FRI VOLUME 5: FRACTIONATION DESIGN HANDBOOK
SURGE
Is s u ed : Revised:
3/23/2011
4.01
SURGE
Surge requirements are fundamental parts of the design of any processing system and determining surge requirements must be done with the same kind of scientific considerations and engineering judgment that is applied to the design of all parts of the processing plant. For fractionation systems surge is provided in the column base, in any chimney trays the the column has, and in the reflux drum. drum. Literature on the subject is is limited and resides primarily in the standard standard of user companies. In the 1990’s the member companies companies of FRI were asked to complete a survey on surge and the results of that survey have been reviewed, analyzed, and published in TR 167. It is recommended that all people producing or reviewing fractionation system system designs read TR 167 to become familiar with practices common in in the industry, as should should people who are preparing preparing or reviewing definitive estimates. For those companies that do not have suitable standards and procedures covering covering this area of interest, TR 167 might also serve as a starting point for the development of those standards and procedures. below. All are based on liquid liquid flows leaving the For scoping studies only , some rough guidelines are offered below. column system. For flows from a column bottom bottom sump to reactors, furnaces, or other columns, columns, surge volumes of 15 minutes are typical; for flows to heat exchangers surge volumes of 5 minutes are typical; and for flows to simple storage surge volumes volumes of 3 minutes are typical. For flow from a reflux drum, a surge volume volume is typically 5 minutes when based on reflux flow or 15, 5 or 3 minutes based on the distillate flow as defined above. For flows from chimney trays and the column base, there is a variety of considerations described in TR 167 that may require more or less surge. Surge volume is defined here as the volume within the normal operating range of the instrument measuring the level, that is, from normal high liquid level to the normal low liquid level. Other criteria will establish how much space or volume is required above the normal high level and below the normal low level.
Page 1 of 1
FRI VOLUME 5: FRA CTIONATION DESIGN DESIGN HANDBOOK
REMOVAL OF ENTRAINED LIQUIDS
Issued:
12/15/1990
Revised:
4.02
REMOVAL OF ENTRAINED LIQUIDS
REMOVAL OF ENTRAINED LIQUIDS................................ ................................................. 1 1.
Summary .................................................. ........................................................ ........................................................................... ................... 2
2.
Process Categories ........................................................ ............................................................................................................. .......................................................2
3.
Equipment Types ..................................................... ............................................................................................................... .......................................................... .2
Page 1 of 2
Issued: 12/15/1990 Revised:
1
REMOVAL OF ENTRAINED LIQUIDS
4.02
Summary
The most commonly used de-entrainment de-entrainment device is the wire wire mesh mist eliminator. In many applications this this provides entirely satisfactory service. In other duties, where mesh pads would prove troublesome or fail, packings and trays are sometimes used. Although data on their performance is limited, reasonable assumptions appear to lead to successful systems. Guidance on wire mesh mist eliminators is given in Section 4.02.1 of this Manual.
2
Process Categories
In the removal of entrained liquid liquid from a gas stream, 100% efficiency efficiency often cannot be achieved. The extent and nature of the shortfall which is acceptable depends on the process objectives, which may be categorized as follows: 1.
Avoid carry over of liquid which could cause damage to downstream equipment, eg. by accumulated liquid slugging into a reciprocating reciprocating compressor. Although collection collection efficiency (especially for large drops) needs to be high, small mist particles (say less than 20 micron) may cause no harm.
2.
Avoid carry over of specific components in the entrained liquid: • • • •
product contaminants (eg. colored molecules), process contaminants (eg. which would poison catalyst in a downstream process), high value circulating process materials (eg. selective solvents, or catalyst solutions), solids in solution or molecules which polymerize - these may cause blockage in downstream equipment.
Removal of the specific components may be required to a very high efficiency, but carry over of liquid which does not contain these components may well be quite acceptable.
3
Equipment Types
Five general equipment types are used for the removal of entrainment caused by mechanical means: a. b. c. d. e.
wire mesh mist eliminators (mesh pads) baffle separators cyclone separators packings trays
The baffle and cyclone types use centrifugal deposition caused by changes in the main direction of gas flow. Mesh pads and dumped packings achieve inertial impaction on collecting elements placed in the flowing gas. On trays, liquid droplets droplets in the entering gas stream are captured by three three mechanisms: inertial impaction on the underside of the tray, contact with the bulk liquid both during bubble formation and in the turbulent froth, and (valve and bubble cap trays only) only) by centrifugal separation separation on the valve or bubble cap. When trays are used, care is required in their design, as poorly designed trays may entrain as much liquid as they capture.
Page 2 of 2
FRI VOLUME 5: FRACTIONATION DESIGN HANDBOOK
WIRE MESH TYPE MIST ELIMNATORS
Issued:
01/15/1992
Revised:
4.02.01
WIRE MESH TYPE MIST ELIMINATORS
WIRE MESH TYPE MIST ELIMINATORS ........................................................................... 1 1.
General .................................................... ............................................................................................................ ............................................................................ .................... 2
2.
Selection ................................................ ....................................................... ............................................................................. ...................... 2
3.
Sizing ...................................................... ............................................................................................................. ............................................................................ ..................... 2
4.
Liquid Removal Efficiency....................................................... .......................................... 3
5.
Pressure Drop .................................................... ........................................................................................................... ................................................................. .......... 3
6.
Location .................................................. ........................................................ ............................................................................ .................... 4
7.
Installation ........................................................ ............................................................................................................... .................................................................. ........... 4
Page 1 of 7
Issued: 01/15/1992 Revised:
1
WIRE MESH TYPE MIST ELIMINATORS
4.02.01
General
Mist eliminators have been widely used in process equipment to remove entrained liquid droplets which are formed during vapor-liquid vapor-liquid contact, evaporation, evaporation, or condensation. condensation. Wire mesh mist eliminators eliminators are fabricated from fine wire which is crimped and flattened to form a double layer. This double layer layer material is then either rolled into into a cylinder or layered to form the desired desired thickness of mesh pad. pad. This + 2 3 arrangement provides free volumes of 90 to 99 % with surface areas from 50 to 600 ft /ft (164-1968 m2/m3). As the vapor carrying the entrained liquid liquid passes through the mesh, the liquid liquid particles strike and adhere to the fine wire, coalesce, and drain by gravity. Mesh mist eliminators are fabricated from a wide variety of materials, and wire diameters ranging from 0.003 to 0.015 inches (.076-.381 mm), with 0.011 inches inches (.28 mm) being the most frequently used. Three 3 general density density ranges ranges are common: 5 lb/ft lb/ft (80 kg/m3), considered anti-fouling; 9 lb/ft 3 (144 kg/m 3), considered standard; and 12 lb/ft 3 (192 kg/m 3), considered high high efficiency. All other things things being equal, the use of finer wire improves separation separation performance, but cost rises rapidly rapidly as wire size decreases. The use of metal wire diameters other than 0.011 inches is only infrequently recommended or, indeed, needed.
2
Selection
Before a particular type of mist eliminator is selected the following criteria should be considered: 1. 2. 3. 4. 5. 6. 7.
Type of service service Removal efficiency Pressure drop available Vapor loadings Nature of the mist mist or spray (drop size distribution) distribution) Fouling and corrosion corrosion Material of construction construction
To remove entrained liquid in the top section of a distillation tower a six inch thick standard type is most commonly used. Carbon steel is not not recommended due to to rusting rusting and and flammability flammability hazards. hazards. For corrosive or fouling services, special consideration must be given on the choice of material, wire diameter, and the void space. space. For severely fouling or extremely dirty services, services, a chevron type mist eliminator should be considered. To purchase any type of mist eliminator the following must be specified: 1. 2. 3. 4. 5.
3
Thickness and density density Material of construction construction Top and bottom support support requirements Removal and installation installation requirements (i.e. from above or below) below) Vessel manhole manhole size size (I.D.)
Sizing
The sizing of mist eliminators follows a "C" factor type approach as follows:
Page 2 of 7
Issued: 01/15/1992 Revised:
WIRE MESH TYPE MIST ELIMINATORS
V − C Where:
4.02.01
D L − DV DV = = = =
V DV D L C
is design vapor velocity in ft/sec is the vapor density in lb/ft 3 is the liquid density in lb/ft 3 is an empirical constant in ft/sec
Although early literature (46) suggests that the "C" factor is a function of the distance between the mist eliminator and the source of entrainment, 0.35 is almost universally used for "standard" designs. Some engineers use lower values for special applications as shown in the following table.
SERVICE
C-FACTOR
Compressor suction KO drums
0.30
Steam, atmospheric and above
0.25
Most vacuum service (0.1-10) psia)
0.20
Turbine feed KO drums
0.175
High vacuum systems
0.15
For general services, a value of 0.40 has been suggested for low density (i.e., 5 lbs/ft 3) mesh and a value of 0.30 is recommended for plastic meshes (48).
