1. Introduction
Distillation operations are carried out in columns, that is, cylindrical pressure vessels with large height to diameter ratios, containing various kinds of internals chosen and arranged appropriately to enable cost-effective separation of binary or multicomponent, mainly liquid, feed mixtures into desired pure products or specified fractions.
As mentioned by Stichlmair and Fair in the preface of Distillationd-Principles and Practice, which focuses on understanding of the principles and their translation into basic design practices, distillation columns are the workhorses of the petroleum, petrochemical, chemical, and related process industries [1]. A complementary, more pragmatic view is given by Kister in Distillation Design, with emphasis on the practical aspects and reliability as encountered and required during the design of distillation columns [2].
The outgoing point for design of a new distillation column is the top and/or bottom product specification (required purity and/or recovery) for a given feed mixture, with known composition, flow rate, and thermal state at a given pressure and temperature. Upon choosing the operating pressure and assuming column pressure drop, detailed column performance calculations deliver optimum stage and reflux requirements, all based on appropriate vapore-liquid equilibrium data. The next, column dimensioning step is concerned with determining column diameter(s) and height for the chosen type of vapor-liquid contacting device. Critical positions are the top and bottom stage of rectification and stripping sections, and the column shell diameter is usually based on the stage with the maximum vapor load; and to complete the design calculations properly, one or more iterations on pressure drop are required.
Upon completing stage and reflux requirement calculations, the internal flow rates of vapor and liquid as well as corresponding values of relevant physical properties (densities and viscosities of two phases and surface tension) are known per stage and serve as the basis for the determination of column diameter(s). However, the number of theoretical plates or equilibrium stages, which is by definition the number of times the mass transfer equilibrium has to be established between ascending vapor and descending liquid streams above and below the feed to achieve desired separation, cannot be directly translated into column height. Namely, along an operating distillation column the equilibrium between vapor and liquid is never fully established. The column height required to produce the degree of mass transfer equivalent to the given number of equilibrium stages will depend on the type and performance characteristics of equipment (internals) chosen to generate interfacial area, that is, to enable mass transfer to occur so that for each component the required number of moles is transported accordingly, from vapor to liquid and reverse within the contact time available.
1.1 Distillation column anatomy (internal structure)
Regarding the equipment used to promote intimate contact of ascending vapor and descending liquid in a distillation column or tower, a distinction is generally made between plate or tray columns and packed columns, and in the case of the latter, between randomly and orderly packed columns, that is, random and structured packings.
Figure 1 shows schematically a distillation column shell containing all relevant equipment as employed in packed and/or trayed columns. From the top to the bottom: (1) a demister, i.e. a device for removal of entrained droplets from the vapor leaving the column through the nozzle at the top of the column; a liquid distributor (2) irrigating a randomly packed bed (3) laying on a V-shaped support plate (4). One should note that normally a hold-down plate (hard to see here) is used to prevent bed expansion in an upward direction due to pressure upsets during operation.
The liquid leaving the randomly packed bed is collected in a chevron (vane)-type liquid collector (5), where it is mixed with the liquid feed entering through a nozzle at the side of the column. The mixed liquid enters via a down pipe the liquid distributor (2), irrigating a bed consisting of two different sizes of a corrugated sheet structured packing (6), supported by a support plate (4). The liquid leaving the structured packing bed is collected using a liquid collector (7) with holes in the bottom. This or similar so-called chimney tray liquid collectors (9) are often used in combination with the packed column vapor inlet device (8) shown here, to ensure a good initial distribution of vapor entering the packed bed above. This particular device, known as “schoepentoeter”, handles with ease two-phase (biphasic) feeds, as delivered by falling film and thermosyphon type reboilers.
In the present case, a chimney tray collector (7) collects and mixes the liquid leaving the packed bed above, and the another one (9) collects the liquid coming from the vapor inlet device (schoepentoeter) (8), and delivers it via a central downcomer to the top tray of a section containing three fixed-valve, two-pass trays (10). The liquid leaving the last tray falls down and is collected in the sump of the column. Above the liquid level there is a vapor inlet nozzle. Usually, no vapor inlet devices are used in tray columns, because the bottom tray generates a large enough pressure drop to ensure a good initial distribution of vapor.
