APPARATUS AND PROCESS FOR THE PRODUCTION OF OLEFINS

The present invention relates to an apparatus, in particular, to an apparatus suitable for the production of olefins by auto-thermal cracking. Autothermal cracking is a route to olefins in which a hydrocarbon feed is mixed with oxygen and passed over an autothermal cracking catalyst. The autothermal cracking catalyst is capable of supporting combustion beyond the fuel rich limit of flammability. Combustion is initiated on the catalyst surface and the heat required to raise the reactants to the process temperature and to carry out the endothermic cracking process is generated in situ. Generally the reactant stream is passed over a supported catalyst to produce the olefin product. Typically, the catalyst comprises a Group VIII metal, preferably at least one platinum group metal, for example, platinum. The autothermal cracking process is described in EP 332289B; EP-529793B; EP-A-0709446 and WO 00/14035.

The product stream typically exits the reaction zone as a gaseous product stream at a temperature greater than 800°C, for example greater than 900 ⁰ C, and it is desired that the product stream is rapidly cooled. This ensures a high olefmie yield because the product cooling step slows down the rate of reaction in the gaseous product stream thus minimising further reactions taking place to form undesired products.

To maximise efficiency of the overall process, downstream of the quench section the quenched product stream is further cooled and heat recovered by a waste heat boiler.

Waste heat boilers, which are essentially heat exchangers which act to remove heat from the product stream by heat exchange with water and generation of steam, are well- known in the art. Typical waste heat boilers may have a "shell and tube" structure including a plurality of tubes in an outer shell and wherein one fluid passes through the tubes ("tube-side") and the other is contacted with the outside of the tubes ("shell-side"). The stream to be cooled may be passed through the tubes, is what is generally known as a tube-side waste heat boiler, as described, for example US 4,288,408, or may have a more complicated flow path through the shell, in what is generally known as a shell-side waste- heat boiler, such as described in US 5,653,282. Whilst the more complicated flow paths such as described in US 5,653,282 can provide long contact times, such structures are not suitable where relatively low residence times are required, for example where the streams to be cooled are prone to cause coking.

Tube-side waste heat boilers are therefore required where it is desired to have relatively low residence times of the stream to be cooled.

It has now been found that the residence time distribution in a tube-side waste heat boiler provided downstream of a product quench zone may be narrowed, the maximum residence time in the waste heat boiler may be reduced, and the efficiency of heat recovery may be improved, by selection of a specific configuration of the tube-side waste heat boiler.

Thus, in a first aspect, the present invention provides an apparatus for reacting a first gaseous reactant stream with a second gaseous reactant stream to form a gaseous product stream, wherein the apparatus comprises at least one first supply means for the first gaseous reactant stream, at least one second supply means for the second gaseous reactant stream, a catalyst zone, a product quench zone and a tube-side waste heat boiler, wherein the product quench zone is positioned downstream of the catalyst zone and comprises a plurality of quench tubes, each of which have an inlet end at the end to which the stream to be quenched is introduced and an exit at the other end, and the longitudinal axes of which are orientated in the direction of flow of the stream to be quenched and wherein the waste heat boiler is positioned downstream of the exits of the quench tubes and connected to the product quench zone without any intermediate zones, and comprises a plurality of boiler tubes, each of which have an inlet end at the end to which the stream to be cooled is introduced and an exit at the other end, and the longitudinal axes of which are orientated in the direction of flow of the stream to be cooled and further wherein the waste heat boiler has a surface on which stream exiting the exits of the quench tubes impinges prior to entering the waste heat boiler tubes. The waste heat boiler of the present invention preferably comprises a plurality of tubes orientated in a parallel direction to each other. Typically, the waste heat boiler comprises at least 10 tubes, for example at least 20 tubes and/or up to 150 tubes.

In the present invention, a surface is provided in the waste heat boiler "below" the exits of the quench tubes ("below" here is defined by reference to the direction of flow of the quenched stream rather than to indicate any overall orientation of the apparatus). The quenched product stream which exits the plurality of quench tubes impinges on the surface

prior to passing into the tubes of the waste heat boiler, which surface thus distributes the quenched product stream across the tubes of the waste heat boiler.