4
Liquid Removal Efficiency
For most applications found in a typical petrochemical plant 98-99 + percent of the liquid droplets can be removed when operating from 30 to 110% of the design vapor velocities. A semi-theoretical equation (47) has been proposed to calculate the removal efficiency, but the lack of information on droplet size distribution and the amount of entrainment make the calculated value at best an approximation.
5
Pressure Drop
Pressure drop through a mist eliminator is usually low (less than one inch of water for a six inch pad) at low liquid rates. However, the pressure drop increases significantly when the liquid loading increases. Pressure drop prediction for mist eliminators can be found in many vendor's brochures. For a conservative estimate of the dry or very low liquid rate pressure drop through a standard six inch, 9 lb/ft 3 density mist eliminator, the following equation can be used.
dp − 0.3 x V 2 x DV Where:
dp V DV
= = =
pressure drop in inches of water velocity in ft/sec the vapor density in lb/ft 3 Page 3 of 7
Issued: 01/15/1992 Revised:
6
WIRE MESH TYPE MIST ELIMINATORS
4.02.01
Location
In locating a meshpad within a tower two vertical dimensions must be considered: the distance from the active operation below the meshpad to the underside of the mesh and the distance from the top of the mesh to the vapor outlet. For trayed columns, the underside of the meshpad should be at least 1.5 tray spacings above the deck of the top tray. For packed columns, the underside of the meshpad should be at least 12 inches (305 mm) above the uppermost active part of the liquid distributor. There should also be at least 12 inches below the mesh of essentially unobstructed space, that is, a volume where the rising vapors have been allowed to achieve the vertical upflow pattern with which they impinge on the mesh. Where the liquid distributor is one that restricts the vapors, and, hence, causes a rapid increase in vapor velocity as the vapor passes the distributor, additional space between the distributor and the meshpad may be prudent to permit vapor flow to become relatively uniform before it reaches the mesh pad. The minimum vertical distance between the top of the mesh and the vapor outlet should be 12 inches or as defined by Figure 1, whichever is greater. Dimension D M is the diameter of a circle through which all the column vapor would flow at a velocity equal to 100% of the calculated design vapor velocity for the mesh pad used (with C-Factor adjusted for a special application if appropriate).
7
Installation
A mist eliminator assembly typically consists of a meshpad and top and bottom support grids resting on a support ring. Spacer rods connecting and welded to the top and bottom grids have been recommended to avoid compression of the mesh and its subsequent disassembly if the pad is otherwise held together with loose clips. Mesh mist eliminators are most commonly attached to the vessel by tie wires, as shown in Figure 2. Some users prefer attaching the meshpad to the vessel by J-bolts and clamps, as shown in Figure 3. If the mist eliminator diameter is smaller than the vessel, it can be placed inside a cylinder through which the gas flow has been directed.
References and Further Readings
45. York, Otto, "Performance of Wire Mesh Demister", CEP, August, 1954. 46. Warner, B.J., and Scauzillo, F., "Design Considerations of Fibrous Filters for Mist Elimination", Socony Mobil Oil Company, paper presented at the Gas Conditioning Conference, University of Oklahoma, 1963. 47. "Fleximesh Design Manual", Divmet Division, Koch Engineering Company. The Following Unreferenced Publications Will Also Be Found To Be Informative And Helpful:
a.
Koch Engineering Bulletin 570, "Fleximesh Mist Eliminators".
b. Glitsch Bulletin 332, "Glitsch Mist Eliminators". c.
York Bulletin 55, "Mist Eliminators".
d. Koch Engineering Bulletin KME-12, "Mist Eliminators".
Page 4 of 7
Issued: 01/15/1992 Revised:
WIRE MESH TYPE MIST ELIMINATORS
Figure 1.
Page 5 of 7
4.02.01
Issued: 01/15/1992 Revised:
WIRE MESH TYPE MIST ELIMINATORS
4.02.01
NOTE-1: Rod spacer may be of welded construction, but must be specified if desired.
Figure 2.
Page 6 of 7
Issued: 01/15/1992 Revised:
WIRE MESH TYPE MIST ELIMINATORS
NOTE-1: Rod spacer may be of welded construction, but must be specified if desired.
Figure 3.
Page 7 of 7
4.02.01
FRI VOLUME 5: FRACTIONATION DESIGN HANDBOOK
TRAYS AND PACKING AS MIST ELIMINATORS
Issued:
12/15/1990
4.02.02
Revised:
TRAYS AND PACKING AS MIST ELIMINATORS
TRAYS AND PACKING AS MIST ELIMINATORS ............................................................. 1 1.
Random Dumped Packing .................................................................................................. 2
2.
Trays ................................................................................................................................... 2
In the following discussion of packings and trays, reference is made to the process categories given in Section 4.02.
Page 1 of 2
Issued: 12/15/199 Revised:
1
TRAYS AND PACKING AS MIST ELIMINATORS
4.02.02
Random Dumped Packing
Packing may be used for both Categories of duty, but there is little point in using it for clean Category 1 duties. Both packings and mesh pads rely on inertial separation on the metal elements in the flow path. By its nature, the mesh pad offers many more elements per foot of depth than a random packing and for the same efficiency the packed bed would need to be considerably deeper. The cost associated with a random packing demister (including packing, support plate and extra shell length) will generally exceed that of a mesh pad. Experience with Category 2 duties appears to be successful. High efficiency is achieved at gas rates approaching the load point and with low rates of clean wash liquid. No data have been found on the entrainment of liquid from the top of a packed bed, but provided a spray distributor is not used, it should be much less than from a spray-irrigated mesh pad. From limited design experience available, it appears that design criteria for an irrigated packed bed distributor are not critical. With liquids which wet the packing surface, rates less and 1 gpm/ft 2 have proved successful. The vapor rate should not exceed the load point of the packing, to minimize the risk of re-entrainment. Conversely, the vapor load should not be so low that the effects of inertial separation are lost. Design for a pressure drop of about 0.4 inches H 2O/ft is suggested. There appears to be no way of estimating the required bed depth, but it appears that a value of around 3 ft should be adequate. For non-wetting liquids (eg. aqueous systems) higher irrigation rates are likely to be required, but no data are available to the Design Practices Committee.
2
Trays
Trays compare unfavorably with mesh pads in most clean Category 1 duties because of the risk of entrainment of the irrigating liquid. They are used for dirty Category 1 as well as Category 2 duties. In principles several tray types can be used, but the usual choice is either a valve tray or bubble cap tray - some users have a distinct preference for valve trays. Both types impose sharp directional changes on the vapor flow and hence a degree of centrifugal separation is achieved. Traditionally bubble cap trays have been chosen, but they may not always have been the best choice since they generate considerably more entrainment than valve trays at low liquid rates. A high rise bubble cap tray may be used, where liquid is deentrained in the cap and vapor does not bubble through liquid in the tray floor; this is for Category 1 duties only. It seems likely that for these tray types the efficiency of capture of droplets into the liquid is high. This is sometimes referred to as the primary efficiency. Values close to 100% appear to be valid for valve and bubble cap trays. However there is some risk of re-entrainment of liquid, so that proper tray design is essential to ensure that the amount of liquid re-entrainment is minimized. By virtue of the flow of irrigating liquid, any re-entrained liquid will have a lower concentration of the materials which must be removed. This may be calculated by material balance if the incoming entrainment rate and concentration and the re-entrainment rate are known or can be estimated. From the specification outlet concentration, the number of trays required may then be estimated. The number will often be in the range 2 - 5; occasionally as many as 8 trays are used. Unless a pumparound irrigation is used, trays used for de-entrainment will generally operate in the spray regime. Careful tray design is necessary, not only to minimize re-entrainment but also to ensure a proper downcomer seal. In particular, bubble cap tray designs are likely to be non-standard. For further guidance on the design of trays for low liquid rates, see Section 1.16 of this manual.
Page 2 of 2
FRI VOLUME 5: FRACTIONATION DESIGN HANDBOOK
EXCHANGERS IN FRACTIONATION SERVICE
Issued:
01/15/1995
4.03.1
Revised:
EXCHANGERS IN FRACTIONATION SERVICE
EXCHANGERS IN FRACTIONATION SERVICE ................................................................ 1 1.
Reboilers ................................................................................................................. ........... 2
2.
Thermosiphon Reboilers................................................... ................................................. 2
3.
Internal Reboilers .................................................. ............................................................. 4
4.
Kettle Reboilers ................................................... .............................................................. 5
5.
Forced Circulation Reboilers ............................................... .............................................. 5
6.
