Patent Application
20080093163
The invention relates to apparatuses intended for performing one or more chemical reactions within a gas, such as catalytic conversion and/or retaining particles suspended in the gas, as well as to silencers incorporating such functions additional to that of suppressing noise.
While previously beds of pellets or the like covered by a catalytic coating were commonly used to perform catalytic reactions in a gas flowing through such a bed, honeycombs with parallel channels have nowadays instead become the most common carrier for catalytic materials, gas flowing inside the channels in close contact with catalytic layers covering the inner surfaces of the channels. When the prime object is to carry out catalytic functions, not retaining suspended particles, the channels will usually be open at both ends, and the walls will be impervious to gas. A catalytic honeycomb can be manufactured in a number of ways, one very common method being to rely on wrapping a corrugated foil into a helical configuration that will appear substantially as a cylinder. Other well-known methods can be used to manufacture honeycombs of other shapes, such as for instance a hexagonal cross-sectional shape that is particularly useful in big catalysers composed of a multitude of smaller honeycomb modules.
Small honeycombs are generally easier and cheaper to manufacture than big honeycombs, especially with some manufacturing processes; an example in point is when a honeycomb is manufactured by extruding a ceramic material. A range of standardised sizes of cylindrical honeycombs, up to certain maximum diameters, have evolved in industrial practice. To use these standard products instead of tailor-made honeycombs of large size, odd shape, etc. in many situations will represent a number of practical advantages. Thus, for instance, a designer of a truck silencer capable of performing exhaust gas after-treatment will often prefer to arrange an array of relatively small, cylindrical honeycombs to be through-flowed in parallel. One advantage is that the price may be kept low, due to mass-fabrication. A second advantage is that there may be several alternative or supplementing sub-suppliers to choose from. A third advantage is that the reliable functionality of standard-manufactured honeycombs is based on a long record of operation, minimising the risk of many sorts of problems, including mechanical integrity of the units.
Filters for retaining particles suspended in a gas, for instance particles of un-burnt fuel, mainly consisting of carbon, suspended in the exhaust from a combustion engine, are also usually designed as honeycombs. The most common type, known as the wall-flow filter, is often manufactured as a ceramic monolith, for instance from cordierite or SiC. Roughly half of the channels will be closed at one end, and the other channels will be closed at the other end. The walls between the channels are made with pores, so that gas can penetrate them, while the pores will retain particles. Since no channel is open at both ends, gas will be forced to pass from a channel being open at the end of incoming gas onto a neighbour channel whose front end is closed, but whose rear end is open.
Combustible particles retained in the pores of a wall-flow filter should be oxidised so that they can escape in gas form, mainly as carbon-dioxide, and build-up of excessive pressure drop of the filter may be prevented or at least delayed. For such purpose, catalytic materials which promote oxidation may be added to a wall-flow filter.
A variety of honeycombs or the like that can perform both catalysis and retaining particles can be manufactured from metal foils that are wrapped up and shaped in various ways so as to guide the gas through the unit in flow patterns that may take many forms, depending on details of the structure. The foils will often be provided with perforations. One particular form of such a honeycomb-like structure may be characterised as a ‘half-open’ filter. By virtue of perforations and local deformations of the foil material adjacent to the holes, gas will both flow in channel-like passages and be forced to flow through perforations in a tortuous way. Heat-resistant mats are arranged within the foiled structure in such a way that gas is forced to impinge upon these mats. Thereby, due to ballistic forces, gas-suspended particles will penetrate into the mats to be captured, while the gas bends and flows further, by-passing the mat.
Although in some cases gas flow may locally be transgressing a wall, a foil or the like, as has been exemplified by the wall-flow filter and the half-open filter, in the following, the term ‘honeycomb’ will refer to a structure that has been designed so that the overall flow pattern is characterised by a flow from an inflow side of the structure to an outflow side, along flow-lines extending inside substantially parallel passages or channels.
Although honeycombs have proven very useful in many respects, there are some problems associated with them:
A very common problem associated with honeycombs, when installed in silencers or in chemical industrial plants, is that there may be deviations from a uniform flow from the inflow side to the outflow side, i.e. there will be a flow maldistribution in the honeycomb or an assembly of parallel honeycomb modules.
Honeycombs are generally arranged inside casings. Pipes or passages leading gas to and from such a casing will normally have transverse dimensions that are significantly smaller than those of the honeycomb or the assembly of honeycomb modules. When space is ample, one can make the casing considerably longer than the length of the honeycomb(s), in which case differences in honeycomb and passage cross-sectional areas may represent no great problem. However, space is often restricted, so that upstream and downstream distances from the openings of the casing to the honeycomb(s) will be short. In particular, a short distance from the inlet passage opening to the honeycomb(s) often represents a problem, since incoming flow of a relatively big velocity will have a tendency to impinge upon the honeycomb(s), causing relatively more gas flow to pass through the part being impinged.
One important application in which this problem is especially difficult is silencers comprising honeycombs. For the purpose of reducing noise of low frequencies, silencers often comprise two or more chambers which are interconnected by passages that must be sufficiently long and of a sufficiently small cross-sectional area compared to the cross-sectional area of the chambers inside which honeycombs can conveniently be accommodated. The relatively small cross-sectional area of the passages will then tend to cause particularly strong impingement of the kind that has been explained above.
Another common source of honeycomb flow maldistribution is that a passage may lead gas to a casing in a direction that deviates strongly from that of the channels of the one or more honeycombs. Thus, for instance, flow may be led in from a perpendicular direction. A silencer feature that has been found very useful in later years, since it allows for a long passage to be accommodated inside a rather small casing, is to accommodate a helically extending passage to connect two chambers. Such a passage may for instance be arranged to extend around a cylindrical honeycomb. In such a case the flow situation in front of the honeycomb may be that of a tangential, radial inward flow all the way around the periphery of the honeycomb.