In order that the stream exiting the exits of the quench tubes can impinge on the surface provided in the waste heat boiler, the waste heat boiler must be connected to the product quench zone without any intermediate zones that impede the flow of the stream, typically by connecting the waste heat boiler directly to the product quench zone. The waste heat boiler and product quench zone are preferably connected such that the surface is provided in relatively close proximity to the exits of the quench tubes, typically less than 100mm from the exits of the quench tubes, such as less than 50mm. The surface is preferably provided in a substantially perpendicular orientation relative to the direction of flow of the stream exiting the quench tubes (which is usually also a substantially perpendicular orientation relative to the longitudinal axes of the quench tubes), by which is meant that the plane defined by the edges of the surface has an orientation within 20° of said perpendicular. (Where the surface is essentially flat this plane is co-planar with the surface itself. Where the surface is conical, for example, this plan© relates to the base of the cone.)

In a most preferred orientation, the entrances to a plurality of the boiler tubes of the waste heat boiler are provided in a plane with a substantially perpendicular orientation relative to the direction of flow of the stream exiting the quench tubes (which is usually also a substantially perpendicular orientation relative to the longitudinal axes of the quench tubes), by which is meant that the plane has an orientation within 20° of said perpendicular.

Most preferably, the exits of the plurality of the quench tubes form a single plane and said exits, the surface and the entrances to a plurality of the boiler tubes are provided in three essentially parallel planes.

Similarly, the longitudinal axes of the quench tubes and the longitudinal axes of the waste heat boiler tubes are most preferably in parallel alignment.

In the absence of the surface according to the present invention, the majority of the quenched product stream flows directly down any boiler tubes immediately below the exits of the quench tubes, and whilst this flow has a very short residence time in this section, these central tubes are not efficient enough to effect the required cooling. In addition, the rest of the flow has been found to have a significantly larger residence time in the

remainder of the waste heat boiler cross-section, which can result in the formation of significant quantities of coke in some of the boiler tubes.

The surface below the quench tubes has been found to act to even the distribution of the quenched product stream across the boiler tubes, and it has been surprisingly found that although this increases the minimum residence time of some of the stream it results in a narrowing of the residence time distribution and a decrease of the maximum residence time. As well as reducing the overall formation of coke, this improved flow distribution also results in improved efficiency of heat recovery in the waste heat boiler because the heat load is more evenly distributed. The surface may be provided in any suitable manner. The surface must have a cross-sectional area of at least the area bounded by the quench tube exits such that the majority of the stream exiting the quench tubes impinges thereon. Suitably, the surface has an area at least equal to, such as equal to, the cross-sectional area of the upstream catalyst zone. Typically, the cross-sectional area of the area bounded by the plurality of boiler tubes of the waste heat boiler (hereinafter referred to as the waste heat boiler tubes cross- sectional area) is greater than the cross-sectional area of the area bounded by the exits of the quench tubes (hereinafter referred to as the quench cross-sectional area).

The cross-sectional area of the waste heat boiler tubes is a result of the competing design requirements of the waste heat boiler. Typically a tube-side waste heat boiler will be designed to have the following characteristics.

• A high superficial velocity in the heat exchanger tubes to minimise the residence time of stream in the tube, reducing the formation of coke.

• An acceptably low pressure loss across the tubes. • Sufficient heat transfer area to accommodate the required duty.

Typically a design which satisfies the above requirements will result in an overall increase in the cross-sectional area of the waste heat boiler tubes compared with the quench cross-sectional area and the upstream catalyst zone cross-sectional area.

Preferably, the waste heat boiler tubes cross-sectional area is at least 50% greater than the quench cross-sectional area, especially at least two times, such as four times the quench cross-sectional area.

Most preferably, the waste heat boiler tubes cross-sectional area is at least two times, such as four times, the cross-sectional area of the upstream catalyst zone.

In one example, the waste heat boiler itself has a solid surface with no tubes in its central section, the area of this central section being at least equal to quench cross-sectional area, and the boiler tubes being provided surrounding this central area.

Alternatively, and preferably, the waste heat boiler may have a more "conventional" distribution of tubes across its entire cross-section, but the quenched product stream is prevented from passing directly through the boiler tubes immediately below the exits of the quench tubes by the provision of a surface above the boiler tubes, for example in the form of a baffle plate, which distributes the quenched product stream. Such a surface may be any suitable shape, for example, a flat upper surface and a cone shaped lower surface to encourage flow of the quenched product stream to the boiler tubes below said surface.

