Parameters characterising the pulsing flow for cocurrent flow of gas and foaming liquid in a pressurised tricklebed

Transcription

1 Parameters characterising the ulsing flow for cocurrent flow of gas and foaming liquid in a ressurised tricklebed reactor Proceedings of Euroean Congress of Chemical Engineering (ECCE-6) Coenhagen, 16- Setember 7 Parameters characterising the ulsing flow for cocurrent flow of gas and foaming liquid in a ressurised trickle-bed reactor D. Janecki a, G. Bartelmus b, A. Szczotka b a Deartment of Process Engineering, University of Oole, Dmowskiego 7/9, Oole, Poland b Polish Academy of Sciences, Institute of Chemical Engineering, Baltycka 5, 44-1 Gliwice, Poland Abstract The aim of the resent work was to exerimentally determine the arameters characterizing the ulse flow of liquids through bed i.e. the frequency of ulsation and the velocity of ulses travelling along the bed and next to work out the correlation equations enabling their assessment. The research was carried out in a ressure reactor (ressures u to MPa) having.5 m in diameter and being 1.5 m long. Physicochemical roerties of the gas and liquid hases were changed in the research and the concentrations of water solutions of alihatic alcohols were chosen in such a way so that the systems would form foams of various lifetime as a result of their contact with the gas in acking. The exerimental data which were obtained as a result of the research were correlated deending on the oerational arameters of the reactor and hysicochemical roerties of gas and liquid hases. The knowledge of the examined in the resent work arameters is extremely imortant to model the oeration of a three-hase system in the ulsing flow regime. Keywords: trickle bed reactors, foaming systems, ulsing flow 1. Introduction Trickle-bed reactors are commonly used in the branches of industry where great streams of substrates are rocessed (among others in etrochemical and kerosene industries) (Al-Dahhan et al. (1997), Dudukovic et al. ()). Physicochemical roerties of gases and liquids, which are either substrates or roducts in the carried out in TBR rocesses, differ from air and water roerties, that is systems which were alied by many authors in their research (Saroha and Nigam (1996), Duduković et al. (1999, )). Besides, a lot of them (alcohols, solvents whose bases are hydrocarbons, kerosene, diesel fuels) form foams as a result of contact with gas in acking. The occurrence of a foam in the system is not a favourable henomenon as it

2 D. Janecki et al. increases gas flow resistance through acking and decreases the amount of liquid resent in bed. Thus, it changes the arameters influencing safety and economics of the rocess unfavourably. Nevertheless, it is a commonly occurring henomenon in a lot of rocesses. Thus, it seems advisable to carry out exerimental research into the hydrodynamics of the systems. The aim of the resent work was to exerimentally determine the arameters characterizing the ulse flow of liquids through bed i.e. the frequency of ulsation and the velocity of ulses travelling along the bed and next to work out the correlation equations enabling their assessment.. Exerimental set-u and measuring system The research was carried out in a ressure reactor (ressures u to MPa) having.5 m in diameter and being 1.5 m long (Fig. 1). PI PI a PI Fig. 1. Exerimental set u: 1 air comressor, filters, 3 air dryer, 4 comressed air tank, 5 flowmeters, 6 acked column, 7 conductivity cells, 8 liquid ums, 9 microbubble searator, 1 liquid tank, 11 float valve, 1 demister, 13 gas liquid searator. Physicochemical roerties of the gas and liquid hases were changed in the research and the concentrations of water solutions of alihatic alcohols (Table 1) were chosen in such a way so that the systems would form foams of various lifetime as a result of their contact with the gas in acking (WF weakly foaming, SF strongly foaming). The method of measuring the changes in conductivity of the two-hase gas-liquid mixture flowing through bed was used to determine the arameters characterizing the ulsing flow of fluids.