In addition to two vapor and two liquid feed lines, Figure 1 shows also two manholes, that is, entrances to the inside of the column for installation during construction and inspection during shutdown. In the case of packed columns, as indicated here, manholes are always placed at the level of liquid distributors. The liquid distributor (2) shown schematically in Figure 1 is a narrow trough type with baffles on both sides that receive and spread the liquid jets coming from the equidistantly arranged orifices in the side walls of the troughs. Detailed drawings and photographs and related performance characteristics of devices shown schematically in Figure 1 can be found together with all other devices belonging to Sulzer-Chemtech separations column portfolio on the website (www.sulzer.com). Similar information and company-specific or proprietary designs can be found on the websites of other well-established distillation equipment manufacturers, i.e. Koch-Glitsch (www.koch-glitsch.com), Julius Montz (www.montz.de), and Raschig (www.raschig.com).
 |
| FIGURE 1 A Computer-Aided Drawing Indicating the Main Types of Vapore-Liquid Contacting Devices and Auxiliary Equipment Encountered in Distillation |
As indicated above, the trays as well as random packings and structured packings with auxiliary equipment represent three essential types of vapor-liquid contacting devices used in distillation operations. Typically distillation columns are equipped either with trays, random packings, or structured packings, but if appropriate, a distillation column can contain each combination of these or even all three of them.
1.2 Methods and main features of distillation equipment operation
In a distillation column, the liquid enters at the top, as reflux, and somewhere along the column as feed, and descends driven by gravity, experiencing more or less pronounced resistance caused by ascending vapor. The main purpose of distillation equipment is to establish an intimate contact between ascending vapor and descending liquid, by creating a large, intensively refreshed interfacial area. The flow pattern and relative orientation of phases during contact depend on the type of the vapor-liquid contactor.
In a tray column the ascending vapor and descending liquid are contacted stagewise, that is, by repeatedly contacting and disengaging two countercurrently flowing phases over and over again under cross-current conditions. Namely, the liquid coming from inlet downcomer flows horizontally across perforated plates (tray decks) and pours over the outlet weir into the outlet downcomer, moving to the tray below, while the vapor flowing upwardly is forced to pass through perforations in the tray deck and a shallow pool of moving liquid a number of times. The interface, being the surface area of bubbles and/or droplets, is created to an extent depending mainly on operating conditions, that is, the governing flow regime and surface tension in conjunction with chosen tray design. Considering the fact that vapor and liquid brought into an intimate contact on a tray usually leave the tray before reaching equilibrium, the number of actual trays should normally exceed the number of equilibrium stages or theoretical plates required to achieve specified separation at given reflux ratio. Therefore, it is a common practice to consider performance of an actual stage or plate or tray in the terms of the deviation from ideal, e.g. to define the “overall tray efficiency” per column section as the ratio of the number of equilibrium stages to the number of actual stages (trays or plates) contained in that section. The resulting column height will depend on the tray spacing chosen, which can differ per section if appropriate. Namely, feed point or stage separates a conventional distillation column into an upper (rectification) section and a lower (stripping) section, which, depending on thermal conditions and/or the nature of feed mixture, can differ significantly in operating conditions and consequently in performance of the internals involved.
In a packed column, mass transfer occurs continuously by countercurrent flow of the liquid and vapor, and the efficiency is commonly expressed as the bed height generating composition change equivalent to that of a theoretical plate (HETP). If the HETP is known, then packed bed height required is simply the product of HETP and the number of theoretical plates (equilibrium stages) required for separation at given conditions.
Nevertheless this is not as simple as it appears. Namely, packing serves as carrier for liquid, and the surface of the liquid film provides interfacial area for mass transfer. Complete wetting is a prerequisite for successful operation. To assure this, a well-performing liquid distributor is required to provide adequate quality of initial liquid distribution, while the quality of liquid distribution within the packed bed will depend on type and size of packing, layout of a layer consisting of segments, operating conditions, and surface tension.
However, due to the efficiency-deteriorating effect of liquid maldistribution, which tends to develop with increasing bed depth, the height of a packed bed is usually limited to a height equivalent to a certain number of theoretical plates, 15 being a typical number. The liquid leaving a bed of given height needs to be collected and thoroughly mixed before being delivered to the distributor of the bed below. The column height required to install a liquid redistribution section is up to 2 m, which means that in case of multistage separations ancillary equipment will contribute significantly to total column height and cost. This is in practice considered to be a good reason to avoid a packed column in a new design, while the manufacturers have often been forced in revamp situations to install taller than usual beds, to ensure that liquid redistribution sections can be accessed through existing manholes. It is well known that this was done with success in numerous cases and that in industrial columns equipped with standard-size packings (220-250 m2 /m3 ) there are wellperforming beds of exceptional height (more than 20 stages and in excess of 10 m). With this, the outgoing assumption that “maximum bed height” extended by 2 to 3 m can add, say, three to four stages as well as the height necessary to compensate for anticipated loss in efficiency, was justified. However, this should not be done at the cost of mechanical integrity of the bed; it will depend on the type and size of packing as well as the construction material.