In this embodiment, the surface may be solid or may be provided with spaces or holes through which the quenched product stream may flow, with the proviso that any spaces or holes therein are not located directly below the exits of any of the quench tubes.

The waste heat boiler may have all boiler tubes of the same dimensions, but preferably comprises boiler tubes with at least two different diameters, to provide what is known as "cold bypass" and allow the waste heat boiler to efficiently handle varying heat loads (for example, if reaction throughput is varied).

In use, the first and second gaseous reactant streams are preferably mixed and preheated immediately before contact with a catalyst in the catalyst zone. Thus, the apparatus preferably comprises a mixing and pre-heating section upstream of the catalyst zone. Any suitable mixing and pre-heating means may be used, but most preferably, the apparatus comprises a mixing and pre-heating section which utilises first and second supply means for the respective reactants each comprising a plurality of outlets, as described in WO 2004/074222, the contents of which are incorporated herein by reference.

Thus, the apparatus preferably comprises at least one first supply means for the first gaseous reactant stream, at least one second supply means for the second gaseous reactant stream, a resistance zone and a catalyst zone,

wherein the first supply means comprises a plurality of first outlets for delivery of the first gaseous reactant stream and the second supply means comprises a plurality of second outlets for delivery of the second gaseous reactant stream, the resistance zone is porous, is positioned downstream of the first and second supply means with respect to the flow of the first and second gaseous reactant streams and is in fluid communication with the first and second supply means, the catalyst zone is positioned downstream of the resistance zone with respect to the flow of the first and second gaseous reactant streams and is in fluid communication with the resistance zone, and wherein the first supply means and the second supply means are arranged such that the first and second gaseous reactant streams are contacted in an essentially parallel manner and mixed prior to contacting the resistance zone.

The plurality of outlets of the mixing device are preferably provided in a regular pattern, such as described in WO 2004/074222, such that the plurality of outlets for the first gaseous reactant stream are equidistant from a plurality of neighbouring outlets for the second gaseous reactant stream and vice versa. This leads to the most efficient mixing and hence the lowest residence time upstream of the catalyst. Preferred configurations to achieve this are hexagonal (where one outlet has 6 nearest neighbours equally spaced from it in a regular hexagon configuration) or square (where one outlet has 4 nearest neighbours equally spaced from it in a square configuration).

The resistance zone is porous and ensures dispersion of the reactants as they pass through the zone, such that they leave the resistance zone substantially uniformly distributed over the cross-sectional area of the resistance zone, and hence substantially uniformly distributed over the cross-sectional area of the subsequent catalyst zone. The resistance zone may be formed of a porous metal structure, but preferably the porous material is a non metal e.g. a ceramic material. Suitable ceramic materials include lithium aluminium silicate (LAS), alumina (Al ₂ O ₃ ), stabilised zirconias, alumina titanate, niascon, cordierite, mullite, silica and calcium zirconyl phosphate. The porous material may be in the form of spheres or other granular shapes. Alternatively, the porous material may be in the form of a foam.

The catalyst zone usually comprises a catalyst bed held in place in the reaction zone in a suitable holder, such as a catalyst basket.

The cross-section of the catalyst zone (CA) is usually at least 0.05 m ² , preferably at least 0.1 m ² .

The cross-section of the catalyst zone is usually less than 1.2 m ² , preferably less than 0.5 m ² . Most preferably, the cross-section of the catalyst zone is in the range 0.2 to 0.3 m ² .

In use the catalyst zone comprises a catalyst. The depth of the catalyst is preferably 0.02 to 0.1 m.

In the product quench zone, the product stream is rapidly cooled, preferably by injecting a quenchant into the product stream in the quench tubes. The quenchant may be a gas or a liquid. The quenchant may be an inert quenchant or may be a reactive quenchant, for example, a hydrocarbon, especially an alkane or mixture of alkanes which could crack to produce olefin. When the quenchant is gas it is preferably an inert gas. Preferably the quenchant is a liquid e.g. water.

The quenchant, such as water, is usually injected at a pressure higher than the pressure of the gaseous product stream, such as 100 barg, and is usually injected at a temperature of between 100-400 ^s C and preferably between 200-350 ⁰ C e.g. 300 ^D C. Injecting the quenchant at high pressure and high temperature ensures that a large proportion of the quenchant instantaneously vaporizes at the reactor pressure and therefore provides a very rapid temperature drop in the gaseous product stream. The product stream is quenched on exiting the catalyst to a temperature of 800 ⁰ C or less, preferably within 40ms and advantageously within 20ms from exiting the catalyst. The product stream may be quenched on exiting the catalyst such that the temperature of the product stream is reduced to between 700 ⁰ C and 800 ⁰ C, or may be quenched to lower temperature, for example 600 ⁰ C or less (again preferably within 40ms, and advantageously 20ms from exiting the catalyst) to minimise further reactions.

