MASS TRANSFER IN COCURRENT GAS-LIQUID FLOW IN A HORIZONTAL TUBE WITH AND WITHOUT TWISTED TAPE INSERTS

Abstract

The two-phase gas-liquid flow in pipes has become increasingly important to engineers in recent years because this type of flow is encountered in large number of situations and a clear understanding of transport rates is necessary for logical and complete design and operations of a very wide variety of engineering equipments and processes. Hydrodynamic and heat transfer studies in tv/o-phase system have been carried out to some extent earlier but the available information about the twophase mass transfer rates under different flow regimes is extremely limited and only during the last two decades some efforts have been made to systematically study the two-phase mass transfer rates in different types of contacting equipments. Literature survey revealed that only very few mass transfer studies have been made in horizontal cocurrent gasliquid flow in the slug flow regime and there has been no reported information about the effect of the presence of twisted tape inserts on mass transfer rates. Therefore, the present investigations were undertaken to understand and analyze the hydrodynamic and xxm mass transfer phenomena in horizontal two-phase gas-liquid flow in slug flow regime with and without twisted tape inserts. Pressure drop and mass transfer studies with and without chemical reaction are carried out for horizontal gas-liquid flow in the slug flow regime. The systems studied are carbondioxide - humidified air mixture as the gas phase and water as the liquid phase for physical absorption and aqueous sodiumhydroxide solutions for chemical absorption studies. The velocity ranges investigated are : 0.13 to 0. 73 m/sec for liquid and 0. 60 to 10.0 m/sec for the gas. Test section consisted of a smooth 13.5 mm internal diameter and about 4.0 metre long perspex tube. Twisted tapes were made from smooth 0. 75 mm thick, 12.50 mm wide and 2.50 metre long stainless steel strips. Experiments were carried out using empty tube and tube with twisted tapes throughout the test section length. The pitch (360 degrees) to diameter ratios of 5.00 and 9.32 v/ere used. The generation of swirl flow in pipe-lines by the use of internal twisted tapes results in the formation of complex flow patterns when used with two-phase gas-liquid flow. The main effect of this "swirling motion " is its ability to increase the dispersion of gas into liquid but at high swirl intensity there is a tendency for the liquid to be thrown to the wall of the horizontal tube and the flow of gas is largely restricted in the central core. Thus, the swirling flow has a significant effect on the interfacial area and liquid phase mass transfer coefficient. XXIV Samples v/ere withdrawn at three points for analysis but some experimental runs could not be used for calculation of required parameters because the concentration of carbondioxide in the liquid phase tended to approach the saturation value at a short distance from the inlet in the case of physical absorption in tube with lower pitch ratio twisted tape insert when high gas flow rates and low liquid flow rates were used. In the case of chemical absorption under similar conditions, almost 100 percent absorption of carbondioxide gas from the gas mixture was observed, The pressure drop per unit length of the test section increased to a high magnitude when twisted tape insert of lower pitch ratio was used. Increase in gas and liquid flow rates also resulted in higher pressure drop. This increase in pressure losses due to the presence of twisted tapes reduces the effective ness of using the twisted tape inserts to certain extent. The pressure drop values were always calculated by using the following modified friction factor equation developed from Smithberg-Landis correlation for Reynolds number range of 2100 to 106 : r , -1.07"i _n f • 0.0014 + L0.125 + 2.51 (H/D - 0.5) JRe r -0.5"\ n = 0.32 L 1 + 0.65 (H/D) J H/D = pitch to diameter ratio for 360 degrees twist Re = Reynolds number z: Djj . V . p / u D« = hydraulic diameter XXV The pressure drop in two-phase gas-liquid flow was calculated using Hughmark method of two-phase Reynolds number and Lockhart-Martinelli method. The values of pressure drop predicted by Hughmark method were always lower than the pressure drop measured experimentally with an average under-prediction of 37.7 percent. The Lockhart-Martinelli method of calculating the two-phase pressure drop was found better since this method under-predicted the pressure drop values for only 80 percent data points while the predicted values for the remaining 20 percent data points were more than the experimental pressure drop. The average under-prediction from Lockhart- Martinelli method is only 11.3 percent. The pressure drop values calculated by Lockhart-Martinelli method can be further improved using the following relationship : 1 018 ( AP/A.L) " 1.134 ( Ap/AL)t' ; standard error = 17.9% where ( AP/AL) = predicted two-phase pressure drop, kN/m3. ( A,P/AL) = two-phase pressure drop values LM estimated by Lockhart-Martinelli method, kN/m3. , o The volumetric liquid mass transfer coefficient, kL a , was observed to increase with an increase in superficial fluid velocities or decrease in the pitch ratio. It was correlated in terms of pressure drop, two-phase Reynolds number and Jepsen's energy dissipation parameter. Jepsen's energy dissipation parameter, £• , was found to be the most suitable J XXV1 for the correlation. The volumetric mass transfer coefficients correlated most satisfactorily by : Empty tube : k°a = 0.0331 6 ' 5| €1^4.0, standard error = 15*8% >-i J J k °a « 0.0234 6t'765 '» €r &4-° >^^dard error - 38.4% Tube with twisted tape inserts ; 0 97 7 kT a = 0.0853 6 ' ; € «$8.0, standard error = 23.6% *-* J J kT°a * 0.0384 £.0'660. ^ > 80, standard error =22.7% L J CJ " where kT ° a is in Sec"* and <6j is in kN/m. sec. From practical point of view it is more desirable to know the effect of twisted tape on volumetric mass transfer modulus, Mk ° , that is, k ° a / (A P/ A L). At lower gas velocities L " the twisted tape with larger pitch ratio ( H/D = 9. 