IChem® 0263~-8762/99/810,00+0.00 © Institution of Chemical Engineers ‘Trans IChemE, Vol 77, Part A, September 1999 PREDICTION OF MASS TRANSFER COLUMNS WITH DUMPED AND ARRANGED PACKINGS Updated Summary of the Calculation Method of Billet and Schultes R. BILLET and M. SCHULTES* Rulr-Universty Bochum, Bootie. Germany *Raschig Gmblt, Ludwigshafen, Germany A good knowledge of the relationship between the two-phase countercurrent flow of a packed mass transfer column is essential for the design of rectification, absorption and. desorption columns. Based on a physical model the authors describe their updated equations for calculating gas and liquid side mass transfer coefficients, pressure drop of dry or irrigated random and structured packings, their loading and flooding points as well as their liquid holdup. Based on one of the largest experimental databases in the world, the calculated results give only small deviations from the database Keywords: packings; mass transfer; mass transfer coefficients: pressure drop: loading point; flooding point; liquid holdup INTRODUCTION ‘nui afew years ago, mass transfer columns with dumped or arranged packings were designed on the basis of empirical calculation approaches which, although they were based on experimental data, could not be traced back to any physical context. These relations always resulted in unproven designs ‘when applied outside ofthe experimental database Modem calculation approaches are based, therefore, on physically proven conditions of fluid dynamics in mass transfer columns and take into account the kinetic laws of mass transfer in flowing media. The calculation model by Billet and Schultes is such a physically proven model forthe advanced calculation of mass transfer columns with dumped for arranged packings. It makes it possible to determine the mass transfer efficiency, the pressure drop, the column holdup and the load limits on the basis of a uniform theory. The results of this calculation method agree well with one of the largest experimental database for results of investigation on. ‘mass transfer columns. This database was compiled under the supervision of Prof Dr-Ing R. Billet of the Ruhy-Universtit Bochum and includes over 3500 measured data, more than 50 test systems and is based on measurements with over 70 types ‘of dumped and arranged packings, see Table |. The basis ofthis calculation approach is a’model which assumes that the empty space of dumped or arranged packings can be replaced for theoretical considerations by vertical flow channels, through which the liquid trickles evenly distributed downwvards while the gas flows upwards in the counterfow. In reality, however, the flow channels deviate from the vertical and are decisively determined by the shape of the dumped packings or of the arranged packing. Since the geometrical shape, however, cannot be only defined by the surface area or the void volume, the 498 ‘model assumes that the deviation of the real flow behaviour ‘of the phases from the vertical flow channels in mass transfer columns can be expressed by a packing-specific shape constant, CALCULATION OF THE MASS TRANSFER EFFICIENCY For a given separation task, the height of a mass transfer column is calculated according to the HTU-NTU model According to this, in the case of systems with predominantly ‘gas-phase mass transfer resistance, the product is formed from. the overall height of a gas-phase mass transfer unit HTUoy and the number of gas-phase mass transfer units NTUoy for the calculation of the height H, see equation (1). The number of mass transfer units is defined by the concentrations of the incoming and outgoing flows of material and by the phase equilibrium