Falling Film Reboilers .................................................. ......................................................5
Page 1 of 12
Issued: 01/15/1995 Revised:
1
EXCHANGERS IN FRACTIONATION SERVICE
4.03.1
Reboilers
Reboilers for distillation columns can be divided into groups in accordance with their hydraulic characteristics. By one way of thinking there would be two categories: 1. 2.
Pool boiling units, such as kettles and internal reboilers; and, High velocity once through and recirculation units, such as thermosiphon reboilers and pump through reboilers.
Since Category 1 exchangers also work by thermosiphoning another way of grouping would be: a. b.
natural circulation; and, forced circulation.
Both natural and forced circulation (external) reboilers (Category 2 above) may be vertical or horizontal. They may also be subdivided by feed system as shown in Figures 1, 2, and 3, and as described below. For purposes of the following discussion, "gross bottoms" is defined as all of the liquid leaving the bottom tray entering the reboiler, and "net bottoms" is liquid of the same composition as the bottoms product of the tower system entering the reboiler.
2
Thermosiphon Reboilers Orientation - Thermosiphon reboilers may be horizontal or vertical. Horizontal units are usually shell side boiling and vertical units are usually tube side boiling.
In the chemical industry horizontal thermosiphon reboilers are generally used when size limitations prohibit the use of vertical units. Horizontal reboilers require separate foundations. They have more expensive piping and increase plot space requirements. Because the process fluid is on the shell side and all process fluids have some fouling tendency, fixed tube sheet design may not be practical. U-tube design should be considered. Horizontal reboilers, on the other hand, have no size limitations. Hydraulic design is less difficult and they allow the use of a fouling heating medium (with straight tubes). The general feeling is that they tend to give better performance in vacuum reboiling of heavy hydrocarbon mixtures because of lower head requirements and better heat transfer coefficients. Longer allowable tube lengths tend to reduce cost and maintenance may be relatively simple. In the petroleum refining industry horizontal reboilers are much more commonly used. Reboilers tend to be large, and the features of long tube length and greater location flexibility (i.e., better access and easier maintenance) are attractive. Vertical reboilers are generally supported on the towers they serve. Because of geometric and hydraulic considerations tube length is limited and for large surface requirements large diameters or multiple units are required. Access is often difficult. The use of horizontal reboilers often permits a lower tower elevation than the use of vertical reboilers. This can be a significant factor in the design of very tall towers. Net Bottoms Feed Systems (Recirculating) - The most common reboilers used are net bottoms feed (recirculating) thermosiphons ( Figure 1). They are simple and have a wide range of applicability.
Net bottoms feed thermosiphon reboilers have several disadvantages when compared to the other available types. The feed to the reboiler is at the same temperature as the net bottoms product, and thus hotter than
Page 2 of 12
Issued: 01/15/1995 Revised:
EXCHANGERS IN FRACTIONATION SERVICE
4.03.1
had the feed come directly from thebottom tray (as in a gross bottoms system). This effect lowers the mean temperature difference (MTD) and may increase thermal fouling and thermal decomposition. The temperature rise is a function of the fluid composition and the recirculation ratio and the significance of its deleterious effects, of course, depend on the particular system. Also, the actual recirculation ratio may be difficult to predict. This type of feed system has the poorest distillation performance. While the reboiler of a gross bottoms feed system (Figure 2) is approximately equal to a theoretical tray and that of a mixed bottoms feed system (Figure 3) is equal to part of a theoretical tray, the reboiler of a net bottoms feed system has essentially no distillation value. This is something that is often overlooked when the results of a process simulation are converted into a column design. The unbaffled net bottoms feed system ( Figure 1-A) has a further disadvantage in that it causes the reboiler to have a varying feed head or the tower to have a rapidly varying net bottoms product flow. If the bottom of the tower has a definite surge function the level will vary and the feed head on the reboiler will not be constant. This may make reboiler operation unsteady, which, in some applications, may be unacceptable. Also, the exchanger must be designed for the worst conditions, that is, the lowest head (lowest level). More surface generally results. These problems can be solved by holding the level constant, but this will negate the surge volume in the tower bottoms. The baffled net bottoms feed system ( Figure 1-B) does not have this last set of disadvantages. It does, however, lose some of the simplicity which is one of its main advantages. It will also, in general, increase the tower length. A less common system of separating the reboiler feed and liquid surge sections employs a horizontal baffle, but this is generally difficult to design, more costly, and the derived advantages minor. Such a system would look something like Figure 3-B. When a net bottoms (or mixed bottoms) feed system is selected, the process engineer must specify the (maximum) percent vaporization (or the recirculation ratio). This is typically 20 - 30 weight percent for hydrocarbons but may be as low as five percent for aqueous systems. Raising the recirculation ratio (lowering the percent vaporization) decreases the temperature rise. This will increase MTD and suppress thermal fouling and decomposition. Raising the recirculation, however, could increase the exchanger size, especially for vacuum service because of the shift of duty to more liquid heating and less vaporization. Net bottoms feed systems should not be used with materials that tend to polymerize as some material may make many passes through the exchanger. Very viscous materials generally prohibit the use of any thermosiphon and design of any thermosiphon in vacuum service is difficult. In both of these cases forced circulation systems, perhaps falling film designs, should be considered. Gross Bottoms Feed Systems (Once-Through) - In the gross bottoms feed system the liquid leaving the bottom tray passes directly to the reboiler inlet ( Figure 2). Here, the reboiler outlet (rather than the inlet as in the net bottoms feed system, above) temperature is the same as the tower bottoms product. MTD's are higher than for the net bottoms feed system and the problems of temperature induced fouling and degradation may be lessened.
Because no material passes through the reboiler more than once, this type of design has found application with materials that tend to polymerize. "Once- through reboiler" is the more common name for this type of unit. The percent vaporization for a once-through reboiler is fixed by distillation requirements as no provisions for recirculation exist. Authorities vary as to the maximum allowable vaporization with values of 30 to 70% being cited. Page 3 of 12
Issued: 01/15/1995 Revised:
EXCHANGERS IN FRACTIONATION SERVICE
4.03.1
Mixed Bottoms Feed Systems - The mixed bottoms feed system ( Figure 3) is a compromise between the net and gross bottoms systems discussed above. Whether or not the reboiler outlet temperature is the same as the bottoms product temperature depends on the hydraulics of the system and the geometry of the baffles. The reboiler inlet temperature will be higher than that of the once-through type but lower than that of the net bottoms feed type. It will depend on the recirculation ratio.
The major disadvantage of this type of system is the complexity of the tower internals required. It is, however, the most flexible design.
3
Internal Reboilers
Internal reboilers are inserted directly into the tower and are frequently the least expensive type of reboiler on a first cost basis (but not necessarily on a long term basis). There is no shell, and external piping is minimized. This is generally only partly counterbalanced by an expensive vessel flange and supports inside the column. However, because internal reboilers are relatively uncommon the body of experience about the design of the systems using them is limited, and the proper design of those systems requires special care and consideration. Internal reboilers present additional maintenance difficulties, including the need to de-inventory the column prior to maintenance. Uncovering the top of the bundle can cause high tube wall temperatures with all its associated problems. Internal reboilers are invariably of U-tube construction with the length being limited by the column diameter. This limitation usually makes the diameter of the internal reboiler larger than the bundle diameter of an equivalent external unit. Tower and bundle diameter considerations limit the applicability of internal reboilers to services where the column size / reboiler size ratio is relatively high. There is, however, no reason why a tower could not have both internal and external reboilers. Figures 4-A and 4-B show two internal reboiler arrangements. In Figure 4-A "simple" design surge is provided by varying the level above the tube bundle. This creates somewhat unsteady operation, often a significant problem, partly because the liquid above the exchanger exists primarily as a froth and its density is difficult to predict and generally varies with operating loads. Level control is therefore a serious potential problem area. If the space above the tube bundle, or above the maximum design liquid level, has not been carefully selected there is a danger that a frothier than expected liquid can reach the bottom tray (or bottom of the packing) while the level meter is incorrectly "seeing" a level that appears to be safely below the danger point. When this happens column flooding is likely and mechanical damage is possible. The risk of this happening with the "simple" design can be reduced by providing enough distance between the boiling liquid and the device into which the generated vapors flow. Such a distance could be calculated by looking at a situation where the effective density of the boiling liquid has been unexpectedly reduced to its lowest possible value while the level meter is calibrated for the normal density. If the distance above the bundle is engineered in this manner the system must be provided with a level meter that measures from the top of the bundle to the top of this engineered distance. Backup would be a separate level alarm high up in this distance that would activate when it encounters a "minimum density" liquid (or froth). Some work in this area was done by Hepp(89) for a number of refinery towers in the 150 to 325 psig (10.3-22.4 bar) pressure range and an F S range of 0.23 to 0.72 (0.28-0.88). Hepp provides some useful insights and proposes guidelines. Because of the narrowness of the ranges of operation studied using the correlation developed much outside those ranges, or with different kinds of systems, would appear to be risky.