The extent of liquid maldistribution and consequently the allowable bed height will depend on the structure of the packed bed, which may be randomly or orderly packed. In other words, if considering a packed column, a clear distinction is made between random and structured packings. The former, often called “dumped packings”, are small, open bodies of various shapes and sizes that are dumped into the column (dry or wet, depending on situation) and settle randomly. Their liquid spreading extent is rather limited, that is, it does not exceed the dimension of the body. The structured packings are fabricated from corrugated sheets tightly packed to each other with corrugations of neighboring sheets oriented in opposite direction, forming a multiplicity of triangular flow channels crossing each other at an angle depending on the corrugation inclination angle (usually 45° or 60° with respect to horizontal axes). The height of a sheet metal packing element or packing layer is usually about 0.20 m (European manufacturers) or 0.25 m or more (American manufacturers), and overall size and weight of individual segments is intended to comply with requirements for installation through manholes. All segments put together form a layer of packing, and each layer is usually rotated by 90° with respect to the previous one. Therefore a packed bed consisting of structured packings represents a highly ordered structure that allows radial spreading and mixing of phases at each transition between packing layers. The extent of liquid and vapor spreading depends on the corrugation inclination angle and is therefore larger in case of common 45° than in case of 60° packings.
The nominal size of random packing is usually given in numbers representing characteristic dimension like the diameter in the case of rings (1-4 inch, i.e. 25-100 mm, for packings used in industrial columns), and the size of structured packings manufactured in Europe is usually given in numbers representing the amount of specific geometric area
Capacity and efficiency of trays, random and structured packings are directly related to pressure drop. Namely, the upward flow of vapor phase is associated with a certain amount of pressure drop across the column internals, that of trays being much higher than that exhibited under the same conditions by random and particularly structured packings. A larger pressure drop creates a larger pressure and, consequently, a larger temperature at the bottom of the column. If excessive, it can detrimentally affect the separation (reduced relative volatility, bottom product degradation by decomposition or polymerization, and/or need for a hotter heating medium). A detailed account on effects of the operating pressure and the pressure drop on distillation operation is given in THIS CHAPTER.
One should note that with going deeper into vacuum the amount of tolerable pressure drop decreases, and in extreme cases (high boiling, thermally unstable chemicals) it may be so low that column internals need to be avoided. In such cases, special, falling film type or spray type columns are considered.
As mentioned above, upon establishing the reflux and stage requirement of a separation by detailed (rigorous) calculations/simulation, the designer must choose between trays and packings and subsequently find the most appropriate one. Namely, trays, random packings, and structured packings come in numerous versions, which differ considerably in their performance characteristics, and in the range of applications where both could be applied a thorough consideration of all relevant factors is required. A recently published paper written by two authors representing a major distillation equipment manufacturer and a major user company, gives a concise account on relevant performance characteristics, providing guidelines when to choose trays and when packings [3]. Therefore, as well as due to the fact that a detailed account of trays, random packings, and structured packings can be found in the Chapters 2, 3, and 4, respectively, there is no need in this chapter to go into too much detail on these topics. In addition, various aspects of performance characteristics of trays and random and structured packings are addressed to certain extent in Chapter 9 (high and low pressure distillation) and Chapter 10 (scale up), as well as in chapters addressing equipment testing and various applications in the book Distillation Operation and Applications [4].
The present, equipment-related introductory chapter addresses and qualitatively compares performance characteristics of standard, well-established tray-type contactors, e.g. sieve, valve, and bubble cup trays, with random and structured packings.
2.1 Trays
Standard trays considered here are cross flow trays with active or bubbling area situated in between inlet and outlet downcomers. Common nonproprietary devices used to facilitate contact of vapor passing through the liquid flowing over the active area are known generally as sieve, valve, and bubble cup trays or decks. Basic geometric features of sieve, valve, and bubble cup trays are shown schematically in Figure 7 in This Chapter. Being essentially perforated thin plates, sieve trays are simplest and cheapest to manufacture, and most importantly, exhibit good efficiency over a wide range of operating conditions. However, if more flexibility is required, then valve trays will be considered. Floating or movable valve trays are preferred in refinery applications where more or less pronounced fluctuations in feed supply occur from time to time. In general, floating-valve trays are more expensive and prone to fouling and mechanical damage. The latter is not the case with so-called fixed-valve trays, which combine advantages of a fully open floating valve tray and a sieve tray at a price not exceeding that of common sieve trays.