3 Parameters characterising the ulsing flow for cocurrent flow of gas and foaming liquid in a ressurised Gas hase Liquid hase ρ µ. 1 5 Medium ρ µ Medium 1 3 σ. L 1 3 [kg m -3 ] [Pa s] (mass ercent) [kg m -3 ] [Pa. s] [N m -1 ] System air methanol ( 53%) WF methanol ( 4%) SF ethanol ( 5%) SF ethanol ( 4%) SF isoroanol ( 56%) WF isoroanol ( 34%) SF Table 1. Physicochemical roerties of the tested system (t=ºc), (P r =.1 MPa) It was stated that, for the tested systems, ulse flow regime is obtained at air ressure in the reactor not exceeding MPa for methanol (WF, SF) and ethanol (SF) solutions, but for 56% isoroanol (WF) only u to.6 MPa. The region of ulsing flow regime occurrence for ressures higher than atmosheric was not found for 34% isoroanol solution (Janecki et al. (5)). 3. Parameters characterizing ulse fluid flow To model ulse flow is necessary to know the arameters charactering it which are the ulsing flow of fluids, namely, the ulse velocity and the frequency of ulsation (Rao and Drinkenburg (1985), Dimenstein and Ng (1986), Dankworth et al. (199), Bartelmus (1993), Belhouwer i ws. ()). Only a few works whose authors estimated arameters characterizing ulse fluid flow can be found in literature. In their works the research was carried out in reactors oerating at ressure close to atmosheric and the tested systems were, first of all, water nitrogen or water air (Weekman and Myers (1964), Sato et al. (1973), Blok and Drinkenburg (198), Rao and Drinkenburg (1983), Tsochatzidis and Karabelas (1995)). It was only in the works of Bartelmus et al. (1998), Burghardt et al. (1999, 3a,b), Szlem et al. (1) and Tsochatzidis et al. (1998) that an attemt was undertaken at quantification of the deendencies of the measured arameters (that is f and V ) on the reactor oerational arameters and hysicochemical roerties of the liquid sraying the bed (glycerol solutions) Velocity of the ulses The methods enabling tracing the changes in the two-hase structure of the gas-liquid mixture flowing through bed in ortions rich in liquid and rich in gas are used to measure the velocity at which ulses move along the bed. Therefore, they are otical methods (Weekman and Myers (1964), Sato et al. (1973)), acoustic methods (Kolb et al. (199), Christensen et al. (1986)) or conductometric methods (Blok et al. (198, 1983), Rao and Drinkenburg (1983), Bartelmus et al. (1998), Burghardt et al. (1999)). In the resent work the velocity of ulses was calculated using the signal roduced by the two conductometric cells laced in the bed at distances of 1. and 1.1m from the to. 3

4 D. Janecki et al. The signals from the cells, after amlifying, were fed into the comuter memory and collected in 9-second series. Samling frequency was fixed at the level of 1 Hz. The registered in the comuter memory signals were standardized to the form of zero mean value and unitary variance. Having standardized data sequences from two conductometric cells at the distance of l=.1 from each other a grah of a crosscorrelation function of both signals was lotted. The correlogram showed a distinct maximum (Fig. ) and the time corresonding to this maximum (termed the signal delay time, τ max ) is the time interval required by a ulse to cover the distance between the uer and lower conductometric cells. Fig.. Grah of the function of signal intercorrelation of signals from two conductometric cells. The exerimentally obtained V values deending on the real velocities of both hases are shown as an examle in Fig. 3. It was stated that ulse velocity was lower than gas real velocity in case of every tested system. It means that gas always enetrates ulse in the systems with low value of surface tension. What follows from the comarison of V values for weakly and strongly foaming systems is that this feature of the system does not influence the velocity of the arising ulses. One has to remember, however, that strongly foaming systems obtain the ulse flow regime at much lower flow velocities of both hases. 4

5 Parameters characterising the ulsing flow for cocurrent flow of gas and foaming liquid in a ressurised V (m/s) v L (m/s) v g (m/s) Fig. 3. The deendence of ulse velocity on the real velocities of both hases; isoroanol 56%-air system; P=.6 MPa. The change in the density of gas hase (ressure) in the reactor did not influence the measured value significantly (Fig. 4) V (m/s) v L = m/s v g (m/s) P (MPa) Fig. 4. The influence of gas density on ulsation frequency; methanol 53%-air system. The data base obtained in the ressurized reactor and in the aaratus working at atmosheric ressure (Burghardt et al. (3a, 3b)) makes it ossible to elaborate an equation correlating the measured values of ulsation velocity. The following deendence was obtained: 5

