?/> Pop/S/iTcr/7l_ ._fc&s./A._____^ jyps &s/0. ></ &*#) zffs/sr _ VcAf PPM w5/. 397/2?... .... 3/22-37__ .My.. 4*7 Ss^P/OjpUpS6<Q.p /P.p.. 39-jdof _____3/22-37 . 3-//______ ______3/22-39____ "79 3/32'52 _____ T-3_____ 3/22-37 _____ .3J62-3P 392a az 3/22- 3r T7c//3,7oT /a< SS5>pP it?o6w /' J2V _J23_ 7 // &7L 3?2m3 3 93a<35 / Z>95'C&3 3/22-35 3/22-363/22-55 fe/- &p#2p?L- .-27*/ ..7X_ XT. jXX -/ 3_<7$po3_____ ______ 3/22-35 __P^sS/L ..._ <5-Xt _____3/2_2-3S_____ 3-7/o 3/0?- /<?*> >73. T/ 3966a Y. 3?pz&P .3/22-36,. .3/22- 3> _/cf $cp*c\ / y, T /5" COLOKITE 017172 \U>0 ; ://5 / i H` | t/l i tei ' Z35'c? IJ : 23 soo 'L .\235to3\. j 23 \2^5t>oS*\ | l*>5oa6\ M^*'r !--P^ J 3,fr3j - n\ ii "\ ;I I 1- . U\ -\ I Us 37/j .. . 3hti rllDlTX i%5po\7 \ f=;Okrti\V! \ ; ;| I 3/2q-ri'-f jj3/2r!~A-^7 I ! !. . I ! :' i '; ; i >1 r!, i ' ! ! i i : I ! | i j yiPo-rtpHs if ti \-- j ' >t // a n it h b'/lje/l, !! h\i *; iP i ! /<!i I 4i 'I I StojeA^, m! ! p n r "i! #/i. H ij PI pp/epz -/, y / / D. ;\I ; , jTT/t; fiJ \!cPu^A7uVr> ! is<z>i\OsJ? ) j-tL.UAAY A. <->Ai 5S~]c/6rF/tSl/Sc?6^ ! 7; 7j v! ! // I- - // : lio` I \fti\ !i ! !" ! i- ' 7,7 7:3i ti !` ' i /( 1: 41 ; "\ ! ! \i I"! !1 ": i // I 1.0 . 37: s.ii i\1ss1 7Zg . jrsr/... .7^ i PP i 4s - P- Ip- - 17 if i ii ft ii !!i it !I `\ !i COLORITE 017173 Zoo'2 -y (5ooc. &s. -> &?) >47?_ _ _______ 0/c.________________ ZZ*/ S'rrir/sZaZ /cm om 5/v^/__ ~7<%r5/*s Z /3-Ov______ 303?-OL-./S____ Sk**/ /95ae/ H 3 /f33a/ tf V 3 /9&e/ ft ft % S'/v / 2 /fe&JL / 3 JOvooz *t 3039V/-/c */ 3/^z'z/ / Zc" o/#.` Mv / ft / O /0jTaa3 os-pcm._______r.__ "_ fate/Ot)__ tv/is- ts$_ (/3aa_'' _ Ati) tO/z.S __________ fay " (o/co tfa) tO/^C V ^ 1/30 f.z" ...._______ //cV yz 00/ ________ m /9-5poj/ _"_^*&. _____ Jo9t^ /J$5tesr__ _______ 3/3/M/^ :_____ ___ "5537^1 rc. /fry. 7.7 ____ ____ 13 & /9Coo&>_ J!.................. ....... ....J^W, 0<y_____ _j:ZjL too COLORXTE 017174 COLORITE 017175 ------------- (50 ] -jot- - -----SsJ ..0.. \ *> i 'yrye~.._ J?:r _Hom.0M.L"l.wjrxz./z yA' 5" .:..................................................................... STPsPPA'fj U PP V V C M vs : ; -------------------------------- pPV PtLObUcJ-r 'V^CM1 ' ; i' 1 1 : !; '. 1 i , vj X COLOR!TE 017176 -o IM_____J3DO________________^QQ_____SCTJ.1________________7?_0.______Q<?._ t'm va&y.ixc ai .^Lem .L.^_ ______________________________ __ _______ _____ . J_______________________________ _________ 2 ____________yu/./rn, Rt=^ID U/1L____RRM___ jjJ___Er_S_//y_.________________________ ______________________VS.____________ -______ PGLV iZl^aJp -y-y rA /=~ COLORITE 017177 sotfv/ncf f aJ tuS PE^ScOO \fcm pv.&S (OS C ^F?// PPG P PC ~sosu) ^flODL'O 1 ripe. \n to K>l>i'A 0 &R 0 F S>4M Pd i a/ts AwF.^'VCff Vcm Sp^O s *4.? cT" C1 G OAaJC.E. SPfm ) AJt> "> 5L*7 H3 0 H32 n 3-7 / ?J <U -> /U t> 72 CG _ mz _ _ mo _/ud- n 55.________ OStS 30J<? .J.3-0. _ 16.^ /u 0 72too--2 3.0 COLORITE 017X78 ag-mg* s-ruo y V\S P^R^iO/J rbs/a/s 2/2//?* Ra&OuAL VC/a (PPM ) __________________ ___________ tjfo /73Q t7& /73 , 1730 \/73Z /73Z \j7vz j/7'/z /7Ji* C7/C- /7sr 1 ! / 27.6 Yo 39 A? ,f 66 33 // 7- 7J c.9 7Z ! / / /Z -- -- 63 3.9 Yz 37 'C.9 3.f 3-7 If i t ' 3> -- ,-- -- -- 3.7 y.a 12 / -- -- -- -- ; 9 -- -- 2<? /7 3 .3 /7 - -- -- --- ! 51 JZJ 3 v5: z _i_ <s AT.o Z-S // A/.0. Yt / -- -- -- 7.9 fa Y* A/A, M-0. -- MA. Ma, M3, a/A 7 A3 A A ---- ---- 3,3 i .....r, a> // _ --~ - 27 3.2 fa ^,jr />7 Z.o 6.% / J,f M.O, .7 y./ A3 Af.O, ______ 6.7 Yc S/A. COLORITE 017179 COLORITE 017180 G-RAPH I COLORITE 0X7181 GRAPH U COLORITE 017182