Owing to the fact that in the case of sieve and floating or moving valve trays as well as fixed-valve trays the liquid is maintained on the deck by kinetic energy of the vapor, these trays are limited at low vapor flow rates, where weeping, if excessive, leads to significant loss of efficiency. If the lower operating limit is a concern, bubble caps, which employ a physical seal that prevents liquid from leaving the tray deck, are considered. Their major drawback is constructional complexity and an appreciably higher price. However, more practical modifications of bubble cap trays such as “tunnel trays” and “Thormann trays” (Figure 2) allow cost-effective implementation of these devices where appropriate (www.montz.de). A distinguishing feature of a Thormann tray is that it can handle the lowest specific liquid flows, down to 40 l/m2 h. This, in combination with a rather low pressure drop, makes it also suitable for vacuum applications.
Other well-established types of trays are: multidowncomer trays and various high-capacity trays, as well as the trays without downcomers, such as dual flow trays and baffle trays. The main feature that distinguishes dual flow trays is the fact that both liquid and vapor are expected to use the same hole or neighboring holes, which means that unlike with other trays, the mass transfer occurs under countercurrent flow conditions. However, we will here consider mainly the characteristics of standard sieve, valve, and bubble cap cross flow trays, and other tray types (see This Chapter) will be mentioned where appropriate, just to indicate their specific benefit with respect to standard cross flow trays.
Regarding the operation of a cross flow tray, one of the important parameters is flow path length, because it affects efficiency. In case of a short distance between downcomers, as encountered in small-diameter columns or on multi-downcomer trays, the liquid is well mixed, that is, its concentration is homogeneous, and thus the driving force for mass transfer is constant and the resulting efficiency is the lowest one encountered in normally operating distillation columns. Longer flow paths allow a uniform liquid flow pattern to develop approaching plug flow condition, which in turn ensures the largest driving force, leading to the highest achievable tray efficiency. However, at large-diameter trays with large liquid flow paths, flow anomalies like stagnant or recirculation zones can be induced by the geometry of the tray, allowing a certain amount of backmixing to occur, leading to the loss of efficiency. Here we speak in particular about the efficiency-deteriorating effect of the liquid maldistribution on a tray. On the other hand, if the depth of froth layer varies along the flow path, this can lead to vapor bypassing, which is a common form of vapor maldistribution observed in practice [2].
The clear liquid height and consequently froth height on a tray are dictated by the liquid flow rate and weir height, and an increase in weir height leads to increased efficiency, which however comes at the cost of increased pressure drop. This is a typical trade-off situation, while the weir height is dictated mainly by the amount of affordable pressure drop. This means lower than normal (50 mm) weir height in vacuum and higher in high-pressure applications. Another important design parameter is the weir length, that is, the related maximum liquid load per unit weir length. If excessive it will affect efficiency adversely. To accommodate larger liquid loads, weir length needs to be extended, which can be done in different ways, but if this proves to be insufficient a tray with the two or more downcomers is chosen.
If in excess of tolerated values, weeping on the low end and entrainment on the high end represent respectively lower and upper limits of tray operating range. The latter is a primary concern because it is related to the achievable approach to the upper operating limit, and it hurts efficiency more then weeping that is more gradual in appearance, and can be tolerated to a greater extent than entrainment. It is interesting to mention here that weeping at the end of the flow path is a minor problem, but that occurring at the beginning is dangerous, because the liquid bypasses the tray and falls on the liquid pouring into the downcomer of the tray below, thus effectively bypassing two trays. This as well as many other peculiarities related to the design of trays needs to be identified and handled accordingly during conceptual tray design, and should get proper consideration during detailed design.
2.2 Random packings
Figure 3 shows a photograph of random packings that represent milestones in the development of modern packed column technology. The first generation is the wellknown Raschig ring, patented in 1911; followed by the second-generation Pall ring introduced in the 1960s; the third-generation IMTP saddle-type ring, introduced in late 1970s; and after nearly 100 years of development, the fourth-generation Raschig Super Ring [5]. There are also other well-established random packings of the third generation (see Chapter 3), like the well-known Nutter ring, that belongs to the Sulzer portfolio (www.sulzer.com). Most recently, Koch-Glitsch introduced a fourth-generation random packing known as Intalox Ultra packing, and performance characteristics as observed in FRI tests make it suitable to improve throughput or product purity in the case of a revamp of a column equipped with thirdgeneration random packings [6].