The "simple" design is generally considered unsatisfactory if the bottoms material tends to polymerize or
Page 4 of 12
Issued: 01/15/1995 Revised:
EXCHANGERS IN FRACTIONATION SERVICE
4.03.1
otherwise has a greater tendency to foul. On the other hand, the separate reboiler compartment in the "separate surge" could become a trap for solids. The Figure 4-B "separate surge" design reduces the problems of unsteady operation, loss of level control, and polymerization and fouling. Compared to an optimistic (i.e., risky) "simple" design the "separate surge" design could be more costly; however compared to a conservative "simple" design, that is not necessarily the case. The distance from the top of the liquid crest overflowing the sides of the boiling compartment to the bottom of the tray or packed section above would typically be in the three to six foot range, depending on column size, reboiler configuration (e.g., number of bundles), and type of mass transfer device used.
4
Kettle Reboilers
Kettle reboilers ( Figure 5) are somewhat similar to internal reboilers but with the services physically separated. They are generally considered more costly than equivalent internal reboilers and some designers feel they should be used only when size or other mechanical limitations prohibit the use of internal reboilers. Others feel that because of the greater flexibility available with a kettle and the greater experience the industry has with them that kettles are preferred. Kettles are generally used only for relatively clean service and, as such, have no real limitations on percent vaporization. Kettles are preferred when available temperature driving forces are very low. They are not considered good for wide boiling mixtures because of the difficulty in predicting internal recirculation and effective temperature driving forces.
5
Forced Circulation Reboilers
Forced circulation (pump through) reboilers are used for viscous service, some vacuum applications, and services where solids are present. Any recirculation reboiler system can utilize forced circulation. Forced circulation provides maximum flexibility. Recirculation can be accurately controlled and very high recirculations are possible. Because relatively high pressure drop is available less expensive exchangers and smaller piping can be used. Hydraulics problems are eliminated and greater layout flexibility is possible. The high available pressure drop allows filtration of the reboiler feed. When fouling is a consideration forced circulation reboilers are often the system of choice. High process side velocities can keep the tubes clean and the high pressure drop available can, if it is desired, suppress vaporization until after the process fluid exits the exchanger. This, of course, requires that all the duty go in initially as sensible heat (i.e., temperature rise) which can lower temperature driving forces or require very large circulations. The initial and operating cost of the pumping system is the major disadvantage.
6
Falling Film Reboilers
Falling film reboilers are vertical units where the process fluid is pumped onto the top tubesheet to flow downward as a thin film on the inside walls of the tubes ( Figure 6). Boiling starts immediately and the generated vapor flows downward in the inner portions of the tubes. The two fluids collect in the lower channel and are conveyed to the column via carefully designed means, generally the special design of the bottom channel. Falling film reboilers are generally used with very heat sensitive materials. They are also used in viscous services, in vacuum services, and with wide boiling mixtures. The only suppression of Page 5 of 12
Issued: 01/15/1995 Revised:
EXCHANGERS IN FRACTIONATION SERVICE
4.03.1
vaporization is the slight pressure drop from the top tubesheet to the column. There is never any liquid head to overcome. The highest temperature the process fluid sees (at the top of the tubes) can often be only fractions of a degree above the outlet temperature if the designer so desires.
Page 6 of 12
Issued: 01/15/1995 Revised:
EXCHANGERS IN FRACTIONATION SERVICE
Figure 1. REBOILER CIRCUITS, NET BOTTOMS FEED
Page 7 of 12
4.03.1
Issued: 01/15/1995 Revised:
EXCHANGERS IN FRACTIONATION SERVICE
Figure 2. REBOILER CIRCUITS, GROSS BOTTOMS FEED
Page 8 of 12
4.03.1
Issued: 01/15/1995 Revised:
EXCHANGERS IN FRACTIONATION SERVICE
Figure 3. REBOILER CIRCUITS, MIXED BOTTOMS FEED
Page 9 of 12
4.03.1
Issued: 01/15/1995 Revised:
EXCHANGERS IN FRACTIONATION SERVICE
Figure 4. INTERNAL REBOILERS
Page 10 of 12
4.03.1
Issued: 01/15/1995 Revised:
EXCHANGERS IN FRACTIONATION SERVICE
Figure 5. KETTLE REBOILERS
Page 11 of 12
4.03.1
Issued: 01/15/1995 Revised:
EXCHANGERS IN FRACTIONATION SERVICE
Figure 6. FALLING FILM REBOILER
Page 12 of 12
4.03.1
FRI VOLUME 5: FRACTIONATION DESIGN HANDBOOK
CONDENSERS
Issued:
01/15/1997
4.03.2
Revised:
CONDENSERS
CONDENSERS ............................................................... ......................................................... .1 1.
Simple Shell and Tube Condensers ................................................... ................................. 2
2.
Internal Condensers ............................................................................................................3
3.
Barometric Condensers............................................... ........................................................4
4.
Pumparound Condensation ................................................................................................. 4
5. Air Cooled Exchangers .................................................. ......................................................4
Page 1 of 8
Issued: 01/15/1997 Revised:
1
CONDENSERS
4.03.2
Simple Shell and Tube Condensers
The most common condenser is the water-cooled fixed tube sheet type, in which water flows through the tubes and condensing takes place in the shell. If the cold medium is not cooling water, but some cleaner fluid, a U-Tube bundle may be used. Condensation almost always takes place in the shell for several reasons. Condensate drains away from the outer surfaces of tubes more completely than from the inside surfaces, offering more effective heat transfer. Pressure drop is generally less on the shell side. Also, the condensing vapor is a clean fluid which is best routed to the side which is difficult to clean. However, if the condensate is corrosive or if the vapor is at high pressure, it may be more economical to condense in the tubes. In this case, especially if the condensate is also viscous, a condenser may be vertical or sloped 10 to 15 degrees from the horizontal to provide better condensate drainage. Some authorities claim that, in general, condensing coefficients will be higher in horizontal condensers than in vertical ones. From experience it has been found that an advantage for one orientation over the other cannot be predicted in advance; two designs must be proposed and heat transfer coefficients calculated before a decision, which is based on heat transfer, can be made. Consequently, the initial selection is usually based on layout and maintenance considerations. This selection may be checked later and changed if heat transfer is markedly better in the alternate setting. Allowable pressure drop across a condenser is typically taken as 10% of the pressure of the entering vapor up to about 5 psi. To reduce pressure drop, shells can be made shorter with larger diameters, or arrangements are made to divide the flow. If condensing takes place over a wide range of temperature, 75°F or more, the shell may require a longitudinal baffle to keep condensate in contact with condensing vapor. This prevents flashing in the accumulator. Vapor flow in a condenser shell must be from side-to-side, since condensate will be trapped between up-anddown baffles. Subcooling occurs in condensers either by design or by accident. When a condenser is overdesigned for the service it is performing some subcooling will occur, even if the design did not call for subcooling. This must be understood and its effects on the operation of the column considered. Some condensers are designed to have subcooling. It may be done to reduce the temperature of the overhead product to better meet the needs of the next processing step, to better satisfy the NPSH needs of the reflux pump, or as part of the column control strategy, or to reduce the load on a refrigerated vent condensation system. Consideration as to whether it is the liquid, the vapor, or both that requires subcooling needs to be considered, and the system designed accordingly. Heat transfer from the condensate in a condenser to the cooling medium will be poor because it moves at low velocity past the tube surface. Therefore, if more than 20% of the total duty is in subcooling, a separate cooler, which will be designed for liquid-liquid heat transfer should be considered. If condensate does not discharge into an accumulator, the shell may need to be continuously vented to prevent accumulation of noncondensable gases. These will impair heat transfer, since condensate vapor must diffuse through them to reach a cold surface. The condensing coefficient for steam, for example, will be reduced about 50% by the presence of only 1 volume percent of air. The potential presence of solids may mandate the use of vertical, tube side condensing, units. Condensers may be located above or below their associated reflux drums. Page 2 of 8
In the latter case particular
Issued: 01/15/1997 Revised:
CONDENSERS
4.03.2
attention must be paid to the pressure balance in the system and the method of process control. condensation system is often the prime means of column pressure control.
The
Experience has shown that many condenser problems can be traced to poor drainage or venting, or inadequate cleaning.