Catalyser honeycombs in which both ends of the channels are open are especially sensitive to flow maldistribution. Wall-flow filters are less sensitive for two main reasons; firstly, the generally higher flow resistance in wall-flow filters helps even out flow maldistribution. Secondly, if some channels are swept by a bigger flow than other channels, there will be a tendency for more particles to accumulate in those channels, which in turn will slow down flow in those channels. In other words, wall-flow filters may to some extent provide an automatic self-repair of a flow maldistribution. Still, flow maldistribution is not without problems even with wall-flow filters. Thus, a continuous oxidation of accumulated particles, known as regeneration, may suffer from flow maldistribution, which contributes to making the regeneration less continuous, and to making the filter operate with a higher time-mean pressure drop.
One known method of promoting a more equal flow distribution across the inlet to a honeycomb configuration contained within a chamber is to insert a perforated plate or a screen between the inlet and the honeycomb(s). The bigger the pressure drop across the plate or screen, the more effective the evening-out effect will be, but there is often a penalty associated with a big pressure drop. In the case of the exhaust system of an engine, there will be an adverse effect on engine performance. In the case of a catalyser of a chemical plant, more fan power will be required to force gas flow through the apparatus.
Another known method relies on the use of a diffuser arranged at the inlet to a casing containing a honeycomb arrangement. Diffusers are attractive in two main respects; due to a widening flow area in flow direction, diffusers will slow down the flow which in itself will promote a more even flow distribution. Secondly, across diffusers that operate as they should (i.e. not exhibiting flow separation within the diffuser) there will be negative pressure drop across the diffuser, known as pressure recovery. The main problem with diffusers is that they tend to require space that may not always be available to the extent one would prefer from a fluid-dynamic point of view.
A third method that is less common, but which has been suggested in prior art, is to shape honeycombs with an inlet surface that deviates from a plane surface. Thus, for instance, it has been suggested to use a conical or tapered inlet shape of honeycombs to adapt them to the shape of an inlet diffuser. Such solutions could deserve consideration in cases of an incoming flow being more or less in line with the overall flow direction inside the honeycomb. However, cost reasons may detract from this type of solution. Also, even theoretic, odd-shaped honeycombs cannot be adapted to all kinds of difficult flow situations.
A common characteristic of the problems described above is that honeycombs, although being very fit for a number of chemical/particle retaining processes as such, and although being available on the market at comparatively low price, pose difficulties from the point of view of attaining optimal flow conditions. This aspect is especially prominent when space is limited, and silencers, due to geometric requirements prompted by acoustic considerations, often present the designer with especially difficult flow problems when honeycombs are to be inserted and operate optimally in respect of pressure, specific inflow or outflow directions, etc.
Grid or netlike catalysers are known to possess a number of appealing characteristics; they can accommodate more catalytic surface per unit volume than honeycomb catalysers which have been utilised, either for the purpose of creating a more compact catalyser or with the scope of attaining a smaller pressure drop within a given volume. Also, such structures have been appreciated for their capacity to promote more intensive kinetic activity adjacent to the catalytic layer, and grid or netlike structures are known to respond more rapidly to temperature changes.
An important aspect of the invention relies on the insight that grid or netlike structures or other open structures, due to a facility of permitting flow to propagate through the structure much more freely in space than in a honeycomb, can solve many of the kinds of flow problems pointed out above.
In this summary we shall be referring to these open structures as ‘matrix structures’; in the claims a strict definition of this concept is given. It is true that classical beds of pellets or the like share a facility for allowing flow to propagate more freely in various directions. However, pellets or the like have to be contained in some way, which makes them less attractive than the kind of coherent matrix structures that are covered by the invention.
A further aspect of the invention is that such matrix structures may be adapted for supplementing honeycomb arrangements, to obtain combinations of better fluid-dynamic flow situation within the honeycombs, to accommodate more chemical treatment capacity, such as catalysis, within a given available space, and a number of additional attractive features that will be demonstrated below.
As an example of this approach, in some cases supplementing a honeycomb by a matrix structure may surprisingly lower the pressure drop across an apparatus.
Furthermore, as will also be illustrated, the invention provides the possibility to arrange matrix structures or elements of matrix structures inside honeycombs to boost their performance in various ways. In such arrangements matrix structures provide hitherto unnoticed possibilities of arranging different catalytic coatings very close to each other, which may be utilised for boosting chemical reactions.
Matrix structures may advantageously be accommodated inside silencers in a variety of ways, some of which have been mentioned already. A still further aspect of the invention, related to silencers, is that matrix structures have been found to have a capacity for affecting the acoustic and fluid-flow performance of passages connecting acoustic chambers in a favourable way.
a and b show a first embodiment of the invention.
a and b show a ninth embodiment of the invention.
a and b show a tenth embodiment of the invention.
a and b show an eleventh embodiment of the invention.
a-c show a twentieth embodiment of the invention.
a and b show a twenty-second embodiment of the invention.
a and b show a twenty-fifth embodiment of the invention.
Within a given total catalyser volume, the amount of catalyst surface for the conversion of gas or particles is restricted by considerations on channel level, and the permissible total volume of catalysers is restricted, especially if a reasonably even distribution of gas among various channels is prescribed. In theory, channels could be made of a very small diameter, but this would produce an excessive amount of pressure drop across the catalysers.
a and b show a first embodiment of a silencer according to the invention. A silencer is shown with an outer, squared casing 1 and inlet and outlet openings 6 and 7 of cross-sectional sizes a1 and a3, respectively. An internal divisional wall 6 divides the interior of the silencer into two acoustical chambers 10 and 11. Both chambers are of a significant size and of a significant cross-sectional area A due to a significant transverse size D of the casing of the outer shell. Inside the silencer, flow is passed from the first to the second chamber via an internal passage 5 of a cross-sectional area a2 and of a length LP. The tail pipe 9 leading gas from the silencer to the atmosphere has a length LT.