Indeed, regarding capacity the above-mentioned packings represent different generations, and a Pall ring really was a big step forward compared to a Raschig ring; and third-generation packings outperformed the Pall ring significantly; while this cannot be claimed for the fourth generation compared to the third generation. Anyhow a shift towards a better overall performance was achieved and this justifies their position as the fourth generation of random packings.
In distillation applications it is a common practice to use stainless steel packings or another alloy with sufficient corrosion resistance. However, in some applications involving corrosive aqueous systems at low enough temperatures (<373 K or 100°C), packings made of much cheaper polypropylene or other suitable plastic material may be considered. Ceramic packings find application in some highly corrosive applications, but appear to be sensitive to damage, i.e. breakage due to operation upsets that can lead to plugging, i.e. reduction of bed void or free area (porosity) and consequently increased pressure drop. To avoid breakage during installation, ceramic packings are dumped under wet conditions, i.e. into a column filled with water.
2.3 Structured packings
Figure 4 shows three layers of a conventional corrugated sheet structured packing, assembled from segments of the size that allows installation through column manholes.
Corrugated sheet structured packings were introduced in early 1960s, by Sulzer, in the form of so called BX packing. This packing was and still is made of perforated wire gauze material, with corrugations inclined 60° to horizontal. At that time it was considered as the proper internal for the very demanding heavy water separation, and later on found thousands of applications in fine chemicals separation/purification. A real breakthrough of structured packings occurred in the late 1970s, with the introduction of the Sulzer Mellapak series, made of much cheaper metal sheets. This as well as similar packings developed by other manufacturers proved their value first as replacements for various trays in vacuum applications, allowing large capacity increases, often accompanied by significant efficiency increase. From that moment on it became obvious that structured packings, due to the lowest pressure drop per stage, are the best option for vacuum applications. With time it appeared that structured packings cannot perform satisfactorily at high operating pressures, where trays, particularly those handling large liquid loads with ease, remain the preferred choice. Random packings can perform well on both ends and in the middle of the range of operating pressures encountered in distillation, but cannot outperform structured packings or trays in their primary fields of application.
Nowadays the main suppliers and innovators in this field, in addition to the pioneer Sulzer, are Koch-Glitsch, Montz, and Raschig. Presently manufacturers from China and India are attempting to enter the global market by offering various packings often at significantly lower prices. While Sulzer Mellapak, Koch-Glitsch Flexipac, and Montz Montz-Pak rely on the established corrugated sheet structure, with textured, perforated, or imperforated corrugated sheets that allow continuous flow of liquid and force vapor to follow the channels, the corrugated sheets of Raschig Superpac are open (see Figure 5), enabling both liquid and vapor to easily switch to the other side. For vapor this means an effective flow angle that is steeper than that of conventional corrugations. The reduced pressure drop brings a capacity advantage that is equivalent to that of advanced designs of corrugated sheet structured packings, those employing a smooth bend towards the vertical on lower (see Figure 6) or both ends of corrugations (see Figure 6 in This Chapter).
The high performance packings of Koch-Glitsch, Montz, and Sulzer are known as Flexipac HC, Montz-Pak M (imperforated) or MN (with holes), and MellapakPlus, respectively. The favorable performance characteristics of the Montz-Pak MN series packings have been demonstrated in tests conducted most recently using semi- and industrial-scale installations [8,9].
Figure 7 shows details of corrugated sheet geometry, including all relevant parameters that allow estimation of specific geometric area, ageo,p (m2 /m3 ), hydraulic diameter for gas or vapor flow, dh (m), and packing void fraction or porosity, ε (m3 /m3 ) using geometry-based expressions given in Table 1.1.
Nevertheless this is not as simple as it appears. Namely, packing serves as carrier for liquid, and the surface of the liquid film provides interfacial area for mass transfer. Complete wetting is a prerequisite for successful operation. To assure this, a well-performing liquid distributor is required to provide adequate quality of initial liquid distribution, while the quality of liquid distribution within the packed bed will depend on type and size of packing, layout of a layer consisting of segments, operating conditions, and surface tension.