2
Internal Condensers
Internal condensers can be mounted vertically or horizontally above the top tray (or above the top of the packing) and horizontally between trays. The simplest and most common arrangements are the generally axially mounted "knock back" configurations shown in Figures 1 and 2. As shown, knock back condensers are vertical units with condensing on either shell or tube side. Knock back condensers are used in primary and vent condenser service and may be mounted directly on a tower or reflux drum. A key feature of the knock back design is that the liquid and vapor are flowing in opposite directions. Depending on the configuration this may or may not be an important factor in the design. Advantages of the fixed tube sheet (tube side condensing) design shown in Figure 1 include that (1.) corrosive process fluids necessitate alloy construction of the tube side only; and, (2.) "counter current" flow (from a heat transfer point of view) of the external coolant and internal process stream is possible. A potentially important disadvantage relates to the counter-current flow of liquid and vapor within the tubes. With the smaller diameter and longer tubes that maximize heat transfer performance and minimize cost, the velocity of the vapor within the tubes could be such that liquid would be entrained upward instead of draining downward, as is desired. This often means that more tubes (or larger diameter tubes) than would otherwise be required must be installed. A major advantage of the U-tube design shown in Figure 2 is that expansion joints are never required. A disadvantage is that with U-tube construction water side drainage requires that the exchanger be removed. Spiral exchangers, because of their compactness, are sometimes preferred in knock back service. However, spiral exchangers are limited in size, available pressure ratings, and by gasketing requirements. If, because they are internal, knock back and other internal exchangers are not provided with surge their use limits control options and decreases column stability. Reflux cannot be measured. If surge is provided the tendency is to provide it inside the column. This requires extra careful attention to the physical arrangement and its effects on the hydraulics and heat transfer characteristics of the system. There are no general standards in this area. Almost every application is unique and how it is handled depends on the reason an internal exchanger was chosen. Two other internal condenser arrangements are shown in Figures 3 and 4. Both could be considered when pressure drop is a critical criterion. Each would have sub options depending on whether the overhead product is vapor, liquid, or both. In Figure 3 horizontal U-tube design special baffling may be required to achieve adequate heat transfer. Internal support could be an important consideration. In Figure 4 special baffling might be required to keep falling liquid out of the vapor product. Figures 5 and 6 illustrate a problem one member experienced. The system in Figure 5 was not performing. Condensate was being blown against the vessel walls and running down the walls, thereby bypassing the collecting pan. The arrangement in Figure 6 solved the problem.
Page 3 of 8
Issued: 01/15/1997 Revised:
3
CONDENSERS
4.03.2
Barometric Condensers
Barometric condensers are direct contact exchangers which find greatest application as steam jet ejector pre-, inter-, and after-coolers. They are less expensive than surface (Shell and Tube) exchangers but have the following disadvantages.
4
1.
Condensables are mixed with the cooling water and generally lost.
2.
Cooling water is generally on a once through basis and makeup requirements are high.
3.
Barometric condenser liquid discharges are often a waste disposal problem
Pumparound Condensation
Pumparound condensation is most commonly found intermediate in a column and is discussed in greater detail in Section 4.03.3. When used for overhead condensation the same general hydraulic and heat and mass transfer considerations apply except that when used for the overhead the condensation is generally more nearly total. There are some special situations in which overhead pumparound condensation is used, but the most common use is to reduce pressure drop in the overhead system and/or reduce the size of overhead vapor pipework, a situation most likely to be encountered in vacuum systems. Since overheads can be condensed traditionally or by pumparound it is generally economic considerations that determine when pumparound condensation is used, although space, per se, is sometimes the driver. Pumparound condensation saves overhead vapor piping and/or pressure drop and often results in a smaller exchanger. On the other hand, it requires a little taller tower, more and generally larger liquid piping, and a much larger liquid pumping system. Pumparound condensation is more commonly found in refinery services. A total drawoff tray is generally provided to allow direct measurement of reflux.
5
Air Cooled Exchangers
An air cooled heat exchanger consists of a horizontal or inverted "V" shaped bundle of tubes which is cooled by ambient air. The air may be blown across the tubes (forced draft) or sucked across the tubes (induced draft) by a fan. The tubes are generally finned. Air cooled exchangers would replace water cooled exchangers where cooling water is relatively expensive. They also eliminate the corrosion and fouling problems associated with cooling water. Variable pitch fans make air coolers relatively easy to control. Leaks are easy to detect and some cooling is available, due to natural circulation, upon power failure. On the other hand, air cooled exchangers are more expensive than water cooled units. They require a structure and they have the various operational ills inherent in mechanical equipment. Also, air coolers are very sensitive to weather conditions, a major potential problem in many installations.
Page 4 of 8
Issued: 01/15/1997 Revised:
CONDENSERS
Figure 1. TUBE SIDE KNOCKBACK CONDENSATION
Figure 2. SHELL SIDE KNOCKBACK CONDENSATION
Page 5 of 8
4.03.2
Issued: 01/15/1997 Revised:
CONDENSERS
Figure 3. INTERNAL CONDENSERS, HORIZONTAL
Page 6 of 8
4.03.2
Issued: 01/15/1997 Revised:
CONDENSERS
Figure 4. INTERNAL CONDENSER, VERTICAL
Page 7 of 8
4.03.2
Issued: 01/15/1997 Revised:
CONDENSERS
Figure 5. KNOCKBACK CONDENSER, INCORRECT LIQUID COLLECTION
Figure 6. KNOCKBACK CONDENSER, CORRECT LIQUID COLLECTION
Page 8 of 8
4.03.2
FRI VOLUME 5: FRACTIONATION DESIGN HANDBOOK
VORTEX BREAKERS
Issued:
02/15/1993
4.04
Revised:
VORTEX BREAKERS
VORTEX BREAKERS .................................................................... ......................................... 1 1.
General ................................................................................................................................ 2
2.
Bottom Outlets .................................................................................................................... 2
3.
Side Outlets.............................................................................................................. ........... 2
Page 1 of 5
Issued: 02/15/1993 Revised:
1
VORTEX BREAK ERS
4.04
General
Vortex breakers are layers of metal grating or flat plates arranged in a cross shape located near liquid outlets to prevent vortex formation. Conservation of angular momentum as the liquid converges toward an outlet nozzle can lead to an intensification of a swirling motion creating a vortex. Vortexing promotes the entrainment of vapor into the liquid drawoff line creating potential drawoff rate limitations, tower instability, or pump cavitation (49). The vortex breaker introduces shear in the vicinity of the outlet nozzle suppressing the swirling motion. Conditions reported to enhance the potential for vortex formation are often present in distillation columns. These include: low liquid level relative to vessel diameter, liquids near their bubble point, high drawoff velocities (>3 ft/s, .9m/s), and swirling induced by tangential vapor and/or liquid inlets. The circular symmetry present at the bottom of most towers also favors vortex formation. For this reason, vortex breakers are highly recommended for bottoms outlets. The highly non-circular geometry of most intermediate liquid drawoffs tends to make them less prone to vortex formation. Accordingly, vortex breakers are less often specified for side outlets. However, vortex formation can occur regardless of the outlet nozzle orientation. One published model suggests vortexing can occur whenever the liquid depth (feet) is less than the nozzle velocity (feet/sec) minus one [H < V-1] (50). The relatively small additional cost of a vortex breaker prompts many designers to specify their use at all liquid drawoff locations. Apart from avoiding the uncertainties of predicting vortex formation, this approach offers greater flexibility for unplanned operations or future expansion. Two basic designs of vortex breakers are in use in the industry today: the cross (or vane) type and the grating type. The cross type is the simpler design, being made from two steel plates welded together at a 90° angle. It is more commonly used for tower bottom outlets but can also be used for side draws. The grating style is generally considered the more effective design (51). It is composed of one to three horizontal layers of one inch vertical bars spaced one inch apart (commonly known as floor grating). Sections of grid packing have also been used. The grating style is widely used for both bottom outlets and side draws.
2
Bottom Outlets Figure 1 illustrates the use of both types of vortex breaker for bottom outlets. The cross type is normally made from heavy steel plate (>0.375 inches thick or 9.5mm thick). A circular or square horizontal cover plate can be optionally welded to the top of the cross.
The height of the cross should be equal to the nozzle diameter but no less than 5 inches (125mm) tall. The width of the cross plates should be four times the nozzle diameter. To avoid restricting flow at the nozzle inlet, the bottom of the cross is spaced half the nozzle diameter above the exit nozzle (52). The cross is typically welded directly to the vessel head. To protect bottom outlets using the grating style, 2-3 layers of grating should be used. The layers are spaced three inches apart and rotated 90° to one another. The bottom layer should be spaced half the nozzle diameter above the nozzle. The square grating sections should be four nozzle diameters in width.
3
Side Outlets Figure 2 illustrates the use of vortex breakers for side outlets.