Inside the first chamber 10a matrix-structured catalyser 50 consisting of a succession of plane screens or grids or nets 52 is arranged to almost fill out the entire space of the first chamber. All surfaces of the screens are covered by a catalytic layer.
A face-view I-I of a screen is shown in
Inside the second chamber 11 another type of grid or net 51 has been arranged. Here, additional elongated members 55 are arranged in the axial direction, as appears from
The first embodiment of the invention illustrates that silencers provide an excellent example of utilising facilities of control of fluid flow inside a catalyser, provided by matrix structures. The better the desired acoustic performance, the bigger should be the ratios between chamber cross-sectional areas A (=D×D in
Due to the openness of both matrix structures, the two chambers 10 and 11 will acoustically function as chambers, not being effectively disrupted by the presence of the matrix structures. The three arrows pointing at a space upstream of matrix 51, matrix 51, and the space downstream of 51, respectively, by way of illustration underscores this aspect. The silencer shown will function as an acoustical low-pass filter according to a well-known formula:
f=(c/(2π)sqrt((a2/LP)(1/V1+1/V2)
where c is the speed of sound, V1 is the volume of chamber 10, and V2 is the volume of chamber 11. If f is low enough, noise reduction will become effective in a broad acoustical spectrum, including relatively low frequencies that are higher than f. f is sometimes termed ‘local natural’ or Helmholtz frequency of the filter.
The volumes of the solid parts of matrices 50 and 51 will surely detract from acoustically effective volumes V1 and V2, respectively, but this effect will not be severe due to the relative openness of the two structures. The main acoustic effect of the matrices will be beneficial, since the matrices will tend to dampen standing waves set up in the chambers, waves that will have pressure nodes inside the chambers and maximum pressure variation at the walls, as the person skilled in the art of acoustics will appreciate.
Honeycombs with channels that are open at both ends, arranged inside chambers may also be viewed as relatively open structures from an acoustical point of view. Thus, if a honeycomb catalyser of this sort substantially covers the entire cross-sectional area A of a chamber, it may be justified to regard the chamber as en entity in acoustical considerations, for instance when applying the above formula.
Wall-flow filters, in which gas is forced through pores at a relatively high flow resistance, and which provide such a great resistance to acoustic waves, may be dubious in this regard, that is, when they are arranged inside a chamber it may be dubious, at least in some circumstances, to regard such a chamber as an acoustic entity. Especially if a wall-flow filter is composed of a plurality of cylindrical modules, being through-flowed in parallel, a pronounced acoustical effect may be obtained from dividing a chamber into two parts. This is because there must be walls or the like between the modules in order to prevent gas flow from by-passing filter modules.
This discussion of to which extent various types of gas treatment equipment will alter acoustic function of chambers illuminates an acoustically attractive feature of matrix structures arranged inside silencers to add chemical treatment of gas to silencing functions: if a certain degree of chemical change is provided within a silencer, such as a certain percentage of reduction of a certain component of gas flowing through the apparatus, a matrix structure, such as the grids or nets shown in the first embodiment of the invention, provides a possibility of affecting acoustically effective volumes of chambers to a minimal degree. The effect on low-frequency performance may even be positive. Sound corresponding to natural frequencies f (according to the formula) often passes through a silencer at a relatively lower degree of damping. The lowest natural frequency may even be slightly amplified by the silencer. Inserting a matrix structure inside a chamber will introduce a small, but not always insignificant damping of such frequencies.
The acoustically favourable effects described may be attributed to two fundamental aspects of matrix structures of the kind covered by the present invention: firstly, the surface-to-volume ratio is relatively big, and secondly, these structures may be adapted to cause greater kinetic activity in the gas close to the surface. An example of this may be if the two matrix structures 50 and 51 are both composed from threads, as the person skilled in the art will appreciate. Both these aspects also account for favourable aspects of matrix structures from a chemical point of view, as has been pointed out already in the last paragraph explaining the background of the invention.
Screens according to structure 50 may be stamped out of sheets of metal, in which case the voids that appear in a face view as shown in
Both structures 50 and 51 are shown to be organised in regular patterns. Alternatively, they could organised in a partly or fully random manner. One example is a metallic ‘sponge’ composed of intermingling treads and having an envelope as a lump or a flat, round sheet. Typically, such truly 3-dimensional structures will have envelope dimensions that are significantly bigger that the biggest cross-sectional dimension of the individual threads, fibres, strips or the like from which the structure are made.
No cross-sectional view of this third embodiment is shown. The walls and the passages could be planes that extend from one side of the casing to the opposite side, in a direction perpendicular to the plane of
At the end of the description, embodiments of the invention will be shown that constitute other matrix structures. These matrix structures also have the feature of plane sheets extending in a longitudinal direction of flow from one side of a matrix structure to the opposite side.
In
In the particular embodiment, the catalyser is an SCR catalyser for promoting conversion of NOx by chemical reactions between NOx and urea, injected via a nozzle 30 arranged inside the pipe 8 leading up to the silencer. In order to avoid condensation of vapour on the inside of the outer shell of the first chamber, the perforated inner cylinder inside the first chamber 10 has been replaced by a solid cylinder 15 and material 14 (for instance mineral wool) providing sound and heat insulation, but no sound absorption. In the second chamber, the sound-absorptive function has been retained by having sound-absorptive material 12 being contained between outer cylinder 2 and perforated, inner cylinder 13.