and is therefore essentially dependent on. the separation task and the thermodynamic properties. On the other hand, the height of a mass transfer unit is determined by the actual design of the column and the load of the intemals, Therefore, the loading capacity of the dumped or arranged packing determines the column diameter and thus the flow velocity and the residence time of the phases in the mass transfer apparatus ‘The geomettic surface area of the column packing and its, ability to create a turbulent flow of the phases also decides the effectiveness of the mass transfer between the phases If the mass transfer resistance is situated mainly in the liquid phase, then the height of the separation column is determined by the product of the overall height of a liquid phase mass transfer unit HTUoy and the number of liquid- Phase mass transfer units NTUo, in accordance withPREDICTION OF MASS TRANSFER COLUMNS WITH DUMPED AND ARRANGED PACKINGS 499 Table 1. Range of material propeties and column loads as well number ofthe examined substance systems forthe determination of fud- dynamic and separation variables Loading and flooding point Column hokiup Pressure diop Mass transfer (Gas load factor Fy ea 047-459 010-278, 021-508 0003-277 Ligui tod ™ me mh 488148 133-828 Osi-60.1 0256-118 Liquid density a kgm? 1750=1026 800-1810 361-1115 361-1237 gui viscosity mi meet 040-104 076-142 04-990 0181.65 Surface tension o Kgs? x10 208-863 07-749 Diffusion coocient Di met x10 029-650 as density : 030-137 006-28 007-97 Gas viscosity % mist <10t Bis-4i5 0.14-106 014-126 Ditfsioncoofcent Dp mts }s10F 029-874 [Number of ystems examined 3 20 2s 46 equation (2). The variables of equations (1) and (2) are linked with one another via the stripping factor, see equations (3) and (4). The stipping factor is deseribed by the quotient formed by the gradient ofthe equilibrium curve ‘mye and the molar phase ratio LIV. H = HTUoy NTUoy a H = HTUo, NTUo. @ HU. = HTUow 2) NTUo, = ¥ NTUoy @ ‘The following statements apply to systems with predominantly gas-phase mass transfer resistance; the results can, however, be applied also to systems with predominantly liquid-phase mass transfer resistance by means of equation (3). The overall height of a gas-phase mass transfer unit HTUpy is subdivided into the height of a gas-phase mass transfer unit HTUy and the height of a liquid-phase mass transfer unit HTU;, which for their part are defined by the 288- and liquid velocities uy and uw, with reference to the empty column cross-sections and the volumetric mass transfer coefficients Byay, and 6,ap,, see equation (5). ‘The volumetric mass transfer coefficients are calculated bby means of equations (6)-(11) and are dependent on the densities of the phases py, p,, dynamic viscosities ny, m, and the surface tension o,, the flow velocities uy, uz, the specific surface area a of the packing and the void fraction ¢ as well, as the packing-specific constants C, and Cy, which can be found in Table 2. The effective liquid velocity 2, is calculated from the velocity uy with reference to the column, cross-section and the column holdup hy, Furthermore, aj, describes the specific interface area between the phases and d, the hydraulic diameter of the dumped packings. HTUpy = HTU, + NHTU, gk © sgn (De) (te Bud = C12" @) a(“) © ra 6) ‘Trans IChemE, Vol 77, Part A, September 1999 4=(nlt ue) for uy yg an ea. ‘The relations apply to load conditions up to the loading point of a tworphase countercurrent flow, ie. ty ts ‘Above the loading point the shear stess of the. gas countercurrent is lage enough fo dam op the falling Am so that the column holdup and te sizeof the intrtace increase and the effective ow velocity