I1.1_ E, The Age Distribution of Fluid, The RTD 267 | Brame 07, FINDING THE RTD BY EXPERIMENT ‘The concentration readings in Table E11.1 represent a continuous response to a pulse input into a closed vessel which is to be used as a chemical reactor. Calculate the mean residence time of fluid in the vessel f, and tabulate and plot the exit age distribution E. Table E111 ‘Tracer Output Concentration, Cre Time ¢, min gmi/liter fluid 10 4 25 30 35 OrNaUUBS SOLUTION: ‘The mean residence time, from Eq. 4, is DACAt, de= constant > 4, DAK =G _5X34+10X5+15X5+20X4+25X2+30X ~ 3+54+54+44+24+1 t= =15 min The area under the concentration-time curve, Area = >) CAr=(3+5+4+54+4+2 + 1)5 = 100 gm-min/liter gives the total amount of tracer introduced. To find E, the area under this curve must be unity; hence, the concentration readings must each be divided by the total area, giving é E=-—— area ‘Thus we have 0 5 0 1 2 2 30 0 0.03 0.05 0.05 0.04 0.02 0.01268 © Chapter 11 Basics of Non-Ideal Flow smoctned oes cine os < E 0.03 . ove oot L 20 Figure E11 Figure E11.1 is a plot of this distribution. [EXAWPEER rivowe rue curve FoR LIQUID FLOWING THROUGH A VESSEL A large tank (860 liters) is used as a gas-liquid contactor. Gas bubbles up through the vessel and out the top, liquid flows in at one part and out the other at 5 liters/s, To get an idea of the flow pattern of liquid in this tank a pulse of tracer (M = 150 gm) is injected at the liquid inlet and measured at the outlet, as shown in Fig. E11.2a. (a) Is this a properly done experiment? (b) If so, find the liquid fraction in the vessel. (©) Determine the E curve for the liquid. (@ Qualitatively what do you think is happening in the vessel? ae few, hy =0.375 z 5 . ea = ee os 2 ¢ ¢ 3 ae Figure Ell.2a SOLUTION: (a) Check the material balance against the tracer curve. From the material balance, Eq. 3, we should have =M _ 150gm _ ,ogm-s _ 9.5gm-min acca Siiters/s ~ °° titer ~ °° titer11.1 _E, The Age Distribution of Fluid, The RTD 269 From the tracer curve aree=A(14¢te dee +) = 0375 (8) = = 0.5 e-min 16 liter These values agree. The results are consistent. (@) (b) For the liquid, Eq. 4 gives _. Liew dt 4 Ay [ea +4xBroxtraxde. ++ | =2.67 min fear Cadt 05 64 Thus the liquid volume in the vessel is V, = iv, = 2.67(5 X 60) = 800 liters and the volume fraction of phases is Fraction of liquid = 522 = 939% sa } ) Fraction of gas = 7% (© Finally, from Eq. 5 we find the E curve, or Coase _ 0.75 = Spite = eS ae Thus the E curve for the liquid is as shown in Fig. E11.2b. © (d) The vessel has a strong recirculation of liquid, probably induced by the @270 Chapter 11 Basics of Non-Ideal Flow The Convolution Integral ‘Suppose we introduce into a vessel a one-shot tracer signal C,, versus ¢ as shown in Fig. 11.15. In passing through the vessel the signal will be modified to give an output signal C,,, versus t. Since the flow with its particular RTD is responsible for this modification let us relate C,, E, and Co. Focus attention on tracer leaving at time about . This is shown as the narrow rectangle B in Fig. 11.15. Welmay) write __ (all the tracer entering t’ seconds earlier than ¢, tracer leaving) _ and staying for time ¢' in the vessel in rectangle B, We show the tracer which enters ’ seconds earlier than ¢ as the narrow rectangle An terms ofthis rectangle te/abovel equation may! be Written ica tracer in \ / fraction of tracer in A (tercerteaving) 3 _{rectanle {which stays or about bon A t' seconds in the vessel, ae In symbols and taking limits (shrinking the rectangles) we obtain the desired relationship, which lee voluson integral Coalt) =f Colt ~ # VEC at! (10a) In what can be shown to be equivalent form we also have Coat) = fi Cale VEC ~ eae 0) We say| that... 18 {i€] CoAUOIMOR| OF EIWith| E, and we write concisely or Cy = Cg E (10e) seconds ier Figure 11.15 Sketch showing derivation of the|@SAWOIIHOH) itera.I1.1 _E, The Age Distribution of Fluid, The RTD 271. Region a Region b Figure 11.16 MOdificatiOn|Of &alinput|EFAcer signal C,, on passing through three succes- sive regions. To illustrate the uses of these mathematical tools ich are closed and connected in series (see Fig. 11.16). Problem 1, If the input signal C,, is measured and the exit age distribution functions E,, E,, and E, are