The number and dimensions of the quench tubes are preferably selected to minimise the time to quench the product stream exiting the catalyst zone.

Preferably, the product quench zone comprises a plurality, N, of quench tubes, each tube having a length, L, a diameter, D, and a cross-sectional area, QA, each quench tube having at least one quenchant inlet per tube which inlet passes quenchant into the tube from the side of said tube, and wherein, D is between 0.04 and 0.10 m,

L/D is at least 5, preferably at least 10, and

(N x QA) / CA is between 0.07 and 0.31, wherein CA is the cross-section of the catalyst zone (CA).

By "passes quenchant into the tube from the side of said tube" is meant that the quenchant is introduced at an angle, suitably at least 45°, and preferably 90°, compared to the longitudinal axis of the quench tube. This provides better mixing, and hence more rapid cooling, than a quenchant inlet injecting quenchant along the axis of the quench tube i.e. in the general direction of flow through the tube.

Preferably, two to four quenchant inlets are provided per quench tube, suitably spaced approximately equidistantly around the quench tube.

Each quenchant inlet may comprise a single nozzle or a number of nozzles, for example 2 to 7 nozzles. Typically, they will be close packed to minimise the size of the inlet nozzle arrangement.

The apparatus comprises a plurality of quench tubes, N. Thus, N is at least 2. The optimum number of tubes depends on the total cross-section of the catalyst zone, and, in general, increases with an increase in the cross=sectional area of the catalyst zone. Thus, the ratio (N x QA) / CA gives the ratio of the total cross-sectional area of the N quench tubes to the cross-sectional area of the catalyst zone. (The cross-sectional area of each tube (QA) is proportional to it diameter, D, according to the equation QA = (0.5 x D) ² x π.) Preferably, (N x QA) / CA is less than 0.25, for example in the range 0.08 to 0.25.

More preferably, (N x QA) / CA is at least 0.1, and, most preferably, in the range 0.1 to 0.2.

Usually N is at least 3. Preferably, N is less than 20, more preferably less than 10.

Preferably D is at least 0.06m and/or less than or equal to 0.085m. Preferably L/D is less than 15, for example from 10 to 15.

In use, each quench tube may be defined by an inlet end, at the end to which the stream to be quenched is introduced and an outlet end at the other end. The quenchant inlet (at least the first nozzle where more than one nozzle is provided per inlet) is generally provided in the portion of each quench tube closest to the inlet end, so that the quenchant can be contacted with the stream to be cooled as quickly as possible after it enters the quench tube.

At the tube diameters of the preferred quench tubes, the L/D of at least 5 has been found to ensure good mixing of the quenchant and gaseous product stream within the length of said tube, which ensures rapid cooling. In contrast, at larger diameters even significantly longer quench tubes may not provide the required quenching, and certainly not in as short a time-scale.

Preferably, the apparatus is designed to operate at elevated pressure, for example at a pressure of greater than 0.5 barg, preferably at a pressure of least 10 barg, and more preferably at a pressure of at least 15 barg. The pressure is preferably less than 50 barg, and more preferably less than 35 barg, for example in the range 15 to 30 barg. The apparatus of the present invention is suitable for autothermal cracking processes or for other processes in which it is desired to rapidly cool the gaseous product stream formed by reacting a first gaseous reactant stream with a second gaseous reactant stream over a catalyst.

The apparatus is advantageously employed to partially oxidize a gaseous feedstock. Thus, the present invention also provides a process in which a first gaseous reactant stream and a second gaseous reactant stream are contacted with a catalyst to produce a gaseous product stream, said process being performed in an apparatus as described herein. Preferably the first gaseous reactant stream comprises an oxygen containing gas and the second gaseous reactant stream comprises a paraffinic hydrocarbon. Most preferably, the oxygen containing gas and the paraffinic hydrocarbon are contacted with a catalyst capable of supporting combustion beyond the normal fuel rich limit wherein autothermal cracking occurs to produce one or more olefins.