ofthe falling film decreses. This condition canbe taken into account by the equations (12)-(16). For vapour loads less than uy.s the Tiga hold-up is according w equation (12) practically hy Hahn = SELB te ae 3) Pee) oe sas (*2) “aay (Mt) (Hea) 0 (BY OE ae for ys = uy < typ, a6) ‘These relations reflect well the results of the experiments if the surface tension along the rectifying column does not, change or if it increases. These systems are termed neutral or positive. In the case of rectification of negative systems, however, which are characterized by the fact that the surface tension,500 BILLET and SCHULTES 5 tonye pete ne “Operating tine ogo Ye rtm in ga Figure 1. y-x-concentraton diagram wo describe the overall concentration Aiference (2°) fom equlibetum-and operating lines for determination ofthe Marangoni number, along the column decreases from the top to the bottom, the Marangoni effect must be taken into account, since additional surface turbulences in the liquid-phase boundary layer reduce the size of the gas-liquid interface. This is taken into account by the Marangoni number according to equation (17) and the interface of negative systems is determined by equation (20). The first. term of the Marangoni number describes the change in the surface tension with the concentration along the column while the expression x— x" takes into account the fictitious concen- tation difference shifted into the liquid-phase boundary layer, see Figure 1. The liquid-phase mass transfer resistance HTUJHTUox can be calculated by means of equations (18) and (19). x HU, Ma, = “dk Dina HT, oF my, __x i as) Maps of DP a 1 at My 0, vi" Di? gM (eh) (yh) (19) (2 )aase® CE) ght 24x10" Mau") 0 Inthe case of exuemely large Tigi lous, th thickness of the liquid film increases to such large values that voids of the dumped packings or arranged packings are filled up with liquid. The column then operates in a fluid-dynamic range in which the liguid flows as a continuous phase while the gas phase bubbles in a disperse form through the liquid layer. ‘The limiting load, at which one reaches this load range, is termed the inversion point and can be calculated according to extensive studies by means of equation (21). The mass transfer at such high column loads can still be calculated according to the above equations if the liquid load is determined according to equation (22), see Figure 2. eI VV. 1, = lgh, = Ody vz 22) Only recently, mass transfer measurements at high pressures =04 for u, > 80m m?h* en Bonn Hilla 8, PP Leh Wh ‘Figure 2. Overall eight of a tansfer unit HTUs a8 function ofthe liquid ga molar low ratio for Rau-Flow rings and Hilow rings at varius iui fads, (ammonis-ai/vate, 293K, 10°Ps), were included in the experimental database. The compar- ison with the calculation approaches showed that the extremely low surface tension of the liquid phase at high Pressures leads (0 a large calculated interface for mass transfer. If, however, one takes into account the experimentally confirmed fact that, below a lower limit for the surface tension of 0.03 kgs", the surface of metals, plastics or ceramic is completely wet, the influence of the surface tension on the wettability of dumped or arranged packings is lost, This was confirmed by the measurements at high pressures of up to 30 bar. ‘The results of the measurements agreed with the calculated results well if minimum value of 0.03kgs~ is used for the surface tension in equations (9) and (15), sce Figure 3. The deviations of the calculation to the experimental results are shown in Table 3 CALCULATION OF THE PRESSURE DROP On the basis of the theoretical model already mentioned, the pressure drop of an untrickled packing can