Page 2 of 5
Issued: 02/15/1993 Revised:
VORTEX BREAK ERS
4.04
A single layer is most often used for the grating style. Additional layers can also be used and are stacked above the locations indicated for the single layer in Figure 2. Adjacent layers are typically rotated 90° from one another. The length of the grating sections is four times the outlet nozzle diameter. The width conforms to the shape at the mouth of the exit sump. In either the partial or total draw configuration, the cross type is located at the nozzle entrance. The size of the cross will be limited by the dimensions of the draw sump. Typically, the cross will be the same size as the nozzle
The Following Unreferenced Publications Will Also Be Found To Be Informative And Helpful:
e.
FRI Design Practices Manual, Volume 5, Section 1.05.
f.
McDuffie, N.G., "Vortex Free Downflow in Vertical Drains", AIChE Journal, Vol. 23, No. 1, January 1977.
Page 3 of 5
Issued: 02/15/1993 Revised:
VORTEX BREAK ERS
Figure 1. VORTEX BREAKERS FOR BOTTOM OUTLETS
Page 4 of 5
4.04
Issued: 02/15/1993 Revised:
VORTEX BREAK ERS
Figure 2. VORTEX BREAKERS FOR SIDE OUTLETS
Page 5 of 5
4.04
FRI VOLUME 5: FRACTIONATION DESIGN HANDBOOK
CASE STUDY #1 – PACKED COLUMN TROUBLESHOOTING
Issued:
10/01/2006
4.50.01
Revised:
CASE STUDY #1 – PACKED COLUMN TROUBLESHOOTING
CASE STUDY #1 – PACKED COLUMN TROUBLESHOOTING..................... ................... 1 1.
Summary .................................................. ........................................................................... 2
2.
Background ............................................... .......................................................................... 2
3.
Problem ............................................................................................................................... 2
4.
Troubleshooting ............................................... ................................................................... 2
5.
Resolution ................................................ ........................................................................... 2
Page 1 of 3
Issued: 10/01/2006 Revised:
1
CASE STUDY #1 – PACKED COLUMN TROUBLESHOOTING
4.50.01
Summary After commissioning, a new glycol purification train could not achieve design capacity. Throughput had to be reduced to 70% of design to enable products to be produced on specification. There were signs that multiple packed columns were experiencing premature flooding. The use of FRI data played an important part in the resulting troubleshooting effort. Following corrective actions the unit was able to operate at design capacity and meet all product specifications.
2
Background Glycol purification units normally produce a range of glycols using a sequential fractionation sequence. These columns are normally packed since they operate under vacuum. A mix of sheet metal and gauze type structured packing was used in this design with the gauze packing typically employed in the upper sections due to the lower liquid load required there.
3
Problem After startup it became apparent that several of the columns suffered from premature flooding. Pressure drop measurements across the packed sections above the feed location confirmed excessive pressure drop indicative of flooding. Gamma scanning confirmed flooding above the feed distributor and below the top distributor. In most cases these sections were equipped with gauze (BX type) structured packing. It appeared that both the vendor and in-house predictions of the packing capacity were too optimistic. Since the service was clean, fouling of the gravity distributors was not expected.
4
Troubleshooting In order to address to apparent discrepancy between the expected and maximum sustainable column capacity, attention was focused on the hydraulic capacity of the gauze packing. During the design phase the vendor had provided data obtained from a 6" (0.15 m) pilot column. Data was also available from the 18" (0.46 m) column operated by the SRP. Suspicion fell on the scale-up characteristics of this type of packing. It was necessary to obtain independent, validated performance data for the packing measured on a commercial scale column. FRI had tested [1] BX-type gauze packing using the system o–xylene/p-xylene at 16, 100 and 300 mmHg absolute as well as atmospheric pressure using 1000, 250 and 70 mm ID columns. The FRI data can be conveniently arranged as a plot of maximum vapor capacity (C s, velocity units) versus flow parameter. By plotting against flow parameter it was possible to make comparisons to the glycol system since the different physical properties are adequately captured within the definition of the flow parameter. This comparison showed that the maximum capacities achieved in the glycol unit and that predicted by the vendor and inhouse methods were similar to that measured by FRI. Packing efficiencies were substantially lower though. As a result of this important conclusion it became clear that it was necessary to shut down the unit and inspect the internals.
5
Resolution Inspection of the column internals showed no fouling, also the gravity distributors were still level. However, Page 2 of 3
Issued: 10/01/2006
CASE STUDY #1 – PACKED COLUMN
Revised:
TROUBLESHOOTING
4.50.01
it was clear from visual observation that several of the distributors had overflowed. Further investigation revealed that the feed and reflux distributors had been swapped during installation (they had similar orifice sizes but different hole counts). This resulted in overflowing of the feed distributor and poor distribution quality to the top bed. After the distributors were moved to their correct locations, the unit performed as per design, both in terms of capacity and product quality. The use of the FRI data played a crucial part in the design verification analysis and the subsequent justification to shut down the unit for inspection
References and Further Readings [1] Plant test Report: PTR 22 2/15/72 Tests of 1000, 250 and 70 Millimeter Diameter Columns with Koch Sulzer Packing
Page 3 of 3
FRI VOLUME 5: FRA CTIONATION DESIGN HANDBOOK
CASE STUDY #2 – AN INTERMEDIATE LIQUID DRAWOFF PROBLEM
Issued: Revised:
10/01/2006
4.50.02
CASE STUDY #2 – AN INTERMEDIATE LIQUID DRAWOFF PROBLEM
CASE STUDY #2 – AN INTERMEDIATE LIQUID DRAWOFF PROBLEM ...................... 1 1.
Type of Unit.................................................... ....................................................................2
2.
Problem Experienced ......................................................................................................... .2
3.
Problem Cause ....................................................................................................... .............2
4.
Review of History and Operation ....................................................................................... 2
5.
Troubleshooting Techniques Used ....................................................... .............................. 2
6.
Recommended Solution and/or Action Taken..................................................... ............... 2
7.
Results .................................................... ........................................................ .................... 2
8.
Lessons Learned ............................................................ .................................................... .3
Page 1 of 4
Issued: 10/01/06 Revised:
1
CASE STUDY #2 – A N INTERMEDIATE LIQUID DRAWOFF PROBLEM
4.50.02
Type of Unit
Refinery Fluidized Catalytic Cracking Unit (FCCU) Absorber-Deethanizer Tower (operates at 120 psig (17.4 kPa(g))).
2
Problem Experienced
The tower has an upper absorber section to absorb C3 and heavier components (C3+) and a bottom deethanizer section to strip C2 and lighter components from FCCU naphtha. An intercooler pumparound circuit is provided in the top section to enhance the absorption of the C3+ components. Limited intercooler rates due to cavitation of the circulation pumps were experienced after the tower was replaced with a new design.
3
Problem Cause
Hydraulic limitations.
4
Review of History and Operation
Problem developed after the tower had been replaced. The replacement tower had the drawoff designed as shown in the attached sketch identified as "Existing". The two intercooler drawoffs remained at the bottom of the center downcomer. Comparison to the prior design revealed that: 1) the outlet weir had been lowered, 2) the inlet downcomer clearance had increased, and 3) the tower diameter had increased. All three would reduce the clear liquid height in the downcomer containing the drawoff nozzles. In addition, the tray spacing above the drawoff had increased allowing the liquid falling into the downcomer to create more turbulence.
5
Troubleshooting Techniques Used • • •
6
Analysis of unit history and operation Visual inspection and records of the tower internals designs Hydraulic calculations.
Recommended Solution and/or Action Taken
The inlet downcomer was shortened and a seal pan was installed above the drawoff to allow any vapor in the downcomer to disengage. Flow baffles were relocated to maintain separation between the hot drawoff liquid and the two 8" diameter distributors for the cooled return stream. The modified drawoff is shown in the attached sketch identified as "Redesigned - Seal Pan Option".
7
Results
Modifications to the intercooler drawoff tray during the next turnaround provided higher intercooler rates for C3+ recovery. Prior problems with pump vibration limited intercooler rates to approximately 20 kB/D (132 m3/h). This circulation rate corresponds to a liquid velocity of 1.5 ft/sec (0.46 m/s) (typical of a drawoff Page 2 of 4
Issued: 10/01/06 Revised:
CASE STUDY #2 – A N INTERMEDIATE LIQUID DRAWOFF PROBLEM
4.50.02
imited by self-venting flow) in the two 10" diameter drawoff nozzles. Modifications to the drawoff tray reduced the possibility of vapor carryunder into the pump suction line. The pump has then able to achieve circulation rates of 45 - 55 kB/D (298 - 364 m3/h) with all vibration readings well below pre-alarm levels. This circulation rate corresponds to a drawoff nozzle velocity of 3.4 - 4.1 ft/sec (1.0 - 1.2 m/s). Operating credits of $90k/yr for improved C3+ recovery were achieved.