The matrix structure 50 may increase the pressure drop of flow passing the helical passage, even though, surprisingly to a person not being very skilled in the art of fluid dynamics, the opposite could in fact be the case, as will be explained below. In any case, there will be minor flow separation on the ‘backsides’ of elements of the matrix, as in the classical example of a rod that has been arranged in a flow field. This has the acoustic advantage of suppressing resonances in this passage, above all the resonance which correspond to a half-wave set up in the passage from inlet to outlet (extending helically, as the passage itself). Also, a pronounced effect of damping Helmholtz natural frequencies f may be obtained. This effect may be explained in the following way: in a first approximation, particularly apt in relation to lower sound frequencies, the system may acoustically be viewed as a mass fixed onto two springs, the mass being that of gas contained within the connecting passage and the two connected chambers representing flexible springs. When the silencer is subjected to a sound of a frequency corresponding to the natural frequency of this spring, there will be a resonance effect that may even seriously compromise the performance of the silencer. But when a matrix is inserted into the passage, the dynamic, frictional resistance caused by the matrix will dampen this resonance.
By selecting a very open structure of the matrix, any pressure drop increase that may be caused by the structure can be reduced to a minimum. Another possibility is to make the cross-sectional area of the helical passage somewhat bigger, optionally to be combined with increasing the length of the passage somewhat, to maintain an unchanged local Helmholtz natural frequency f.
When a matrix structure, as has been explained above, is inserted into a helically extending passage, the relative increase in flow resistance will be smaller than if the same matrix structure had been inserted into a straight passage. The reason for this is that in the empty, helical passage, secondary swirl, within a cross-section, will be set up to increase the flow resistance compared to a corresponding, straight passage. Inserting a matrix structure in the empty helical passage will dampen such swirl or even eliminate it. Since circular passages are more prone to such secondary swirl, the fluid flow performance of such passages (instead of the rectangular passages of
In some cases insertion of a matrix structure inside a diffuser, straight or curved, may improve the fluid-flow performance dramatically. This happens when the diffuser has been made with a too great increase of flow area in flow direction, such as when the length is given, and the ratio between inlet and outlet passage cross-sectional area is chosen to be too large. In that case, flow separation will occur, that is flow will not really widen, but will cling to one of sides, as is well-known from elementary fluid mechanics. A matrix structure inserted in such a diffuser can eliminate such flow separation, provided the degree of area increase is not too big. Thus, if a passage between two chambers is made up of one part of a constant cross-sectional area and a second part being a diffuser, an optimal solution may be to insert a matrix structure in the diffuser part only.
a and b show a ninth embodiment of the invention.
Gas is led into a cylindrical casing 1 via inlet pipe 8, the direction of which coincides with a tangent of the fictive cylinder, i.e. gas is led into the casing in a tangential direction. Thereby, a rotational flow field is created upstream of the catalyser. If the catalyser had been of the type with elongated, parallel channels instead of the matrix structure, the fluid flow transition from rotation to co-axial flow inside the catalyser would have been rather abrupt, and the centrifugal force upstream of the catalyser would have created a tendency for more flow to pass through channels adjacent to the casing shell, compared to channels closer to the centre of the catalyser.
Now the catalyser is instead of such a structure that flow inside the catalyser in addition to co-axial flow will be both tangential and radial, which is advantageous. Due to radial flow inside the catalyser there is an evening-out effect, so that the upstream tendency to a flow concentration toward the outer parts of the catalyser is gradually being reduced so that a more even radial flow distribution will exist at the outlet from the matrix. The rotational flow is gradually being ‘braked’ inside the matrix, by which dynamic energy of the rotation is being utilised for promoting catalytic activity, due to more intense gas kinetic activity between the catalytic layers and the bulk gas flow between the matrix structure solid elements.
a and b show a tenth embodiment of the invention, once again applied to a casing with a cylindrical shell 2. A curved, helically extending passage 6 connects an outer, annual chamber 10 with an inner, circular-cylindrical cavity 12 being part of a second acoustic chamber also comprising a cavity 14. The cylindrical cavity contains a matrix-structured catalyser 50. A cylinder 4 forms both an outer casing of this cavity and the inner, cylindrical part of the helical passage. Due to a tangential inflow to the outer chamber via pipe 8, gas will, in addition to moving axially, rotate around the axis of the silencer. Gas leaves the passage along a helically extending outlet section 13 from which it passes on to a flat, cylindrical cavity 14, where flow turns 180 degrees in an inward, radial movement, the gas flow being guided by vanes 15. The vanes will tend to take away the rotational movement, so that flow towards the catalyser, in an axial direction to the left in
The general arrangement of this embodiment will appear akin to some designs known from prior art, related to other catalysers than a catalyser designed as a matrix structure according to the invention. However, the matrix-structured catalyser allows for both tangential and radial movement inside the catalyser, once again causing an evening-out of the flow inside the catalyser.
Instead of vanes forcing the flow to stop its rotational movement around the axis of the silencer, less fluid-flow forcing vane shape could have been adopted, or one might even leave out the vanes; in the case of a honeycomb replacing the matrix structure, just leaving out the vanes may simply be unacceptable because of resulting flow maldistribution within the honeycomb.
A cavity 14 may be contained within a cap part 15 of the silencer that by not shown details, related to the transitional zone 16 of the cylindrical casing, can be arranged so that the cap is de-mountable.
a and b, representing an eleventh embodiment of the invention, once again show an arrangement which in its overall arrangement resembles prior art, but with the important distinction of the incorporation of two matrix-structured catalysers 50 and 51 instead of using honeycombs with parallel flows inside aligned channels. The differing matrix-structured catalysers will improve the function of the apparatus, partly due to effects already described in relation to the two foregoing embodiments of the invention. In addition, the catalyser 51 will cause an evening-out of the inflow to the honeycomb 30, which will improve its performance, compared to a design without the pre-catalyser 51.