Thus, in a particular embodiment of the process of the present invention, a paraffinic hydrocarbon and a molecular oxygen containing gas are contacted with a catalyst capable of supporting combustion beyond the normal fuel rich limit of flammability to produce a gaseous product stream comprising olefins, said process being performed in an apparatus as described herein.

On contacting of the paraffinic hydrocarbon and molecular oxygen containing gas with a catalyst capable of supporting combustion beyond the normal fuel rich limit of flammability, combustion of the paraffinic hydrocarbon is initiated on the catalyst and the heat required to raise the reactants to the process temperature and to carry out the endothermic cracking process to produce olefins is generated in situ.

The catalyst for autothermal cracking may be unsupported, such as in the form of a metal gauze, but is preferably supported. Any suitable support material may be used, such as ceramic or metal supports, but ceramic supports are generally preferred. Where ceramic supports are used, the composition of the ceramic support may be any oxide or combination of oxides that is stable at high temperatures of, for example, between 600 ⁰ C and 1200 ⁰ C. The support material preferably has a low thermal expansion co-efficient, and is resistant to phase separation at high temperatures.

Suitable ceramic supports include cordierite, mullite, lithium aluminium silicate (LAS), alumina (e.g. α-Al ₂ C> ₃ ), stabilised zirconias, alumina titanate, niascon, and calcium zirconyl phosphate. The ceramic supports may be wash-coated, for example, with 7-Al ₂ O ₃ . The support is preferably in the form of a foam or a honeycomb monolith. Preferred hydrocarbons for autothermal cracking are paraffmic hydrocarbons having at least 2 carbon atoms. For example, the hydrocarbon may be a gaseous hydrocarbon, such as ethane, propane or butane or a liquid hydrocarbon, such as a naphtha or an FT liquid. Where a liquid hydrocarbon is to be reacted it should be vaporised to form a gaseous reactant stream for use in the present invention.

The oxygen containing gas may be provided as any suitable molecular oxygen containing gas, such as molecular oxygen itself or air.

Preferably, hydrogen is co-fed to the autothermal cracking reaction. Hydrogen co- feeds are advantageous because, in the presence of the autothermal cracking catalyst, the hydrogen combusts preferentially relative to hydrocarbon, thereby increasing the olefin selectivity of the overall process. The amount of hydrogen combusted may be used to control the amount of heat generated and hence the severity of cracking. Thus, the molar ratio of hydrogen to oxygen can vary over any operable range provided that the autothermal cracking product stream comprising olefins is produced. Suitably, the molar ratio of hydrogen to oxygen is in the range 0.2 to 4, preferably, in the range 0.2 to 3.

The hydrocarbon and oxygen-containing gas may be contacted with the catalyst in any suitable molar ratio, provided that the autothermal cracking product stream comprising olefins is produced. The preferred stoichiometric ratio of hydrocarbon to oxygen is 5 to 16, preferably, 5 to 13.5 times, preferably, 6 to 10 times the stoichiometric ratio of hydrocarbon to oxygen required for complete combustion of the hydrocarbon to carbon dioxide and water.

Preferably, the reactants are passed over the catalyst at a pressure dependent gas hourly space velocity of greater than 20,000 h ^'1 barg ^'1 and, most preferably, greater than 100,000 h ^'1 barg ^"1 . For example, at 20 barg pressure, the gas hourly space velocity is most preferably, greater than 2,000,000 h ^"1 . It will be understood, however, that the optimum gas hourly space velocity will depend upon the nature of the feed composition.

The autothennal cracking step may suitably be carried out at a catalyst exit temperature in the range 600 ⁰ C to 1200°C. Suitably the catalyst exit temperature is at least 720 ⁰ C such as at least 750 ⁰ C. Preferably, the autothermal cracking step is carried out at a catalyst exit temperature in the range 850 ⁰ C to 1050 ⁰ C and, most preferably, in the range 850 ⁰ C tO lOOO ⁰ C.

The autothennal cracking step is usually operated at a pressure of greater than 0.5barg, preferably at a pressure of least 10 barg, and more preferably at a pressure of at least 15 barg. The pressure is preferably less than 50 barg, and more preferably less than 35 barg, for example in the range 20 to 30 barg. The catalyst for autothermal cracking is capable of supporting combustion beyond the fuel rich limit of flammability. The catalyst usually comprises a Group VIII metal as its catalytic component. Suitable Group VIII metals include platinum, palladium, ruthenium, rhodium, osmium and iridium. Rhodium, and more particularly, platinum and palladium are preferred. Typical Group VIII metal loadings range from 0.01 to 100 wt%, preferably, between 0.01 to 20 wt%, and more preferably, from 0.01 to 10 wt% based on the total dry weight of the catalyst.