However, due to the efficiency-deteriorating effect of liquid maldistribution, which tends to develop with increasing bed depth, the height of a packed bed is usually limited to a height equivalent to a certain number of theoretical plates, 15 being a typical number. The liquid leaving a bed of given height needs to be collected and thoroughly mixed before being delivered to the distributor of the bed below. The column height required to install a liquid redistribution section is up to 2 m, which means that in case of multistage separations ancillary equipment will contribute significantly to total column height and cost. This is in practice considered to be a good reason to avoid a packed column in a new design, while the manufacturers have often been forced in revamp situations to install taller than usual beds, to ensure that liquid redistribution sections can be accessed through existing manholes. It is well known that this was done with success in numerous cases and that in industrial columns equipped with standard-size packings (220-250 m2 /m3 ) there are wellperforming beds of exceptional height (more than 20 stages and in excess of 10 m). With this, the outgoing assumption that “maximum bed height” extended by 2 to 3 m can add, say, three to four stages as well as the height necessary to compensate for anticipated loss in efficiency, was justified. However, this should not be done at the cost of mechanical integrity of the bed; it will depend on the type and size of packing as well as the construction material.
The extent of liquid maldistribution and consequently the allowable bed height will depend on the structure of the packed bed, which may be randomly or orderly packed. In other words, if considering a packed column, a clear distinction is made between random and structured packings. The former, often called “dumped packings”, are small, open bodies of various shapes and sizes that are dumped into the column (dry or wet, depending on situation) and settle randomly. Their liquid spreading extent is rather limited, that is, it does not exceed the dimension of the body. The structured packings are fabricated from corrugated sheets tightly packed to each other with corrugations of neighboring sheets oriented in opposite direction, forming a multiplicity of triangular flow channels crossing each other at an angle depending on the corrugation inclination angle (usually 45° or 60° with respect to horizontal axes). The height of a sheet metal packing element or packing layer is usually about 0.20 m (European manufacturers) or 0.25 m or more (American manufacturers), and overall size and weight of individual segments is intended to comply with requirements for installation through manholes. All segments put together form a layer of packing, and each layer is usually rotated by 90° with respect to the previous one. Therefore a packed bed consisting of structured packings represents a highly ordered structure that allows radial spreading and mixing of phases at each transition between packing layers. The extent of liquid and vapor spreading depends on the corrugation inclination angle and is therefore larger in case of common 45° than in case of 60° packings.
The nominal size of random packing is usually given in numbers representing characteristic dimension like the diameter in the case of rings (1-4 inch, i.e. 25-100 mm, for packings used in industrial columns), and the size of structured packings manufactured in Europe is usually given in numbers representing the amount of specific geometric area
Capacity and efficiency of trays, random and structured packings are directly related to pressure drop. Namely, the upward flow of vapor phase is associated with a certain amount of pressure drop across the column internals, that of trays being much higher than that exhibited under the same conditions by random and particularly structured packings. A larger pressure drop creates a larger pressure and, consequently, a larger temperature at the bottom of the column. If excessive, it can detrimentally affect the separation (reduced relative volatility, bottom product degradation by decomposition or polymerization, and/or need for a hotter heating medium). A detailed account on effects of the operating pressure and the pressure drop on distillation operation is given in THIS CHAPTER.
One should note that with going deeper into vacuum the amount of tolerable pressure drop decreases, and in extreme cases (high boiling, thermally unstable chemicals) it may be so low that column internals need to be avoided. In such cases, special, falling film type or spray type columns are considered.
2. Performance characteristics of vaporeliquid contactors
As mentioned above, upon establishing the reflux and stage requirement of a separation by detailed (rigorous) calculations/simulation, the designer must choose between trays and packings and subsequently find the most appropriate one. Namely, trays, random packings, and structured packings come in numerous versions, which differ considerably in their performance characteristics, and in the range of applications where both could be applied a thorough consideration of all relevant factors is required. A recently published paper written by two authors representing a major distillation equipment manufacturer and a major user company, gives a concise account on relevant performance characteristics, providing guidelines when to choose trays and when packings [3]. Therefore, as well as due to the fact that a detailed account of trays, random packings, and structured packings can be found in the Chapters 2, 3, and 4, respectively, there is no need in this chapter to go into too much detail on these topics. In addition, various aspects of performance characteristics of trays and random and structured packings are addressed to certain extent in Chapter 9 (high and low pressure distillation) and Chapter 10 (scale up), as well as in chapters addressing equipment testing and various applications in the book Distillation Operation and Applications [4].
The present, equipment-related introductory chapter addresses and qualitatively compares performance characteristics of standard, well-established tray-type contactors, e.g. sieve, valve, and bubble cup trays, with random and structured packings.