6 D. Janecki et al. V v g = Re. 857 gr. 13 ρ ρ g airn. 46 Re. 358 Lr (1) which with the mean ercentage error of 1.5% describes the measured V values (R=.957) (Fig. 5) % 1. (V/vg)calc.8.4 -% (V /v g ) ex Fig. 5. The comarison of V values estimated exerimentally and calculated from the equation (1) 3.. Pulsation frequency Pulsation frequency was determined by analysing the signal from the uer conductometric cell laced in bed ~1.m below the to of bed. The rocedure described in detail in the work of Burghardt et al. (1999) was used to obtain the aim. It enables obtaining a set of ulse duration and the number of ulses with the given duration for every 9-second realization. Next, robability density function was calculated from the test and its chart (histogram) was drawn. A log-normal distribution was acceted as a statistical model as a result of the analysis of its shae. A mean ulse duration was determined from the first moment of this distribution, whose converse is the looked for value of ulsing frequency. While analysing the obtained in the resent work findings of the exeriments for systems forming foams it was stated that it is very difficult to define the influence of real velocities of both hases on the ulse frequency as this deendence is different for systems forming weakly and strongly foams (Fig. 6). 6

7 Parameters characterising the ulsing flow for cocurrent flow of gas and foaming liquid in a ressurised f (1/s) v g (m/s) v L (m/s) Fig. 6. The influence of real velocities of both hases on the ulsation frequency; ( ) ethanol 5%-air system, ( ) ethanol 4%-air system, ( ) isoroanol 56%-air system, P=.6 MPa. As well as this, the change in gas/ressure density in the reactor causes the change in the number of ulses formed in the column; the higher the gas density in the reactor, the higher the ulsation frequency (Fig. 7)..8 v g = m/s f (1/s) v L (m/s) P (MPa) Fig. 7. The influence of gas density on ulsation frequency; ethanol 4%-air system. 7

8 D. Janecki et al. The ψ flow arameter was introduced into the correlation equation to catch hold of influence of the hysicochemical roerties of the tested solutions on the measured f values. The following equations have been roosed: - for strongly foaming systems (34 exerimental oints): f = ψ Re gr We Lr () which with the mean ercentage error where e Y =33.4% (R=.936) aroximates the exerimental data. - for weakly foaming systems (387 exerimental oints): f = Re gr +, 6ψ Re gr We Lr (3) where e Y =8,3% (R=,84). The region of low ulsation frequency is the most interesting one for industrial uroses because only then, from the economic oint of view, the increase in resistance of gas flow through acking is comensated for by a great increase in mass transfer coefficients. In the initial hase of ulsation the frequency increases in a linear way together with the increase in the real velocity of the liquid hase. For a reactor oerating at an elevated ressure the majority of exerimental oints was located in the region of linear deendence of the ulsation frequency on v L. In addition, it was noticed that the directivity factor of these straight lines changes both with the change in gas hase velocity and ressure in the reactor. However, changes in the hysicochemical roerties of liquid are take into account in the real velocity of the liquid hase. In the considered region the following deendence was obtained for weakly foaming systems ( exerimental oints): f ( Re ) v. 41 = (4) gr L which with the mean relative error not exceeding 18% (R=,937) aroximates the exerimental data (Fig. 8). 4. Summary The results of the research whose aim was to determine arameters characterizing ulsing flow of fluids in a trickle-bed (gas-liquid-solid) reactor with a fixed bed were resented in the work. The research was carried out for a secific grou of measuring systems systems which formed foams in the ulse flow regime. The aim of the exeriments was to find the answer to the question how the change in hysicochemical roerties of fluids and the forming of foams in the system influence the value of the measured arameters. 8