Where a Group VIII catalyst is employed, it may be employed in combination with a catalyst promoter. The promoter may be a Group IIIA, IVA, and/or VA metal. Alternatively, the promoter may be a transition metal; the transition metal promoter being a different metal to that which may be employed as the Group VIII transition metal catalytic component. Preferred promoters are selected from the group consisting of Ga, In, Sn, Ge, Ag, Au or Cu. The atomic ratio of Group VIII metal to the catalyst promoter may be 1 : 0.1 - 50.0, preferably, 1: 0.1 - 12.0.

For the avoidance of doubt, the Group VIII metal and promoter in the catalyst may be present in any form, for example, as a metal, or in the form of a metal compound, such as an oxide.

The catalyst may be prepared by any method known in the art. For example, gel methods and wet-impregnation techniques may be employed. Typically, the support is impregnated with one or more solutions comprising the metals, dried and then calcined in air. The support may be impregnated in one or more steps. Preferably, multiple impregnation steps are employed. The support is preferably dried and calcined between each impregnation, and then subjected to a final calcination, preferably, in air. The calcined support may then be reduced, for example, by heat treatment in a hydrogen atmosphere.

Although the catalyst has been described above in terms of a single catalyst, the catalyst may alternatively be present as a sequential catalyst bed, as described, for example, in WO 02/04389.

The gaseous product stream, in addition to olefins, will generally comprise unreacted paraffinic hydrocarbons, hydrogen, carbon monoxide and methane, and may comprise water, and small amounts of acetylenes, aromatics and carbon dioxide, which need to be separated from the desired olefins. The required separations on said stream, after quenching and cooling, may be performed by any suitable techniques, such as an amine wash to remove carbon dioxide and any water, a demethaniser, to separate hydrogen, carbon monoxide and methane, a deethaniser, to separate C3+ hydrocarbons from ethane and ethylene, and a C2 splitter to separate ethylene from ethane. The present invention is particularly useful for apparatus and processes at a commercial scale. "Commercial scale" will depend on the process itself, but the reactor/catalyst bed will typically be sized to process at least 50 ktpa of hydrocarbon (per reactor where more than one reactor is present), preferably at least 100 ktpa of hydrocarbon (per reactor). For example, for the production of olefins in an autothermal cracking process, a commercial scale is typically sized to produce at least 25 ktpa of olefins (per reactor), preferably at least 75 ktpa of olefins (per reactor).

The invention will now be illustrated by way of Figures 1 and 2, and the examples, wherein: Figures 1 and 2 show in schematic form sections through the quench zones and waste heat boilers respectively of two sets of apparatus according to the present invention.

In particular, with respect to Figure 1 there is shown a quench zone (1), with a plurality of quench tubes (2), each with an exit (3) at the base of the quench zone. The diameter, d, corresponds to the diameter of the upstream catalyst zone (not shown).

Downstream of the quench zone (1) there is provided a waste heat boiler (4), which comprises a plurality of boiler tubes (5a, 5b) of at least two different dimensions. The central section of the section of the waste heat boiler comprising the boiler tubes has no tubes therein, and hence provides a solid surface (6), below the exits (3) of the quench tubes (2). In use, quenched product stream exits the quench tubes (2), impinges on this surface (6) and is distributed across the area of the boiler tubes (5a, 5b) below. With respect to Figure 2 there is shown a quench zone (1), with a plurality of quench tubes (2), each with an exit (3) at the base of the quench zone. The diameter, d, corresponds to the diameter of the upstream catalyst zone (not shown).

Downstream of the quench zone (1) there is provided a waste heat boiler (4), which comprises a plurality of boiler tubes (5a, 5b) of at least two different dimensions. The diameter, D, is approximately twice the diameter of the catalyst zone, d, and thus the cross- sectional area of the section of the waste heat boiler containing the plurality of boiler tubes is four times the cross-sectional area of the upstream catalyst zone.

