GB2181664A - Packed beds - Google Patents

Abstract

Packed beds having a main zone packed with particles having a size or characteristic such as to give the desired characteristics of the bed as a whole, and a second zone adjacent a wall of the reactor and packed with particles selected to have different size and/or characteristics to those of the main zone exhibit many advantages over conventional packed beds, especially in radial heat transfer, and especially in catalytic processes operating at high Reynolds number. The wall-channelling effect may be reduced by using smaller particles in the interstices between the larger particles of the main zone in the second zone adjacent the bed wall. Particles having differing shapes and/or loaded with differing amounts of active catalyst material may also be used.

Description

SPECIFICATION Packed beds / This invention concerns improved packed beds and their preparation and use, with particular reference to packed bed reactors, eg catalytic reactors.

Hereinafter, it will be generally described that the packed bed is located within a tubular reactor, but it should be understood that other packed bed devices, and other reactor shapes, eg on the shell side of shell-and tube reactors are included.

The use of packed bed reactors according to previously known methods, to perform solid-catalysed reactions, suffers from three main disadvantages.

Firstly, in the neighbourhood of the tube wall there exists a substantial resistance to the radial transfer of heat. For reactions with large heat effects, this resistance can cause difficulties associated with excessive bed temperatures (for exothermic reactions) orexcessivetube-skin temperatures (for endothermic reactions). Secondly, there exists a degree of axial dispersion in the bed which can lead to reduced conversion and, in the presence of competing reactions, to reduced selectivity. Thirdly, the average path length for radial heat transfer is sufficiently long that substantial differences in temperature are established between the tube wall and the neighbouring particles and between those particles and the particles nearerthe bed centre.

All these effects are attributable, in large part, to the existence nearthe wall of the bed of a zone of high void fraction. This zone supports a low rate of radial heattransferbecause ofthe reduced presence ofthe heat transfer mechanisms associated with the presence of solid particles, and it is this reduced rate which is observed as a resistance near the wall. The high void fraction in the zone nearthe wall also in vites flow to occur preferentially near the wail and it is this "channelling" or "by-passing" which contributes to the apparent axial dispersion observed in such beds. The high voidage in thezone meansthat the bed particles are, on average, at a high distance from the tube wall, and so the average path length for heat transfer is correspondingly long.

The present invention provides a packed bed reac torcomprising a pl ural ity of zones of packed bed material including a main zone comprising particles selected to have such dimensions and/orchar acteristicsto provide the required characteristics of the bed as a whole, and a second zone adjacent a wall ofthe reactor and arranged in a radial orsubstantially radial mannerand comprising particlesofsuch dimensions and/or characteristics as to (i) more fully fill said second zone than would the particles of said main zone and/or (ii) more obstructfluid flow through said second zone than would the particles of said main zone and/or (iii) further promote radial heat transfer in said second zone than would the particles of said main zone.The invention also provides the use of, and a method of producing, the reactor of the present invention.

The bed material is preferably a catalyst, and will generally be described hereinafter as such, butthe invention has application also to materials normally found in packed bed devices, including inert materials, absorbents, adsorbents, etc. In the case of a catalyst, it may be of any form, whether on a support or not, including mixed catalysts orcatalystsdiluted with inert material. The invention is preferably applied to beds of lowtube-to-particle diameter ratio; the ratios implied in the Examples below (5.6:1 and 190.1 ) fall within the range of interest.

According to one embodiment of the invention, both main and second zones comprise particles of substantiaily similarshape and size, and the inten- tions of the invention are achieved by arranging, interstitially, smaller particles in the second zone. Because the resistance to radial heat transfer near the wall decreases with a decrease in particle size near the wall, heattransferto (orfrom) the wall will occur at a higher rate than would be the case with a bed of particles of substantially uniform size and ofthe sameshape asthe main zone particles, atthesame mass flux, bed void fraction and tube diameter.

Moreover, the bed pressure drop would be lower, at the same mass flux, than in a bed (ofthe same diameter) packed with particles of substantially uniform size, and the same shape, such that the same heat transfer rates are obtained as in the bed according to the invention.