9 Parameters characterising the ulsing flow for cocurrent flow of gas and foaming liquid in a ressurised 6 5 fcalc 4 3 +% 1 -% f ex Fig. 8. The comarison of f values exerimentally determined and calculated from equation (4) The change of the hydrodynamic regime of the aaratus to ulse flow regime and next calculating the roductivity and selectivity of the rocess requires the knowledge of arameters characterizing this hydrodynamic regime (the velocity of ulses and their frequency). The carried out in the resent work research and the correlations formulated on its basis make it ossible to calculate the values of these arameters and as for system forming foams they have been the only available deendencies so far. Notation d article diameter (m) N 1 Yex,i Ycal,i ey = 1% N Y i= 1 ex, i f frequency of ulsation (s -1 ) P ressure (Pa) R correlation coefficient R cross-correlation function x~, 1 x~ v α real velocity of α hase (ms -1 ) V ulse velocity (m) Greek letters µ viscosity (Pas) ρ density (kg m -3 ) τ time (s) σ surface tension (N m -1 ) mean ercentage error (%) 9

10 D. Janecki et al. Dimensionless numbers vα d ρα Re α r = Reynolds number µ We Lr v = L α ρ d σ L L σ µ ρ. 33 Weber number W L W ψ L = flow arameter σ L µ W ρ L Subscrits g gas L liquid max maximum N normal conditions ulse w water α α-hase, α = g, L References Al-Dahhan M.H., Larachi F., Duduković M.P., Laurent A., (1997) Industrial Engineering Chemistry Research, 36, Bartelmus G., (1993) Chemical Engineering and Processing, 3, Bartelmus G., Gancarczyk A., Stasiak M., (1998 Chemical Engineering and Processing, 37, Boelhouwer J.G., Pieers A.A., Drinkenburg A.A.H., () Chemical Engineering Science, 57, Blok J.R., Drinkenburg A.A.H., (198) Chemical Engineering Journal, 5, Blok J.R.. Varkevisser J., Drinkenburg A.A.H., (198) Chemical Engineering Journal, 5, Blok J.R., Varkevisser J., Drinkenburg A.A.H., (1983) Chemical Engineering Science, 38, Burghardt A., Bartelmus G., Gancarczyk A., (1999) Chemical Engineering and Processing, 38, Burghardt A., Bartelmus G., Janecki D., (3a) Chemical and Process Engineering, 4, Burghardt A., Bartelmus G., Janecki D., (3b) Chemical and Process Engineering, 4, Christensen G., McGovern S.J., Sundaresan S., (1986) AIChE Journal, 3, Dankworth D.C., Kevrekidis I.G., Sundaresan S., (199) AIChE Journal, 36, Dimenstein D. M., Ng K. M., (1986) Chemical Engineering Communications, 41,

11 Parameters characterising the ulsing flow for cocurrent flow of gas and foaming liquid in a ressurised Dudukovic M.P., Larachi F., Mills P.L., () Catalysis Review, 44, Dudukovic, M. P., Larachi F., Mills P. L., (1999) Chemical Engineering Science, 54, Janecki D., Bartelmus G., Krótki T., 7th World Congress on Chemical Engineering Proceedings, 1-14 July 5, Glasgow, Scotland. Kolb W.B., Melli T.R., de Santos M.J., Scirven L.E., (199) Industrial Engineering Chemistry Research, 9, Rao V.G., Drinkenburg A.A.H., (1983) Canadian Journal of Chemical Engineering, 61, Rao V.G., Drinkenburg A.A.H., (1985) AIChE Journal, 31, Sato Y., Hirose T., Takahashi F., Toda M., Hashiguchi Y., (1973) Journal Chemical Engineering of Jaan, 6, Sorosha A.K., Nigam K.D.P., (1996) Review Chemical Engineering 1, Nos Szlem A., Bartelmus G., Janecki D.; (1) Chemical Engineering Science, 56, Tsochatzidis N.A., Karabelas A.J., (1995) AIChE Journal, 41, Tsochatzidis N.A., Ntamegliotis K.J., Karabelas A.J., (1998) Chemical Engineering Communications, 166, Weekman V.W., Myers J.E., (1964) AIChE Journal, 1,