From inlet pipe 8 flow enters the outer, annular chamber 10 by a tangential motion, so that a rotational movement is set up inside this chamber, upstream of matrix structure 50, and flow is distributed across the inlet to this structure. Inside this catalytically covered matrix structure, the rotational flow is being ‘braked’ (similar to what has previously been described), causing an added catalytic activity. From this first catalyser flow moves axially to the right inside annular cavity 12 towards a chamber-interconnecting member 20. This member contains two spirally extending passages 13 and 14, separated by spirally extending walls 15. Hereby, there will be a gradual inward, radial movement of gas.
As in previously shown silencers with helically extending passages connecting two acoustical chambers, gas will have to move a longer distance, compared to a simple, rectilinear movement inwardly; this improves attenuation of low-frequency noise. The chamber-interconnecting member 18 is delimited to the left by a wall 16 and to the right by a wall 19, both being parts of an end cap unit, containing helical passages 13 and 14 and being de-mountable as a separate unit, due to not shown details that the person skilled in the art will readily be able to supplement. The first wall 16 is provided with two openings 17 and 18, as indicated in
In the centre part 23 of the chamber-interconnecting member 20, gas leaves the two passages 13 and 14 in a tangential direction, which sets up a flow field of rotation in this centre part. In the case of a conventional arrangement with a central honeycomb only, for flow to pass from this flow field to an axial inflow in the left direction, to parallel channels of it would be necessary to arrange for some distance to the catalyser. Still, evening-out of flow will not be perfect, so that there will be different flow rates in differing channels of the catalyser.
By contrast, in the inventive embodiment shown here flow evening-out takes place in a matrix structure 51, arranged upstream the conventional honeycomb catalyser 30, and there is no need for a cavity. Thus, more catalyser can be accommodated within a given total space, and the conventional catalyser 30 will perform more effectively as a catalyser.
Dividing the central catalyser into two parts, as is done in this embodiment, provides the engineer with more freedom when optimising the catalytic effect. For instance, the pre-catalyser 51 may serve the purpose of finalizing mixing of urea that in the case of SCR-catalysis may have been injected into the silencer upstream of, or directly into passages 13 and 14 (such injections are not shown in the figure but are easily imaginable). Another feature of the embodiment will be that the matrix structure, due to its big catalytic surface-to-solid-volume ratio will respond more quickly to changes in mass flow rates, a feature that will be particularly useful when the apparatus is applied to vehicles, where the engine sometimes changes its mode of operation vary quickly, due to speed changes of the vehicle. The case of upstart from a standstill condition with a cold exhaust system is a particular case in point in this respect.
Instead of a honeycomb whose channels are open at both ends, the honeycomb 30 could have been a wall-flow filter. In that case, both catalysers 50 and 51, or one of them, could perform oxidation of NO into NO2 that will promote regeneration of the filter. Since two catalysers are present, they may alternatively be differently coated, so that one of them is especially suited for oxidising other elements than NO. There might even exist a situation where a silencer, incorporating a honeycomb resembling the design shown in
Gas is led into the passage via inlet 1 and out via outlet 3. In the centre of the honeycomb a hole 7 has been indicated. Such a hole may be utilised for accommodating an internal tailpipe, i.e. a pipe that will extend perpendicularly to the plane shown, providing a connection from a chamber to the external environment. In such an arrangement, parts of the geometry may connect such a tailpipe with second chamber 11.
The curved passage has been formed within the residual space between the outer shell geometry 2 and the outer geometry 31 of the honeycomb. As can be seen, the cross-sectional area of this passage varies; in the ‘corners’ 7 the area is greatest. Here, lumps of matrix structures 50 have been inserted. Due to the micro-separational flow surrounding threads of the matrices, there will be no major flow separation in the corners, a phenomenon that most probably would have occurred in the absence of the matrices. Such major separation would cause significant pressure drops within the curved passage. Apart from preventing major flow separation there are two benefits associated with ‘micro-separation’ associated with this phenomenon: firstly, when the matrices are supplied with catalytic coating, such as a coating promoting oxidation, micro-separation will enhance this oxidation, due to lively kinetic activity around solid elements of the matrices. Secondly, there will be damping which will have beneficial acoustic effects.
It should be added that the embodiment shown in
As the person skilled in the art will appreciate, a number of the previously described configurations, illustrated by application to silencers, may also be adopted in arrangements where chemical reactions are in focus and there may be no specific intention of suppressing noise, even though dampening of noise or pulsations in flow may occur as a by-product process.
Now some further embodiments will shown, without specific reference to silencers. However, the person skilled in the art will appreciate that they can all be accommodated in silencers.
Gas enters a chamber 1 containing a cylindrical honeycomb 30 whose channels are open at both ends. The cross-sectional area A of the honeycomb is much bigger than the areas of not only the passages leading gas to and from the chamber, but also compared to the outlet area of the diffuser 7. As is often the case when space is narrow, the inlet to the honeycomb has been arranged at a relatively small distance from the inlet to the chamber. Thus, there will be a tendency for the incoming flow to impinge onto the central part of the honeycomb, so here relatively more flow will pass through the channels. In other words, the residence time of gas within the channels is greater in the more peripherally positioned channels than in the central one.