Further, the particles in the bed according to the invention are, on average, closerto the wall than would be the case in a conventional bed ofthesame void fraction, and diameter, comprising substantially uniformly sized particles of the same shape. The pre senceofsmallerparticles increasestheflowresist- ance in the second zone. This may be expected tore- sult in a more even distribution of flow within the bed and thus reduce axial dispersion. If the heat transfer gains are substantial, then itwouid suffice merely that the axial dispersion is not substantially greater than in a bed of substantially uniformly sized particles.

It may be beneficial to use particles having different catalyst loadings on a support, varying with particle size. Typically, it would be advantageous for the smaller particles to carry the greater loading since they are less vulnerable to reduced conversion or selectivity due to transport resistances.

In a further embodiment, the interstitial volume may be reduced and a greaterfluid obstruction achieved in the second zone by using particles of dif ferent size and/or shape from those used in the main zone. For example, if spherical particles were used, the diameters ofthose in the second zone may be less than the diameters of those in the main zone.

According to another aspect ofthe invention, the zones of particles in the bed may be arranged such that the extent of interstitial packing, or the transition from particles of one characteristic to another, may show gradation in a continuous orsubstantially continuous radial manner across the bed so that there are no definite zone boundaries.

The present invention is to be distinguished from two prior art practices, namely (i) incorporating into the bed voids a powder which isfluidised by the upflow of fluid through the bed, enhancing radial heat transfer; (ii)fillingtheinnermostzoneofthebedwith inert particles while filling the outermost zone with very small particles held in place by a screen. The first introduces difficulties associated with moving particles, not arising in the present invention, while in the second, the outermost zone does not support bulkflow; rather,transport is by molecular mechanisms.

The invention is expected not only to improve performance of existing reactors, but also to permit the design of new reactors with tube diameters or cata lust types which are currently precluded by conventional bed packing practices. Typically, it will allowthe use of larger diameter tubes, thus requiring fewer tubes for any given total reaction mixture flowrate, and often requiring a shorter packed length for given fractional conversion. These changes offer reduced reactor capital costs. Moreover, the invention permits the use of higherflowrates through existing reactor tubes. The invention has particular application at high Reynolds numbers forfluid flow through the reactor, and it is preferred that the fluid is a gas orvapour.

In the packing of the beds according to the invention, it may be desirableto use known methods of vibrating,tapping ortamping, orto pre-fill the reac torwith a gas,vapourorliquid.Asuitable method of packing the bed isto position within the empty reactortube anothertube, called a "loading tube", coaxially. For instance, if an outer zone of the bed is to contain large spheres, with smaller spheres located interstitially, then the outer diameter of the loading tube may be chosen such thatthe annular gap between it and the reactorwall is slightly larger than the diameter ofthe larger sphere. A mixture of large and small spheres is charged into the annular gap, and a suitable packing, eg the larger spheres, is charged into the loading tube.The loading tube is then withdrawn until its base isjustsubmerged in the bed so made, and the charging and withdrawing process is repeated until a bed of suitable depth is completed.

To avoid excessive re-structuring ofthe bed on withdrawal of the loading tube, the wall thickness of that tube should be as small as convenieni' y possible. If the reactor tube is so long that there is some difficulty in locating the loading tube co-axially, the loading tube may be provided with radially projecting arms or perforated discs at suitable axial positions, to assist location.

In orderto avoid restructuring of the bed upon withdrawal of the loading tube, an alternative is that the loading tube is destroyed in place, eg by solution, evaporation, combustion, or some other method.

Deleterious re-structuring ofthe bed consists of particles from the wall zone, such as the smaller spheres, moving into the main zone, wherethey add to the bed pressure drop without enhancing the radial heattransfer. A design of loading tube which assists in holding the wall zone packing in position as the loading tube is withdrawn, while allowing the main zone packing to move outwardstowardsthe wall, is a loading tube which increases in crosssection towards its base. The lower part ofthetube maythus have, eg, a conical shape, a bell shape or the shape of a trumpet bell.

Afurther method of preventing wall zone particles from tumbling inwards as the loading tube is withdrawn, comprises passing a continuous stream of fluid downwards through the loading tube, with the base ofthe reactor shut-off, so that the fluid must then pass out of the loading tube base, flow outwards and then return upwardsthrough the annular gap and leavethe reactortube. The Flow pattern of the fluid in the neighbourhood ofthe the loading tube base acts to hold back the particles near the wall, while promoting the downwards and outwards movement ofthe main zone packing, contained in the loading tube, which therefore comes largely to fill the space vacated by the loading tube as it is withdrawn.