8
Lessons Learned
The problem experienced with the original design of the intermediate liquid drawoff could have been avoided had the designer realized that pump suction drawoffs in some high pressure towers (such as absorberdeethanizers) should not be taken directly from downcomers. The potential for water condensation leading to foaming conditions aggravates a froth disengagement problem at a drawoff location. The guidelines in ection 1.05.1 of the FRI Design Manual should be followed to avoid intermediate liquid drawoff limitations, i.e., use a downcomer seal pan above a liquid drawoff sump to help disengage vapor from the liquid, especially for high pressure services and for liquid drawoffs to pump suctions.
Page 3 of 4
Issued: 10/01/06 Revised:
CASE STUDY #2 – A N INTERMEDIATE LIQUID DRAWOFF PROBLEM
Page 4 of 4
4.50.02
CASE STUDY #3 – HOW TO DESIGN AN FRI SMALL SCALE COLUMN
FRI VOLUME 5: FRACTIONATION DESIGN HANDBOOK Issued: Revised:
10/01/2006
4.50.03
CASE STUDY #3 – HOW TO DESIGN AN FRI SMALL SCALE COLUMN
CASE STUDY #3 – HOW TO DESIGN AN FRI SMALL SCALE COLUMN ...................... 1 1. Internals Design ................................................... ................................................................ 3 2. Economic Consideration............................................ ................................................... .......3
Page 1 of 9
Issued: 10/01/06 Revised:
CASE STUDY #3 – HOW TO DESIGN AN FRI SMALL SCALE COLUMN
4.50.03
A small deisopentanizer column is required to better stabilize the condensate mix in a LNG Fractionation Unit. The tower is simulated with 14 stages with 2 additional stages for the condenser and reboiler. A summary of the internal vapor/liquid traffic and physical properties is summarized in Table 1 & 2. The tower operates at 37.7 psia (2.6 bara) and 143°F (61.5°C). A flashing feed is introduced to the column on stage 8. The reboiler is a kettle type and a total condenser provides liquid reflux to the top tray. The new column was generously sized at 850mm in diameter (33.5”) at less than 72% FRI flood per Figure 1. An interim flood calculation was evaluated on a column diameter of 800mm (31.5”). The smaller column diameter produced a flood value of 81.5%. Although this percent of flood is acceptable, the relatively small economic benefit of the smaller column does not justify the loss of upward capacity and flexibility that the larger 850mm diameter column provides.
TABLE 1 – Internal Tower Vapor/Liquid Loadings Tray Number Condenser 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Reboiler
Vapor Okg/h 8150 kg/h 8420 kg/h 8400 kg/h 8370 kg/h 8330 kg/h 8260 kg/h 8140 kg/h 7890 kg/h 3213 kg/h 3240 kg/h 3268 kg/h 3295 kg/h 3323 kg/h 3350 kg/h 3364 kg/h
Liquid 5145 kg/h 5415 kg/h 5395 kg/h 5370 kg/h 5325 kg/h 5255 kg/h 5135 kg/h 4884 kg/h 17020 kg/h 17050 kg/h 17080 kg/h 17105 kg/h 17135 kg/h 17160 kg/h 17175 kg/h 13810 kg/h
The column capacity is being controlled by the top section. The bottom section was rated at only 33.3% flood. Thus, a larger size packing of 1.5 Nutter Rings (NR) was chosen for the top, while a smaller, 1.0 NR was selected for the bottom. Structured packing could also be used, but due to operational experience, random packing was selected. Based on previous operating experience, the tower liquid hold-up was sized for a residence time of 3 minutes between the LLL (Low Liquid Level) and HLL (High Liquid Level), with an additional 1 minute residence time between the LLL and the emergency shut down LLLL (Low Low Liquid Level). The residence time added significant height to the vessel with a diameter of 850mm. To reduce column height and assist with mechanical skirt design, a larger bottom 1400mm (55”) diameter section was specified. TABLE 2 – Physical Properties
Stage
MW
Vapor Density
Condenser
71.80
5.957 kg/m3
7.357e-003 cP
72.03
587.3 kg/m3
0.1672 cP
11.58 dyne/cm
1 2 3 4 5 6 7 8 9 10 11 12 13 14 Reboiler
72.03 72.12 72.22 72.35 72.59 73.01 73.75 75.13 75.13 75.13 75.14 75.15 75.18 75.30 76.13
7.306 kg/m3 7.349 kg/m3 7.391 kg/m3 7.432 kg/m3 7.471 kg/m3 7.505 kg/m3 7.526 kg/m3 7.512 kg/m3 7.559 kg/m3 7.606 kg/me 7.654 kg/m3 7.701 kg/m3 7.749 kg/m3 7.797 kg/m3 7.794 kg/m3
7.569e-003 cP 7.578e-003 cP 7.591e-003 cP 7.610e-003 cP 7.637e-003 cP 7.680e-003 cP 7.753e-003 cP 7.899e-003 cP 7.906e-003 cP 7.913e-003 cP 7.919e-003 cP 7.926e-003 cP 7.933e-003 cP 7.945e-003 cP 7.995e-003 cP
72.18 72.32 72.54 72.91 73.58 74.80 77.17 82.80 82.78 82.77 82.76 82.75 82.77 82.96 84.81
578.5 kg/m3 578.4 kg/m3 578.8 kg/m3 579.7 kg/m3 581.7 kg/m3 585.3 kg/m3 591.9 kg/m3 603.3 kg/m3 603.0 kg/m3 602.7 kg/m3 602.4 kg/m3 602.1 kg/m3 601.8 kg/m3 601.8 kg/m3 604.6 kg/m3
0.1567 cP 0.1566 cP 0.1569 cP 0.1577 cP 0.1592 cP 0.1623 cP 0.1683 cP 0.1809 cP 0.1805 cP 0.1801 cP 0.1797 cP 0.1793 cP 0.1789 cP 0.1787 cP 0.1814 cP
10.81 dyne/cm 10.80 dyne/cm 10.81 dyne/cm 10.83 dyne/cm 10.89 dyne/cm 11.01 dyne/cm 11.21 dyne/cm 11.55 dyne/cm 11.52 dyne/cm 11.49 dyne/cm 11.47 dyne/cm 11.44 dyne/cm 11.42 dyne/cm 11.40 dyne/cm 11.46 dyne/cm
Viscosity
MW
Page 2 of 9
Density
Liquid Viscosity
Surface Tension
Issued: 10/01/06 Revised:
CASE STUDY #3 – HOW TO DESIGN AN FRI SMALL SCALE COLUMN
4.50.03
The calculated HETP for the top section was 585mm (23”) and 332mm (13”) for the bottom. A safety factor of 15 to 20% is typically suggested for new applications. Based on a 15% safety factor, the resultant bed heights are 4750mm (15.5 ft) for the top section and 3000mm (10ft) for the bottom section. Please refer to the FRI Packed Tower Program rating sheets at the end of this document.
1
Internals Design A tower layout was developed per Figure 1. The internals were designed and selected per the recommendations developed in this design practice. Handholds were specified at the bottom of each bed to assist with removal of packing without having to dismantle the bottom body flanges. A total of 3 body flanges are needed to facilitate installation of the internals. The nozzles were sized per design practice guidelines. The reflux nozzle was sized for a maximum velocity of 3 ft/s. (1 m/s), to minimize the inlet momentum to the distributor. The mixed phase & the vapor return was sized to limit the inlet momentum to between 3000-4000 lb/ft s 2 (4000-6000 kg/m s 2) as per section 1.06. A pan distributor was selected in order to allow easy leveling upon installation. This was deemed necessary for the low to medium liquid rate of 6.75 gpm/ft 2 (102 m3/h/m2). The bottom bed liquid rate is much higher at 20.6 gpm/ft2 (310 m3/h/m2). A pan distributor was also selected for the bottom bed, to reduce gasketing, and provide a means for leveling. An orifice plate distributor would also be a good selection, but it would require a continuous support ring with gasketing. This is a very difficult task to perform in a small diameter column. A flashbox configuration was selected for the mixed phase inlet based on previous operating experience for this unit. A flash gallery is another alternative internal. A distance of 450mm (18”) was allowed between the top of the flashbox hat to the support plate for good vapor distribution. The flashbox should be designed with ample vapor free space for vapor/liquid disengagement. When designing the liquid distributor and the downpipe exiting the flashbox, one needs to accommodate for a liquid that is not completely vapor free. The liquid distributor should have sufficient space between the maximum clear liquid level and the top of the vapor risers. The downpipe should be designed for self-venting flow. Note that a 200mm (8”) vessel nozzle is needed to accommodate the 150mm (6”) mixed phase distributor. A total drawoff chimney tray is provided at the bottom of the tower to provide liquid feed to the reboiler. This tray is completely seal welded to avoid leakage. The drawoff nozzle is sized as self-venting per the equation found in Section 1.05.1. The risers provide sufficient pressure drop to ensure good vapor distribution, even though the vapor reboiler return nozzle from the kettle reboiler is flush. The chimney tray and drawoff nozzle are located in the transition of the cone section to save vessel height. When locating internals, the designer must always consider mechanical restrictions, such as the minimum distance between the bottom liquid level tap and bottom head, location of drawoff nozzle N4 with transition cone knuckle, and location of nozzles and internals to weld seams. Consultation or review by both a vessel and materials specialist are recommended prior to release for purchase. A grating type vortex breaker is specified in the bottom of the column as a precaution against vapor breakthrough. The liquid bottom outlet was sized for a liquid velocity of less than 3 ft/s (1m/s). At normal operating conditions, the liquid should not vortex, even at LLL.