also be precalculated according to equation (23). It is dependent on the geometric surface and the void volume of the column packings, the gas load factor Fy =u /py which is calculated from the gas velocity and the sqlare root of the gas density and the wall factor K which, according t0 equation (24), takes into account the increased void fraction at the column wall, The resistance coefficient Yo, see equation (25), is dependent on the Reynolds number Rey of the gas stream, which depends, in accordance with equation (26), on the particle diameter dp, see equation (27), and the variables mentioned above. (24) (25) (26) ‘Trans IChemE, Vol 77, Part A, September 1999PREDICTION OF MASS TRANSFER COLUMNS WITH DUMPED AND ARRANGED PACKINGS S01 Table 2a, Characteristic data and constants for dumped packings Size w a fi Dumped Packings mo tm mG Ge Raschig SoperRing Mel =—«0.3—=«180000-=31s 9803560 «2340 0750 0.760 1500 0450 05 145000 «280 a7 3350-2200 0620 OTD 150 O30 1 32000 1600980 «3.491 «2200 0.750 05001290 440 2 9500 976 O98S 3326 «2096-0720 ASI «13230400 3 300 80982-3260 ©2100 062004300850 0.300 Phstic 2 90001000960 «3326 +2096 amm0 037712500537 Ralu Flow Phase 33000 —«165—=SC9]«36IZ~«2 OT OGD COMBS LARS _0360 2 4600 10009453412 217406400380 «12700320, Pal ing Met 50 on 26 09st 2725 ~«1s80 O77 L192 Oal0 35 19517 139409652629 .679 Gtk 0967 1012S 25 53900 22350954 26272083719 0957 LHD 0.336 Phsic 50 6165 19192816 AS70593 698 1.259068 35 yoo S195. 2658742718 0927 08560380, 25 52300 72500887 2696-206 «528 ORGS 905 aaa ‘along Meal 50 300 «105 =O =—=«TSCSO OTe OTH 192 OMS 38 450 1356s 229 679 Ot L003. L277 OSL 23 51000 215.960 «26272083714 9871840 0.336 Phase $0 57 95298323 LBIZ GH Od 1520 0.308 38 13500150, 09%) 28438120640 O67 13000333 25 36000 190d 21989719 aM 1300333, NOR PAC ting Plastic 50 7390 «68ST «2980 «1786 ~—«OS 0380 —«1080 ~~ OaRD 35 vaso MIS O94 3.79 «22857 OT! 07560825 25 5000020200988 3277 24726010397 OBS 0366 25 4897019790900 2.862083, 0383 0976 oto Hinowsing Meal 30 S000 237m snes ONT OaM—«LNGR OME 2 40790-2028 0.962 «2918-2777 O.68 Kat O402 50 oes 710925 Dada 10S8 0327) LaTe OMS Plasic Oye. 6590-1184 .925— 2894 LTE Osi 1353 036@ 508 050 2026s «702s SI4 1219 OE 2 4610019451828. 989 om 1577 0390 Ceramic 50 si0 97.808 Less 03381377 0379 38 eats 1930) Osa 16896 m Wiss 28620758 aS 2410 LST NTAK 6 isch ving Mel 30PMK 79200 ~—«180S «0975 —«24a4 «1.900 0830 OAS «1.920 —04s0 30P 3110016400989 2564 1.760 OS1—1.036 S77 0398 Giieh CMR ring Mel 1.5 coms «ACTA Ona is'T 6357 KOT —-2790 1870827 io" 15846723257 2.703, «1996 Lo Oat Os" seobi1 35600982 26M 2178 088 2038 0495 ‘TOP Pak ing Au 50 eel —«*10SS—=—«08S6~—=«ASTS «STD ORL ODM —1.806 O89 Raschig ring ‘Ceramic 50 5900 (9S0~=«C«OR30 ==. 14160210 25 477001900 ——oe80—24S4 18990571329 L361 aie VSP sing Meal 50 71 «1k ==] =A Cassada 2 3634 1996 07S ——2755 «197 1369 07? «137 0.405 Bnvi Pac ring Plastic 80 2000 ~=~«OSCOSSSSSCRNSSCSD OWA OSE «4030287 0 6800 5849612987 L6H 0.798 ORE 1522 0296, 2 3300013890936 =——2944 20121039 OSH9 SIT 059. Bialeci ring Mew 50. ie 12100866 «2916 «18950798 Tm 0302 3 18200 1550067 2.753 88S ORT 1412 0390 25 48533-2100 0.956 «25211856682 ORS1_—1.461— «0331 Telerene Plastic 25 37007 —«1900 =a —=«9TS «2S? OSIH ORDO Hackene Puasic 45 1200095 ONE—=«2RD—«18KG~~«CBES (080 ‘Rafa cing Plasic 15 asa 3009084 «2825-2400 OMB] «0585 —«1913 370 ‘Ben sadae Ceamic 25 80080 (m0 «O8 0620 12460387 B esos S450 0.650 0833 13640232 DINPAK Plasie 70 9763 —«MN07-—=«0838~STO«CL8ID 0981 OaTR—saT One a7 2s L293 29299817354 1690 0384 ‘Trans IChemE, Vol 77, Part A, September 1999502 BILLET and SCHULTES Table 26. Characteristic data and constants fr regular pockings te 7. Dumped Pckngs nm tm? mmm Paling Cenmis 