The loading tube need not be a rigid tube. A sock or an inflatable bag may be used, inserted within the reactortube. Appropriate bed particles are charged, or pre-loaded, within the sock or bag. The annular gap between the fabric and the reactor wall is loaded with appropriate packing. After deflation (if necessary) the base of the bag is opened and the bag is withdrawn.

In each of the above-described methods, it may proveuseful,whilethe loading tube (or bag orsock) is being withdrawn, to hold one or both sections of the bed down in position by pressing against that bed section some suitable restraining device, eg a suitably-shaped perforated plate located on the end of a rod or rods.

In a further embodiment, there are cases in which the above methods are complicated by the presence of other elements of a reactor, eg a tube bundle which has to be pre-assembled before being fitted in a reactorshell. In such a case, it is possible to utilise a matrix-forming material to locate an amount of an appropriate bed material. For example, a reactor tube (or one of a bundle) may be mounted in a device in which it may be spun at high speed about its axis, eg in a lathe, while suitably capped at both ends. Be fore being so spun, or during the spinning,the tube is loaded with the particles intended for the wall zone, and with an amount of a material which can form a matrix atthe reactor wal I which wil I hold the particles in position there when the spinning ceases.For example, if the tube and particles are pre-heated, and wax is used as the matrix-forming material, the wax and particles are thrown outwards to the wall and, as thetube is allowed to cool, the wax solidifies to form the desired matrix. The spinning may be terminated andthetube mounted vertically, and the main zone loaded with particles. The matrix must be removed at some stage, and this may be conveniently done in the case of wax by heating the whole until the wax melts, and the wax is drained off. Other substances used to make a matrix may be removed by appropriate methods such as solution, volatilisation or reaction. The invention will now be illustrated by the following examples.

Example 1 Atube of 70.8 mm inside diameter is packed using a sleeve (or "loading tube") of uniform outside dia meter of 36.1 mm and ofwall thickness3 mm. The annular gap is packed with a mixture of 75% by volume of ceramic catalyst support spheres of 12.7 mm diameter and 25% spheres of 7 mm diameter.

Approximately 6 cm depth of bed is packed at a time, and a similar depth of 12.7 mm spheres is packed into the main zone defined bythe loading tube, and the loading tube is partially withdrawn and the filling or packing process is repeated. When the desired total bed depth (eg 50cm) is reached, the packed reactor "A" is compared to an identical tube packed to the same bed depth with spheres of 12.7 mm diameter, reactor "B". At bed Reynolds number (based on the larger particle diameter and superficial velocity) greaterthan approximately 1100, the heat transfer rate, as represented by h1,the one-dimensional bedto-wall heat transfer coefficient, is measured to be about 15% higher in bed "than in bed "B".

Afurtherconventional bed, bed "C", was made up in identical mannerto bed "B", but using 7 mm spheres. Bed "A" and bed "C" have essentially equal void fractions. At high Reynolds number, for the same pressure drop per unit length of bed, bed "A" has a value of ha approximately40% higherthan bed "C". Additional tests carried out at the same value of h1 for both beds, demonstrate that bed "A", according to the invention, has a pressure drop 50% lower than bed "C".

Example2 Asimulationofan industriallyimportantexothermic reaction was carried out using a digital computer in order to illustrate the advantages of the bed according to the present invention, including the advantage of a shorter mean heat transfer path length, which is not quantifiable from Example 1. The reaction chosen for simulation is the oxidation of ethylene to ethylene oxide, performed in a great ex- cess of ethylene, and operating at a low per-pass conversion. The reaction kinetics used are representative of a particular industrial catalyst, but the operating conditions and bed dimensions used are illustrat iveonly, and are not claimed to be those used in industrial practice. The comparison is made between beds "A" and "C" of Example 1.A mathematical model was devised to include the radial void fraction distributions, radial profiles offluid velocity and heat transfer rates of the two cases. The comparison was performed at the same feed composition of 8% 02.