Such flow maldistribution may be disturbing in any case of a honeycomb catalyser performing one or more chemical processes. The actual effects of flow maldistribution will depend on the specific reactions. In some cases, where chemical reactions depend on the degree of turbulence inside channels, there will be less activity in channels with a slower flow. Other processes, Depending less upon the degree of turbulence than on time, will suffer from a too short residence time of gas in the core region, resulting in less chemical conversion inside channels.
Tailoring matrix structures to the various types of reactions taking place will prompt different strategies. In the former case one can shape a matrix or two matrices, up and downstream of a honeycomb so that there will be relatively more added catalytic material in series with channels where flow is slower. Conversely, if the reaction is more time-controlled, one may fit in relatively more added catalyser capacity in series with channels of less residence time.
Between
As can be seen, the total pressure drop across the entire arrangement, as taken from the inlet point upstream the diffuser to a downstream point inside the pipe leading gas from the casing, in the inventive embodiment has not surprisingly become smaller.
This can be attributed to the more widening diffuser. The comparison is fair, since considering the length of the diffuser to be given, in the prior art configuration it would not be realistic to adopt the more widening diffuser, since then flow inside the diffuser would be subjected to major separation, that it, it would not be possible for the flow to adhere to both the upper and the lower diverging contours. Due to the flow resistance arranged immediately in front of the diffuser outlet in
The more widening diffuser, with outwardly rounding contour immediately upstream the interface with the casing, represents a third feature of assisting more even flow distribution across the inlet surface to the honeycomb.
a and b shows a fourteenth embodiment of the invention. Here, gas flow is led tangentially into a casing 1 in which a honeycomb 30 has its axis oriented perpendicular to the axis of the inlet pipe 8. As in
a and b show a sixteenth embodiment of the invention, being part of a silencer. Inside a cylindrical casing 1, a helically extending passage 5 is wound around a cylindrical honeycomb 30. Flow exits the passage by flowing into annular spacing 6. From this spacing flow enters a circular matrix structure 50 with built-in guide vanes 7, shaped in a more flow-friendly way than the simple, straight vanes adopted in
An arrangement of this kind opens up for many possibilities of tailoring a catalyser in various ways. Three examples will be given here in broad terms: First, the thin thread within the channel represents are small thermal mass per catalyser surface unit, which helps speed up the thermal response of the assembly. Second, two different catalytic coatings can be arranged very close to each other by simple manufacture. Third, by deliberately arranging the thread very close to the wall of a channel or even to touch it, one might ‘short-cut’ chemical reactions that would otherwise be performed in separate steps if gas were to meet first one type of catalytic coating outside in one catalyser and then another type in a downstream catalyser. In other words, arranging a thread inside a channel opens up for concurrent chemical reactions without resorting to complicated procedures of arranging varying catalyser coatings onto inner surface walls of a honeycomb, such as for instance a checkered pattern of alternating coating that would be rather costly to manufacture.
a-c show twentieth embodiment of the invention. Here, a grid 51, being an example of a matrix structure, has been arranged right at the inlet to a honeycomb catalyser 30. As can be seen from
The catalytic material of the threads can so that they will cause oxidation, such as oxidation of NO to NO2. The walls of the filter may be catalytically coated, or that may not be considered necessary for a sufficient oxidation of accumulated particles and other combustible elements by NO2 and O2 to become effective. One might even envisage that singular oxygen molecules O in such an arrangement can be created, and will combine directly with combustible matter before combining with NO. If this is possible, a very efficient way of combusting accumulated particles of a wall-filter will have been achieved. This provides an example of what has above, as related to a honeycomb with both ends open, been designated ‘shortcut’ of chemical reactions or concurrent reactions.
a and b show a twenty-second embodiment of the invention, where thin threads 50 shaped as helical springs have been arranged inside channels of a wall-flow filter, to be held fixed by way of friction within the channels. As is commonly known, if a spring is subjected to torsion in one direction of turning, its outer diameter will shrink. By taking advantage of this, a simple tool for arranging the spring inside a channel can be designed. The tool can be composed of a cylindrical rod whose diameter is slightly smaller than the inner diameter of the spring when it is initially left free, arranged around this rod. A wrench can then apply torsion onto the spring whose diameter will shrink until it comes into contact with the rod. Then the assembly of the spring and the rod can be pushed into the channel. When the spring has been positioned axially, the torsion is released, the spring will expand onto the walls, and the rod can be taken out. If the channels 32 are not circular, but for instance squared as shown in
Naturally, when elongated members or threads extend into channels of a honeycomb, they will increase the flow velocity inside the channels, which in some cases can be beneficial in that it may promote kinetic activity and thereby chemical processes. On the other hand, the higher flow velocity will increase the pressure drop. Especially in the case of a catalyser honeycomb with both ends open, where the pressure drop across the honeycomb without any insertions can be low, this may be a draw-back. Therefore, when threads are inserted in a honeycomb that has already been designed for a specific purpose, at will be advisable to use elongated members or threads that occupy only a certain part of the cross-sectional area of a channel, such as at the most 50%, or more preferably at the most 25% or even more preferably at the most 10% of the cross-sectional area.
On the other hand, when a honeycomb catalyser is being designed with the concept of inserting elongated members or threads, more channels or channels with somewhat bigger cross-sectional areas can be selected, to make a compensation that will lower the pressure drop across the combined arrangement.
The concept of having threads penetrating into a honeycomb offers a possibility of selecting shapes of individual threads, such as thickness or length of threads individually for the various channels, for the purpose of compensating for any mal-distribution of flow across the channels of the honeycomb.
The concept of arranging threads inside channels of a honeycomb represents a very space-economical way of enhancing the performance of a honeycomb, whether it be a catalyser with open ends and impervious walls, a wall-flow filter or any other kind of honeycomb.