2.1 Trays
Standard trays considered here are cross flow trays with active or bubbling area situated in between inlet and outlet downcomers. Common nonproprietary devices used to facilitate contact of vapor passing through the liquid flowing over the active area are known generally as sieve, valve, and bubble cup trays or decks. Basic geometric features of sieve, valve, and bubble cup trays are shown schematically in Figure 7 in This Chapter. Being essentially perforated thin plates, sieve trays are simplest and cheapest to manufacture, and most importantly, exhibit good efficiency over a wide range of operating conditions. However, if more flexibility is required, then valve trays will be considered. Floating or movable valve trays are preferred in refinery applications where more or less pronounced fluctuations in feed supply occur from time to time. In general, floating-valve trays are more expensive and prone to fouling and mechanical damage. The latter is not the case with so-called fixed-valve trays, which combine advantages of a fully open floating valve tray and a sieve tray at a price not exceeding that of common sieve trays.
Owing to the fact that in the case of sieve and floating or moving valve trays as well as fixed-valve trays the liquid is maintained on the deck by kinetic energy of the vapor, these trays are limited at low vapor flow rates, where weeping, if excessive, leads to significant loss of efficiency. If the lower operating limit is a concern, bubble caps, which employ a physical seal that prevents liquid from leaving the tray deck, are considered. Their major drawback is constructional complexity and an appreciably higher price. However, more practical modifications of bubble cap trays such as “tunnel trays” and “Thormann trays” (Figure 2) allow cost-effective implementation of these devices where appropriate (www.montz.de). A distinguishing feature of a Thormann tray is that it can handle the lowest specific liquid flows, down to 40 l/m2 h. This, in combination with a rather low pressure drop, makes it also suitable for vacuum applications.
Other well-established types of trays are: multidowncomer trays and various high-capacity trays, as well as the trays without downcomers, such as dual flow trays and baffle trays. The main feature that distinguishes dual flow trays is the fact that both liquid and vapor are expected to use the same hole or neighboring holes, which means that unlike with other trays, the mass transfer occurs under countercurrent flow conditions. However, we will here consider mainly the characteristics of standard sieve, valve, and bubble cap cross flow trays, and other tray types (see This Chapter) will be mentioned where appropriate, just to indicate their specific benefit with respect to standard cross flow trays.
 |
| FIGURE 2. Photographs of the Top View of a 3.4 m Diameter Thormann Tray and of a Cap Containing Vapor Directing Openings That Push the Liquid across the Tray Following the Pattern Shown in the Attached Drawing on the Left Side |
Regarding the operation of a cross flow tray, one of the important parameters is flow path length, because it affects efficiency. In case of a short distance between downcomers, as encountered in small-diameter columns or on multi-downcomer trays, the liquid is well mixed, that is, its concentration is homogeneous, and thus the driving force for mass transfer is constant and the resulting efficiency is the lowest one encountered in normally operating distillation columns. Longer flow paths allow a uniform liquid flow pattern to develop approaching plug flow condition, which in turn ensures the largest driving force, leading to the highest achievable tray efficiency. However, at large-diameter trays with large liquid flow paths, flow anomalies like stagnant or recirculation zones can be induced by the geometry of the tray, allowing a certain amount of backmixing to occur, leading to the loss of efficiency. Here we speak in particular about the efficiency-deteriorating effect of the liquid maldistribution on a tray. On the other hand, if the depth of froth layer varies along the flow path, this can lead to vapor bypassing, which is a common form of vapor maldistribution observed in practice [2].
The clear liquid height and consequently froth height on a tray are dictated by the liquid flow rate and weir height, and an increase in weir height leads to increased efficiency, which however comes at the cost of increased pressure drop. This is a typical trade-off situation, while the weir height is dictated mainly by the amount of affordable pressure drop. This means lower than normal (50 mm) weir height in vacuum and higher in high-pressure applications. Another important design parameter is the weir length, that is, the related maximum liquid load per unit weir length. If excessive it will affect efficiency adversely. To accommodate larger liquid loads, weir length needs to be extended, which can be done in different ways, but if this proves to be insufficient a tray with the two or more downcomers is chosen.
If in excess of tolerated values, weeping on the low end and entrainment on the high end represent respectively lower and upper limits of tray operating range. The latter is a primary concern because it is related to the achievable approach to the upper operating limit, and it hurts efficiency more then weeping that is more gradual in appearance, and can be tolerated to a greater extent than entrainment. It is interesting to mention here that weeping at the end of the flow path is a minor problem, but that occurring at the beginning is dangerous, because the liquid bypasses the tray and falls on the liquid pouring into the downcomer of the tray below, thus effectively bypassing two trays. This as well as many other peculiarities related to the design of trays needs to be identified and handled accordingly during conceptual tray design, and should get proper consideration during detailed design.