For the first comparative simulation, the beds are 10 m long and contain the same mass of catalyst. The mass flux in each is 21 000 kg/m2hr. Forthe same hot spot temperature of 226.5"C, bed "C" has a feed tem- perature of 200"C and achieves an ethylene conversion of 1.95%, whereas bed "A" has a feed temperature of 204.5"C and achieves a conversion of 2.60%, an increase of over 30% in conversion.It is possible to calculate that not only has bed pressure drop been reduced by 50% as in Example 1, but that recycle ethylene flow to the reactor has been decreased byapproximately25%. Moreover, the cost pertonne of separating the product ethylene oxide from the reactor effluent will be reduced, since its concentration in the effluent will be higher.

In a further comparison, using the same feed conditions, the bed according to the invention can achieve the same 1.95% conversion as the conventional bed, in a reactorofonly7.17 m packed length. In this case, the pressure drop is found to be 36% of that of the conventional bed.

lnayetfurthercomparison,thebedaccordingto the invention may be operated at a higher mass flux than the conventional bed. With a mass flux raised by 10% and the hot spot temperature retained at 226.5"C, a 10 m length of bed according to the invention gives a conversion of 2.40%.

The calculations mentioned above, and many others, show clearly that the novel packed beds according to the invention offer many advantages over conventional beds, including, interalia, exploiting the better heattransferto reduce hot spot tempera- tures (or tube skin temperatures, as appropriate), or ballast gas flow, or by holding heat transfer rates constant, to reduce pressure drop, or by using shorter beds or using larger diameter beds, or any suitable combination of such gains may be utilised. In other examples, the invention may offer betterselec- tivityamongstcompeting reactions.

Claims (17)

1. A packed bed reactor comprising a plurality of zones of packed bed material including a main zone comprising particles selected to have such dimensions and/or characteristics to provide the required characteristics ofthe bed as a whole, and a second zone adjacent a wall of the reactor and arranged in a radial orsubstantiallyradial mannerand comprising particles of such dimensions and/or characteristics as to (i) morefullyfill said second zone than would the particles of said main zone and/ or (ii) more obstruct fluid flowthrough said second zonethan would the particles of said main zone and/ or (ii) further promote radial transfer of heat in said secondzonethan would the particles of said main zone.

2. A reactor according to claim 1, wherein the second zone comprises particles of the same size and shape as those ofthe main zone, in admixture with smaller particles.

3. A reactor according to claim 2,wherein the smaller particles are arranged interstitially amongst the larger particles in the second zone.

4. A reactor according to claim 1,2 or 3, wherein the bed material is a catalyst.

5. A reactor according to claim 4, wherein smaller particles in the second zone carry a higher catalyst loading than larger particles in the main zone.

6. A reactor according to any of the preceding claims, wherein the transition from the second zone tothemainzoneisgradual.

7. A reactor according to claim 1, substantially as hereinbefore described.

8. The use of a reactor according to any one ofthe preceding claims.

9. The use of a reactor according to any one of claims 1 to 7, in a catalytic gas phase process at high Reynolds number.

10. Amethod of producing a packed reactoraccording to anyone of claims 1 to 7, comprising the formation of a second zone adjacent the wall of the reactor by the use of a loading tube, sock or bag within the reactor or by the use of a matrix-forming material and providing a quantity of particles selec ted to satisfy any of the characteristics (i) to (iii) of claim 1 in a manner such that it is retained within the second zone and packing a main zone in the remain derofthereactorwith particles selected to have the desired characteristics of the bed as a whole.

11. A method according to claim 10, wherein a loading tube is used which is withdrawn in stages as the second zone and main zone are packed.

12. Amethod according to claim 11, wherein a flow of liquid is passed downwards through the main zone and upwa rds throug h the second zone to avoid restructuring ofthe bed.

13. A method according to claim 11, wherein the loading tube is wider at its base than at its top.

14. A method according to claim 10, wherein a loading tube is used and is destroyed in situ after packing the reactor.

15. A method according to claim 10, wherein a non-rigid or inflatable sock or bag is used to define the second zone.

16. Amethod according to claim 10,whereina fluid matrix-forming material is distributed around the wall ofthe reactor is used to retain particlesfor the second zone, the material is solidified, the main zone is packed and the material forming the matrix is removed or destroyed.

17. A method according to claim 10, substantially as hereinbefore described.