known, then C, is the convolution of E, with Ci, and so on, thus G=Gek. G=C#E, Cy=G*k, and on G6mbining, Coat = Cin * E, +E, *E, ay ‘Thus we can determine the output from a multiregion flow unit. Problem 2. Ifwe measure C,, and C,,,and know B, and E, we can then extract the unknown E,.. This type of problem is of particular importance in experimentation where the entrance region and collection region for tracer are both large com- pared with the experimental section. It is a straightforward matter to convolute; however, to deconvolute, to find ‘one of the distribution functions under the integral, is difficult. Thus Problem 2 is harder to treat than Problem 1 and requires using a computer. * By independence we mean that the fluid loses its memory as it passes from vessel to vessel. Thus, ‘a faster-moving fluid element in one vessel does not remember this fact in the next vessel and doesn’t preferentially flow faster (or slower) there. Laminar flow often does not satisfy this requirement of independence; however, complete (or lateral) mixing of fuid between units satisfies this condition,272 Chapter 11 Basics of Non-Ideal Flow In some cases, however, we can in essence deconvolute. This special situation is considered at the end of Chapter 14 and is illustrated with an example there. Example 11.3 illustrates convolution; Example 14.4 illustrates deconvolution. (lEanPETES, convotution Let us illustrate the use of the convolution equation, Eq. 10, with a very simple example in which we want to find C,,, given C,, and the E curve for the vessel, as shown in Fig, E11.3a. Tracer input curve RTD for the vessel E curve 04 E Area=1 02 Came See eee ea ' Figure E113a SOLUTION First of all, take 1 min time slices. The given data are then i Ge ' o£ 0 0 5 0 Note: The area under the 1 0 6 00s E curve is unity. 2 8 7 0.50 3 4 8 035 A fs 9 0.10 50 10 0 Now the first bit of tracer leaves at 8 min, the last bit at 13 min. Thus, applying the convolution integral, in discrete form, we have t Cua 7 0 0 8 8 x 0.05 = 04 9 8x05 + 4x 0.05 42 10 8X 035 +4X 05 +6 x 005 = 5.1 see 8X 0.10 +4035 +6 05 = 52 2 4X 0.10 + 6 X 0.35 25 13 6X 0.10 06 14 011.2 Conversion in Non-Ideal Flow Reactors 273 The C,,, E, and C,y curves, in discrete and in continuous form, are shown in Fig. E11.3b. Note that the area under the C,,, curve equals the area under the Cj, curve. Ecune Calculated 11.2, CONVERSION IN NON-IDEAL FLOW REACTORS To evaluate reactor behavior in general] We) Have) {6) kAOW four factors.) the kinetics of the reaction the RTD of fluid in the reactor the earliness or lateness of fluid mixing in the reactor whether the fluid is a micro or macro fluid nee in plug or mixed flow we have developed the equations in the For intermediate flow we will develop appropriate models in Chapters 12, 13, and 14. To consider the early and late mixing of a microfluid, consider the two flow patterns shown in Fig. 11.17 (ASRS In (a) the reactant starts at high concentration and reacts away rapidly because n> 1.In (6) the fluid drops immediately to a low concentration. Since the rate of reaction drops more rapidly than does the concentration you will end up with a lower conversion. microfluids Late mixing favors reactions where n > 1 Early mixing favors reactions where n <1 a2) 5 POA magne US| HUAI HG staying for different lengths of tuinetn the reactor Cpivea by the E functioa)| Sane274 Chapter 11 Basics of Non-Ideal Flow —a-(&)- —- + - © © Figure 11.17 This shows the/[a{Stand (HE) ESHESt MINING we can HAVEHORS given RTD. So the mean composition in the exit stream will have to account for these two factors, the kinetics and the RTD.{Injwords) then concentration of 5 5 ‘mean concentration reactant remaining Sastion of ent of reactant inanelementof || ore he in exit stream aidemens | age between ||" 48° Petweens attic | a8 between and t+ dt In symbols this becomes C, = (C, (& -f (& “Edt wT 40" Such ottaeraget or in terms of conversions ¥ a= [em Ed on or in a form suitable for numerical integration Ct, CG ay (2) “Et Cao sage \Ca0/ coment ‘ners From Chapter 3/G8IBtGH|FESGIGH we have , Ca * forfirst-order reactions (—* (a4) "aa/ element 5 Cy 3 for second-order reactions ( es) “TEs (as) Cc, © for an nth-order reaction (2) = [1+ (n-1)Cxp ee (16) ‘a0 element