Immediately below the exits (3) of the quench tubes (2) there is provided a solid surface (6), in the shape of a cone. Although not shown, this may be attached to the base of the quench zone or to the top of the boiler tubes below. In use, quenched product stream exits the quench tubes (2), impinges on this surface (6) and is distributed across the area of the boiler tubes (5a, 5b) below. Examples

Computational fluid dynamics (CFD) has been used to model the residence time distributions of a number of configurations of waste heat boiler.

As a Comparative Example a waste heat boiler with a diameter of 500mm is provided below the exits of three quench tubes, with a pitch of 107mm (the area bounded by these tubes is a triangle of approximately 50 cm ² ). The waste heat boiler comprises a chamber of height 100mm separating the exits of the quench tubes from the entrances of the waste heat boiler tubes. No surface according to the invention is provided. The product exits the quench tubes at a velocity of approximately 75m/s, which reduces immediately on entering the chamber of the waste heat boiler.

Example 1 is comparable to the Comparative Example, except that the central section of the waste heat boiler is a solid surface (with no tubes in or below), 100mm below the exits of the quench tubes and with a diameter of 250mm. The diameter of the waste heat boiler is increased to 560mm, to give an equivalent area of quench tubes.

Example 2 is the same as Example 1, except that the chamber of the waste heat boiler is reduced to 50mm in height (hence also reducing the distance from the exits of the quench tubes to the surface).

The results are given in Table 1. Table 1

The Comparative Example shows a reasonable mean residence time in the chamber, but this is due to the fact that a significant amount of the flow proceeds directly down the central tubes of the waste heat boiler, and the variation in flow through the tubes is relatively high.

Example 1 shows a slightly increased mean residence time, but the maximum residence time is approximately the same as in the Comparative Example and the variability in flow rates in the tubes is effectively removed (variability less than +/- O.lm/s)

Example 2 shows that the surface may be closer to the exits of the quench tubes without affecting the flow rate distribution (again variability than less than +/- 0. lm/s), whilst having the further advantages of reducing the mean and maximum residence times in the waste heat boiler.

Examples 3 and 4 present results from a larger scale waste heat boiler. In these Examples the product gas exits an upstream catalyst zone and is quenched in a series of seven quench tubes (in an overall hexagonal shape with six outside tubes and one tube in

the middle) with a pitch (distance between adjacent tubes) of approximately 200mm (the area bounded by these tubes is a hexagon of approximately 1040 cm ² ).

The product exits the quench tubes at a velocity of approximately 75m/s, which reduces immediately on entering the chamber of the waste heat boiler. In Example 3, according to Figure 1, the central section of the cross-section of waste-heat boiler is a surface of diameter 600mm (shown as 6 in Figure 1), provided 50mm below the exits of the quench tubes, and the overall diameter of the waste heat boiler, D, is 1410mm.

In the Example 4, according to Figure 2, a surface in the shape of an inverted cone (6) with diameter of 600mm is provided 50mm below the exits of the quench tubes, with a waste heat boiler of overall diameter 1280mm and tubes across its entire cross-section provided underneath (total area of boiler tubes is the same as Example 3, the entrances to boiler tubes are 100 mm in total below the exits of the quench tubes).

Table 2 shows the mean and maximum residence times of the gas exiting the quench tubes in the chamber for each example.

Table 2

Although at larger scale, Example 3 is comparable to Example 2, with slightly lower mean and maximum residence times due to the flared configuration of the waste heat boiler at the exits of the quench tubes compared to the cylindrical chamber of Examples 1 and 2.

As with Example 2, both Examples 3 and 4 provide an even distribution of flow rates at the entrances to the boiler tubes, whilst having the advantages of reducing the mean and maximum residence times in the waste heat boiler compared to configurations without a surface. Although Example 4 shows slightly higher mean and maximum residence times than Example 3, Example 4 has a larger height chamber above the waste heat boiler tubes

and the advantage, compared to Example 3, of using a waste heat boiler of overall smaller diameter.

These results show that providing a surface according to the process of the present invention, and in particular the configurations of Figures 1 and 2, can provide a more even flow distribution across the boiler tubes of the waste heat boiler, with the minimum flow rate in to the waste heat boiler tubes increased and the maximum flow rate reduced. This results in more even heat load across the waste heat boiler and improved efficiency of heat recovery.

In addition, the present invention can provide a narrowed residence time distribution in the chamber, with reduced mean and maximum residence times in the chamber.