32 ) gave the maximum value of Mk^ °a at superficial liquid velocity upto 0. 8 m/sec. But at higher gas velocity the maximum value of M, o • was found to be limited to a much lower value of liquid kL a velocity. The effective interfacial area, a , was observed to increase with an increase in superficial gas velocity or decrease in the pitch ratio. Interfacial area increased initially with an increasein superficial liquid velocity upto a value of 0.4 m/sec. but further increase in superficial liquid velocity, in general, decreases the interfacial area. The effect of difference in pitch ratio ( H/D = 5.0, and 9. 32 ) was found to be marginal but the observed values XXV11 of interfacial area v/ere 1,5 to 4.0 times the values obtained in empty tube. The effective interfacial area is correlated by Banerjee energy dissipation parameter, two-phase Reynolds number and Jepsen's energy disspotion parameter. Area creation rate, first defined by Kasturi and Stepanek, was also correlated by prescure drop. The correlations based on Jepsen's energy dissipation parameter, CT , and area creation rate, ( a Qt) / E-r , were found to be most suitable. The effective interfacial area correlations with Jepsen's energy dissipation parameter are : Empty tube : C. 313 a = 163.9 €j' ; standard error = 31.0% Tube with twisted tape inserts a = 253.9 £°«231 . standard error • 31.2% »j 2 3 where a is in m'"' / m . Kasturi and Stepanek's area creation rate correlations with pressure drop are given as : Empty tube : ( a QL ) / E = 0.0224 (Ap/AL) ' ' ; standard error • 32.3% Tube with twisted tape inserts (aQL )/ EL =0.00778 (AP/ AL) '° standard error = 32.8 % where a Qr / Et is rate of area creation in m /sec. XXV111 From practical consideration it may be more desirable to estimate the area modulus, a/ (Ap/Al), or area creation rate modulus , ( a QL / EL ) / (Ap/AL), Area modulus was found dependent on superficial gas and liquid velocities, but the value of area creation rate modulus can be found quite easily by : Empty tube : ( a Q / ET ) 0.034 h k - 0.0224 ( A P/^L) Si 0.0224 ( AP/ A L) Tube with twisted tape inserts ( aQr / ET ) 0.064 - ^ = 0.00778 ( A P/ A L) CrT 0.00778 ( A P/ A L) The liquid phase mass transfer coefficient, k , was calculated from the volumetric liquid phase mass transfer coefficient, kT°a, and the value of interfacial area, a, at the similar hydrodynamic condition, that is, at the same value of two-phase Reynolds number. k,° was found to increase with an increase in superficial gas velocity and decrease in the pitch ratio. o k was correlated using the correlating parameters L proposed by Jepsen, Banerjeo, Lament and Scott and Kasturi and Stepanek. The Jepsen energy dissipation parameter and the method of Kasturi and Stepanek were found most satisfactory. These correlations are given as below : XXIX Empty tube and tube with twisted tape inserts kL° = 1.500 xlO"4 £0-375 . standard error = 28.8% where k, is in m/sec. Empty tube : _ 4 ShL = 0,1243 PcL EuLSc^ 2 ; standard error =30.5% Tube with twisted tape inserts i ShL = 0.07863 PeL EuLSC]^ 2 ; standard error - 31.2% where ShL , Pe , EuL and Sc^ are liquid phase Sherwood, Pedet, Euler and Schmidt numbers respectively. It may be noted that in most of the cases the numerical values of the constants in all correlationr for tube with twisted tape inserts were found to be independent of the pitch-to-diameter ratio of the twisted tapes investigated. However, these values differed considerably when compared with those for empty tube. Following values selected from the range of parameters studied in the present investigation, show the effectiveness of the use of twisted tape inserts on mass transfer coefficients and interfacial area : XXX Effectiveness Factor £ , k-N/mi1 sec. J 0.1 0.4 1 4 10 40 100 (kT °a) u twisted tape 4.45 3.29 2.42 1.68 1.33 1.19 1.02 (k °a) J-1 empty tube (a) tv/isted tape 1.88 1.67 1.53 1.40 1.28 1.12 1.07 (a) x empty tube (k °) v L 'twisted tape 1. 72 1.43 1.27 1.09 0.88 0.89 - 0.90 O + *-• empty tube The value of effectiveness factor greater than unity proves the utility of tv/isted tape inserts, and the above values clearly show that the use of twisted tape inserts is attractive to increase the mass transfer rates and interfacial area only upto a limited increase in liquid and gas velocities. The results of the present investigation clearly show that the mass transfer rates can be increased by using the twisted tape inserts and this increase is maximum, at low gas and liquid velocities. The effectiveness of twisted tape inserts is maximum at low values of increases to a sufficiently high value the use of tv/isted tape inserts may even retard the mass transfer rates when compared with those for empty Jepsen's energy dissipation parameter, £ , and as £ T XXXI tube. Thus, absorber volume can be reduced considerably by using twisted tape inserts but at a considerably high pressure drop. When pressure loss considerations are important it is recommended to use twisted tapes with very large pitch to diameter ratio especially at lower liquid and gas velocities. It is hoped that the correlations developed in the present investigations will be useful in the design of tubular absorbers and reactors.