The person skilled in the art will appreciate that this idea can be extended to having threads or any shape of elongated members being arranged into any kind of honeycomb structure, even kinds not described above. The company Emitec GmbH has invented a great many honeycomb designs created by shaping metal foils in many ways. One group of such honeycombs are represented by the previously described filter with mats inserted for capturing particles. But in a many disclosures by patent applications this company and other companies have shown that metal foil honeycombs can be adapted in numerous ways, to tailor the flow pattern through the honeycomb according to many design requirements. In almost any such design insertion of threads or elongated members can boost performance, for instance by the virtue of a thread responding more rapidly to changes in gas temperature, due to the small thermal mass of a thread, compared to a metal foil.
A last embodiment that is once again a silencer will be explained below. This embodiment represents a modification of a prior art silencer that will first be explained, as a background.
a-c show a prior-art silencer with inserted catalysers adapted for performing SCR DENOX onto an exhaust gas being silenced as well. The apparatus outwardly appears with a box-like shell 1. Exhaust gas from a combustion engine is led to the apparatus via pipe 2 and from the apparatus via outlet pipe 3. Quite a distance upstream of the apparatus a nozzle 30 has been arranged for adding urea to the gas. The pipe leading up to the apparatus will allow for urea to mix with gas and evaporate before entering the apparatus. The interior has three cavities 4, 5, and 6, being created by flat division walls 7 and 8. In line with previous considerations, cavities 5 and 6 can be regarded as constituting a single acoustic chamber. An internal pipe 9 leads gas from the first cavity 4 through cavity 5 into cavity 6, the pipe providing full separation between internal pipe flow and chamber 5 by having impervious walls where the pipe leads gas through chamber 5. Pipe 9 inside chamber 4 has a perforated inlet section 10 with holes 11 through which gas enters the pipe from chamber 5.
In total 6 honeycomb catalyser modules 31 to 36 are have been arranged to be through-flowed in parallel. They are held fixed onto division wall 8 that has been provided with in total seven holes 37 of diameters adapted for arranging the modules as well as pipe 9. Each catalyser module consists of a central honeycomb 38, a cylindrical metal casing 39, and an intermediate heat-resistant and flexible layer 40. Further, each catalyser module has been divided into a first and major part 41, where catalyser channels have been supplied with a catalyst adapted so as to perform DENOX. A rear and small part, 42 has been provided with a different catalyst adapted for reducing ‘slip’, that urea rest products that have not been consumed in the upstream DENOX process.
Cavity 6 is a combined distribution and flow reversal cavity. Here, gas from pipe 9 is being divided into six part flows through each catalyser module, at the same time reversing its flow direction, as seen in planes B-B and C-C of
An arrangement of the kind shown is quite common in prior art. Of course, more catalytic surface could have been built into the apparatus by selecting a rectangular honeycomb instead of the six cylindrical modules, but some honeycomb technologies at present simply do not permit designing for such a large and odd-shaped a honeycomb at a reasonable price, and even with those honeycomb technologies that offer better possibilities of adapting the shape of a honeycomb, there will be a significant price reduction associated with using smaller cylindrical modules that have been standardised in catalyser industry; the company designing the apparatus will have a whole range of alternative catalyser suppliers to select from when choosing the specific sub-supplier and catalyser technology.
a-c show a twenty-fourth embodiment of the invention, representing a modification of the prior-art apparatus of
The urea nozzle 30 has been arranged to supply urea inside the apparatus instead of into inlet pipe 2. An inlet diffuser 21 has been added to inlet pipe 2, to provide a transition form this pipe to the inlet to chamber 4. A further diffuser 22 has been added at the downstream, right-hand end of internal pipe 9. The six end caps 43 upstream of the catalyser modules have been removed. Each catalyser module is no longer divided into an upstream DENOX part and a downstream slip reducing part, but the entire catalyser modules has its channel walls covered by the same catalytic layer performing DENOX only. Three groups of matrix structures, 51, 52, and 53, have been inserted into each of the three chambers 4, 5, and 6.
Matrix structure 51 inside chamber 4 mainly serves two purposes: First, it ensures mixing of urea with gas as well as urea evaporation, due to the higher degree of local turbulence of gas flow created around elements of the matrix structure. Second, a sub-part 54 of matrix structure 51, arranged in front of the central zone of diffuser 21, ensuring that three will be no flow separation, as in explained in an akin arrangement shown previously. Matrix structure 52 has its interior surfaces in contact with through-flowing gas provided with a catalytic layer adapted for performing ‘slip’ elimination, that provides substitute for the downstream catalyser parts 42 of the prior art apparatus. Matrix structure 52 has been shown to be of a simple outer shape, to be of regular structure and not to extend to be left beyond outlets of the catalyser modules. A more refined matrix structure could have been made according and more matrix catalyser volume could have been accommodated to perform more slip reduction and to shape the flow pattern inside chamber 5 so as to enhance the performance of matrix structure 52. On the other hand, the simplicity of matrix structure 52 in itself is appealing, and it may sufficient and work well, depending on circumstances. The individual screens making up the matrix can be manufactured from a succession of sheets of nets that are cut out from bigger nets to their outer contours, and the seven holes needed for inserting the nets in-between pipe 9 and the six catalyser modules can be cut or stamped out of the nets.
Matrix structure 53 is more complicated. It can be divided into nine parts: Two parts 55 and 56 being arranged adjacent to each other, a third circular part, 57, being arranged in front of diffuser 22 and an total six circular parts 58 to 63. The diffuser slows down the inflow of gas to cavity 6, which in eases gas flow distribution inside the cavity, apart from recovering static pressure. To the right of the matrix structure there is a flat, empty space 64. Part matrix structure 57 helps prevent flow separation within diffuser 22.