2.2 Random packings
Figure 3 shows a photograph of random packings that represent milestones in the development of modern packed column technology. The first generation is the wellknown Raschig ring, patented in 1911; followed by the second-generation Pall ring introduced in the 1960s; the third-generation IMTP saddle-type ring, introduced in late 1970s; and after nearly 100 years of development, the fourth-generation Raschig Super Ring [5]. There are also other well-established random packings of the third generation (see Chapter 3), like the well-known Nutter ring, that belongs to the Sulzer portfolio (www.sulzer.com). Most recently, Koch-Glitsch introduced a fourth-generation random packing known as Intalox Ultra packing, and performance characteristics as observed in FRI tests make it suitable to improve throughput or product purity in the case of a revamp of a column equipped with thirdgeneration random packings [6].
 |
| FIGURE 3 Photograph of the Representatives of the First (Raschig Ring), Second (Pall Ring), Third (IMTP Saddle), and Fourth (Raschig Super Ring) Generations of Random (Dumped) Packings |
Indeed, regarding capacity the above-mentioned packings represent different generations, and a Pall ring really was a big step forward compared to a Raschig ring; and third-generation packings outperformed the Pall ring significantly; while this cannot be claimed for the fourth generation compared to the third generation. Anyhow a shift towards a better overall performance was achieved and this justifies their position as the fourth generation of random packings.
In distillation applications it is a common practice to use stainless steel packings or another alloy with sufficient corrosion resistance. However, in some applications involving corrosive aqueous systems at low enough temperatures (<373 K or 100°C), packings made of much cheaper polypropylene or other suitable plastic material may be considered. Ceramic packings find application in some highly corrosive applications, but appear to be sensitive to damage, i.e. breakage due to operation upsets that can lead to plugging, i.e. reduction of bed void or free area (porosity) and consequently increased pressure drop. To avoid breakage during installation, ceramic packings are dumped under wet conditions, i.e. into a column filled with water.
2.3 Structured packings
Figure 4 shows three layers of a conventional corrugated sheet structured packing, assembled from segments of the size that allows installation through column manholes.
Corrugated sheet structured packings were introduced in early 1960s, by Sulzer, in the form of so called BX packing. This packing was and still is made of perforated wire gauze material, with corrugations inclined 60° to horizontal. At that time it was considered as the proper internal for the very demanding heavy water separation, and later on found thousands of applications in fine chemicals separation/purification. A real breakthrough of structured packings occurred in the late 1970s, with the introduction of the Sulzer Mellapak series, made of much cheaper metal sheets. This as well as similar packings developed by other manufacturers proved their value first as replacements for various trays in vacuum applications, allowing large capacity increases, often accompanied by significant efficiency increase. From that moment on it became obvious that structured packings, due to the lowest pressure drop per stage, are the best option for vacuum applications. With time it appeared that structured packings cannot perform satisfactorily at high operating pressures, where trays, particularly those handling large liquid loads with ease, remain the preferred choice. Random packings can perform well on both ends and in the middle of the range of operating pressures encountered in distillation, but cannot outperform structured packings or trays in their primary fields of application.
 |
| FIGURE 4. Photograph Illustrating the Layout of a Large-Diameter Corrugated Sheet Structured Packing Bed |
The high performance packings of Koch-Glitsch, Montz, and Sulzer are known as Flexipac HC, Montz-Pak M (imperforated) or MN (with holes), and MellapakPlus, respectively. The favorable performance characteristics of the Montz-Pak MN series packings have been demonstrated in tests conducted most recently using semi- and industrial-scale installations [8,9].
 |
| FIGURE 5 Photograph of a Sheet of Raschig Super-Pak |
 |
| FIGURE 6. Photograph Showing the Orientation of Subsequent Elements of a High-Capacity, B1-250M Packing, with the Bottom of Corrugations Bent to Vertical |
Figure 7 shows details of corrugated sheet geometry, including all relevant parameters that allow estimation of specific geometric area, ageo,p (m2 /m3 ), hydraulic diameter for gas or vapor flow, dh (m), and packing void fraction or porosity, ε (m3 /m3 ) using geometry-based expressions given in Table 1.1.
0 comments:
Posting Komentar