30 oe ) ile rng Meal 35_—~—«2OG~—«ems aS 06900460 1405037 Rao Meal_-¥C250 Ee aio 1338 ons elo Meat_250" 25000970 ~S1S7 ast —05H 0am Gompack Meal AzESO8 20 0s77 «298620 osm _—o3ee inpue ping ‘Neal 250 i009 26101996 Oa 0RR O98 Oa Ceri 108 ot Oss sot e859) Oat al O32 Mowzpocing ‘Moat Di-200 20000079 S16 2339 ose? 035s aml 0x60 32300 50000930 «5088 26k aE O95 LHS oan sie C1200 mn00 ose 04 Lo oa12 0 000500 aes tg aa 0039 Biota Posies RHO 1100088607 197s OST) 0250 OST ONT in We righ Bane, dBm, TA Va When dumped packings or arranged packings are Custer, 6 abe, V1, 5 0 inigated, the ftee cross-section forthe gas flow is reduced = by the column holdup, and the surface structure is changed Bes 713 25 the result of the coating with the liquid film, Thus the ze Exped de column holdup hy, according to equation (11) must be taken g@? lo into account by equation (28) and the resistance coefficient 2s | of the two-phase flow ¥, defined by equation (29). It is, S22 f—} dependent on the Froude number of the flow of liquid and on gE: Made coteulted curve ze — os SCT TY : Vopour capacity foctor Fylm™® s kg!) ‘S0-ame Willow rings, plastic, d;+0.3 mm, 2= 100 Wm? a corn dee ai, 235, 5 wah, ed 0 : Bes © H tagcat aio ‘ kz ga. 7 ei ett cates cre Ba oly ww aw wu we wo Ges cnc fcr Fin 5° Soe Wow Age A GT om TOT sama nt eon ne Fe: 3 Eagan ae ae? Biz eecoeets ae, aor uw m2 2B a a Wo fos ein fcr Fy 5g!) igure 3. Comparison of the load-fficiency relationship caealated from the model conelstions with hat determined in reeteation, absorption and constant 2 diameter, m Aifusion cooticien, Fas or vapour capacity factor, Pa gravitational seceleration, ms Height, m {HITU height ofa transfer unt, m i, colum holdup, am K wall factor Ky overall mass transfer coefficient 1” ass flow of the liquid, gh" 1 molar flow ofthe gud, kath! M molecular weigh, zkmol-* yg slope ofthe equilibsum curve, kmolkol"! packing density 1, "> NTU tomber of tansfer unite exponent ‘4p pressure dop, Pa te) Velocity with ference to the free column ros-secton, mm? mean effective velocity, ms“! Vass flow ofthe gas or vapour, kgh-* v ‘molar ow ofthe gas or vapour, noth ‘molar concentation inthe iid, kl"! Greck eters ‘mass transfer coetReent, ms! © Void faton, mm dynamic viscosity, kems A Stripping Factor kinematic viscosity, m="! density, kgm? surface tenon, kgs"? resistance cosfcient Indices FL flooding point hydrate Liguid total particles Ph terface 5 loading pone S column V gas or vapour Wier REFERENCES Billet, R., 1983, Festechrift commemeratve publication) (Fakult fr Maschincnbay, Ruty-Universiat Bochum), p.24—31 Bilt, R, 1987, 1ChemE Symp Ser No, 108, . A171-A182. Billet, Rand Schulte M, 1987, IChemE Symp Ser No 10, p.B255— 266, 4. Billec R. and Sehults, M., 1991, In Beige eur Verfarens. und Unvelechnit, (Rube-Universtt Bochum, p. 88-107 and 108-122 5, Bille, R- and Schulte, M, 1991, Chere Bg Technol, 1: 89-95, 6, Bille, Rand Schulte, M, 1992, Chom Symp Ser No, 128, p, BI29-— B39, 7. Schltes, M., 1990, Doctoral dsseraton, (Rahe-Universtt Bochum, Fortctut-Berichte VDI Series 3 No, 230, VDLVetla, Disseldor, Billet, R, 1992, Chemlng-Technol, 64 401-~ 9. Bill Rand Scholtes, M, 1983, Chem Eng Technol 16 1-8. 10. Billet, Rand Schultes, M1993, Chem fing Technol, 16: 370-315. HL Billet, Rand Schattes, M1995, Chem Eng Technol 18: 371-39, I Billet, R., 1995, Packed Towers in Processing and Environmental Techvology, (VCH). ADDRESS Correspondence conceming this poper should be addressed to Dr-Ing. M. Scholtes, Raschig GmbH, Mandenbeimersrsse 100, D.67061 Ludw shafen, Germany. The manuscript was received 11 February 1999 and accepted for publication after revision 10 June 1999. ‘Trans IChemE, Vol 77, Part A, September 1999