Six similar matrix parts 58 to 63 provide extra flow resistance in front of the catalyser modules to promote an equal flow distribution among the modules, as well as suppressing tendencies to skewed flow across the inlets to the modules. These six matrix parts can be seen as direct replacements for the six caps 41 in the prior art apparatus. As can be seen, the extra flow resistances within part structures 57 to 63 has been attained by inserting extra, masked, circular sheets between bigger sheets. As can be seen from
The empty space 64 assists in easing flows to modules arranged at a relatively greater distance from inlet to chamber 6, as well as ensuring that, as viewed for each module, channels being situated further away from chamber inlet than other channels of the same module, will receive gas, i.e. counteracting skewed inflows to individual catalyser modules. In addition to modelling details of matrix structure 53 to optimise flow as has been described, all or some of matrix parts can be coated by a catalytic coating that will boost DENOX performance, either the same coating as that used inside the honeycomb catalyser modules, or another coating, adapted for the matrix, drawing upon its characteristics that differ from those of the honeycomb, or to perform a chemical reaction that differs from that being performed inside the honeycombs. Due to the bigger surface-to-volume ratio of fine-graded matrix elements, as compared to the walls of the honeycombs, a faster thermal response can be expected from the matrix structure. This can be taken advantage of to improve the SCR DENOX performance of the apparatus when subjected to varying operating conditions.
It has already been pointed out that inside cavity 5, to the left of gas outflows from modules, in the spacing between modules and division wall 7, more matrix structure could have been added. As an alternative to utilising this for adding more slip catalyser capacity, one could add more DENOX catalyser capacity by adding screens provided with a coating suitable for performing DENOX, upstream of slip reducing screens arranged as in the embodiment shown. Also, the structure of such added matrix could be adapted to compensate for any variation in the NOx concentration at outlets from individual channels from all six modules.
When deciding how to design details of matrix structure 52 one could proceed as follows: An apparatus will be designed with all the parts of the embodiment described, except for matrix 52. A range of velocity, NOx, and ammonia sensors is arranged at outlets from a number of channels, for instance three channels of each module being provided with anemometer velocity sensors, three other channels with NOx sensors, and three ammonia sensors at outlets of further channels of the same module, so that in total 54 sensors are inserted. They should be small enough and such arranged that their disturbance caused to the flow pattern inside chamber 5 will be minimal. If this cannot be achieved, the flow and chemical composition pattern can be mapped by transversing sensors; it will take more time to measure, but will disturb the flow less and will probably be cheaper to install.
From the patterns recorded, modifications improving the pattern can be adopted, perhaps by modifying the amount of urea added and making changes of matrix structures 51 and 52. Having performed a number of modifications, some minor deficiencies in flow and chemical composition pattern may remain. Having this knowledge will provide good input for designing matrix structure 52. As an overall consideration, one can subdivide matrix structure 52 into a DENOX part and a slip elimination part, the sizes of the two parts being tuned in conjunction with selecting the amount of urea that is being added. The optimisation procedure described may be refined to include, not only a single operation condition, but a test cycle, such as one of the standardised test cycles prescribed for testing the NOx and particulate emission performance of a particular engine provided with a certain exhaust emission treatment equipment.
Comparing pressure drops across the two apparatuses of
In the embodiment shown in
Prior to the procedure of physical experimentation there may have been a phase of computer modelling performed to start out experiments on the basis of a better design than a design in which tailoring matrix structures has been made mainly on crude calculations and qualitative reasoning. Such simulation could start by pure computerised flow simulations, not including chemical reactions, according to one of the many commercial codes for such simulation that exist today on the market. A next step could be to include computer modelling of chemical reactions.
As the person skilled in the art will appreciate, optimisation of matrix structures operating inside a given apparatus, without or (as in the example above) in conjunction with honeycombs and/or foiled structures can be designed to be combined with experimentation in many ways, similar to many other previously made optimizations of systems performing chemical conversion be catalysis or otherwise, sometimes combined with silencing noise.
By adopting the method according to the claims when designing, constructing and/or using an apparatus and/or a silencer according to any of the claims an optimisation procedure will become possible to find clues for network configurations that one could hardly have figured out in more conventional ways. By drawing upon the vast capacity of modern computers one could afford, within reasonable means to test myriads of virtual configurations, even to include randomly defined matrix structures as a way of experimenting. Modelling can include one or a few calculated variables, such as for instance mean downstream NOx and slip concentrations. In this way it can be possible to distinguish more preferable matrix structure configurations from less preferable configurations.
a and b show a twenty-fifth embodiment of the invention, in which one or more metal foils have been adapted to constitute a matrix structure according to the invention.
The structure receives gas flow from the left. In
It can be seen that due to the off-set arrangement of holes in rows 52-54, flow within a plane is forced to follow tortuous paths. This is combination with the previously mentioned fact of holes in the vertical plane being arranged in-line results in the structure providing a greater flow resistance in along the layers than in the perpendicular direction.
To arrange the opposite effect one could simple omit holes. The rear part of the assembly shown in
A circular arrangement according to this embodiment could be adopted as a substitute for the previously described matrix structure, that is, to help distribute flow in a radial, outward direction. As with the previously described grid-like structure, such an effect can serve a triple purpose: firstly, to even out flow for the benefit of a downstream honeycomb catalyser, secondly, to allow for a more widening diffuser without major flow separation, and thirdly, to add catalyser capacity within a space that would otherwise be wasted from the point of view of cramming as much catalyser surface as possible into a limited space available.
A matrix structure according to
| Number | Date | Country | Kind |
|---|---|---|---|
| PA 2004 01640 | Oct 2004 | DK | national |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/DK05/00690 | 10/26/2005 | WO | 7/26/2007 |