2
Economic Consideration Although this is a clean service with very good operational history, operations were concerned with maintenance and vessel inspection of the tower. To inspect any of the feed points, the vessel must be broken Page 3 of 9
Issued: 10/01/06 Revised:
CASE STUDY #3 – HOW TO DESIGN AN FRI SMALL SCALE COLUMN
4.50.03
down to its 3 sections. While this is typical of a small tower, operations were not happy with the prospect of dismantling the tower. A crane would be required, and more man-hours would be needed. An economic study was initiated, to compare the cost of a slightly larger tower to accommodate conventional vessel manholes versus the smaller flanged-body column. The study showed that it was more economical to increase the tower diameter to 1000mm (40”) and use trays at a tray spacing of 450mm (18”). The 1000mm diameter column can now accommodate a more conventional 450mm (18”) nominal manhole. For the packed tower version shown in Figure 1, the cost of the packing and the body flanges made this design uneconomical, even though it resulted in a smaller tower diameter. A larger difference in tower diameter, larger operating pressure, and more theoretical trays would make the packed tower version more economical. This case study shows that if the tower diameter is 30” (750mm) or greater, with low operating pressure, a tray option should be considered.
Figure 1. CASE STUDY #3 DEISOPENTANIZER ELEVATION Nozzle Schedule Nozzle.
Size Diameter
Qty.
Description
N1 N2
6” (150 mm) 3” (75 mm)
1 1
Vapor Outlet Reflux
N3
6” (150 mm) 8” (200 mm)
1 1
Mixed Phase Feed Sparger Vessel Nozzle
N4
6” (150 mm)
1
Reboiler Draw
N5
4” (100 mm)
1
Vapor Reboiler Return
N6 N7 N8
6” (150 mm) 6” (150 mm) 3” (75 mm)
1 1 1
Liquid Reboiler Return Liquid Product Outlet Drain
N9
3” (75 mm)
1
Utility
N10 N11 T1
2” (50 mm) 2” (50 mm) 2” (50 mm)
1 1 1
Min. Pump Back Return Vent TW Connection
P1-3 L3 L4 L1, L2 M1 M2A-B
2” (50 mm) 2” (50 mm) 2” (50 mm) 24” (600 mm) 10” (250 mm)
3 2 2 1 2
PI/PDT Connection LT Connection (not shown) LG/LT (not shown) Manhole Handhole
Page 4 of 9
Issued: 10/01/06 Revised:
CASE STUDY #3 – HOW TO DESIGN AN FRI SMALL SCALE COLUMN
Page 5 of 9
4.50.03
Issued: 10/01/06 Revised:
CASE STUDY #3 – HOW TO DESIGN AN FRI SMALL SCALE COLUMN
4.50.03
FRI Packed Column Rating Detailed Output Case Identification Project: File Path:
Top Version 1.3.0 (5/21/2006 9:02:59 PM) Case Study Deisopentanizer ‐ Top
Maximum Rates: System Limit: Flood (Constant L/V) Flood (Constant Liquid) Useful Capacity Pressure Drop/Unit Length HETP
% % % % mmH2O/m mm
45.0% 72.1% 71.5% 77.9% 23.6 584.9
Settings: Packing Type Packing Name Packing Material Input Units Output Units
Design Loads:
Correlations Used
Vapor Volumetric Rate Vapor Mass Flow Rate Vapor Load (VLD) Superficial Capacity Factor
m3/s kg/h m3/s m/s
0.318 8,420.0 0.036 0.0636
System Limit Jet Flood Pressure Drop Efficiency
Fs, Factor Vapor Superficial Velocity Liquid Mass Flow Rate Liquid Volumetric Flux Liquid Superficial Velocity
m/s(kg/m3)^0.5
1.520 0.561 5,415.0 16.50 0.0046
IDs
Liquid Load (LLoad) Flow Parameter
Random NR1.5 Stainless Steel Metric Metric
m/s kg/h m3/h/m2 m/s m3/s
TR 136 TR 147 TR 147 TR 152
Process Hardware Output Model
Top Pack_D Top RPMetal
No. of Warnings and Messages:
0.0026 0.072
Calculation Messages Input Warnings
0 0
Tray Efficiency: HTUOV HTUV HDUV HTUL
mm mm mm mm
748.7 245.6 5.5 352.2
0.1
System Limit Capacity Factor: V Load Capacity Design as % Capacity
m3/s m/s %
0.08 0.1414 45.0%
Pressure Drops: Static Pressure Drop Dry Pressure Drop Dynamic Pressure Drop Dynamic Press. Drop/Stage Bed Dynamic Pressure Drop Liquid Holdup (Parameter)
mm H2O/m mm H2O/m mm H2O/m mm H2O mm H2O
23.6 13.8 109.8 0.033
Constant L/V Flood Capacity: V Load Capacity Design as % Capacity
ft3/s m/s %
0.05 0.0882 72.1%
Constant Liquid Flood Capacity: Flooding Safety Factors & Margins: VLD (Sys Limit)/VLD(Design) VLD (Const L/V Flood)/VLD (Design) VLD (Const Liq Flood)/VLD (Design)
2.22 1.39 1.40
Page 6 of 9
V Load Capacity Design as % Capacity
m3/s m/s %
0.05 0.0889 71.5%
Issued: 10/01/06 Revised:
CASE STUDY #3 – HOW TO DESIGN AN FRI SMALL SCALE COLUMN
4.50.03
FRI Packed Column Rating Detailed Input Process Properties Process Identification:
Top
Vapor
Liquid
Scale Factor Scaled Mass Flow Rate
1.000 kg/h
8,420.0
Scale Factor Scaled Mass Flow Rate
1.000 kg/h
5,415.0
Density
kg/m3
7.349
Density
kg/m3
579
Volumetric Flow Rate
m3/s
0.318
Volumetric Flow Rate
m3/h
9.36
cP
0.0076
cP
0.157
Surface Tension
dyne/cm
10.81
Molecular Weight
kg/kg mol
72.2
Viscosity Molecular Weight
kg/kg mol
72.1
Viscosity
Efficiency Parameters Vapor Diffusivity
cm2/s
Vapor Schumidt Number Liquid Diffusivity
0.0149 0.692
cm2/s
Liquid Schmidt Number
6.4410E ‐05 42.0558
Lambda (m V/L)
1.547
m‐Value
0.994
Hardware Details Hardware Identification:
Pack_D
Column Geometry: Diameter Packed Bed Height Cross‐Sectional Area
mm
850
m
4.65
m2
0.567
Random Packing Dimensions: Diameter Specific Area
Structured Packing Dimensions: mm
38.00
Block Height
mm
m2/m3
124.7
Specific Area
m2/m3
0.978
Void Fraction
Void Fraction Wall Thickness
mm
Crimp Angle
Page 7 of 9
°
Issued: 10/01/06 Revised:
CASE STUDY #3 – HOW TO DESIGN AN FRI SMALL SCALE COLUMN
4.50.03
FRI Packed Column Rating Detailed Output Case Identification
Bottom
Project:
Version 1.3.0 (5/21/2006 9:10:26 PM)
File Path:
Case Study Deisopentanizer ‐ Bottom
Maximum Rates: System Limit: Flood (Constant L/V) Flood (Constant Liquid) Useful Capacity Pressure Drop/Unit Length HETP
% % % % mmH2O/m mm
17.0% 48.0% 33.1% 51.2% 6.1 332.3
Settings: Packing Type Packing Name Packing Material Input Units Output Units
Design Loads:
Correlations Used
Vapor Volumetric Rate Vapor Mass Flow Rate Vapor Load (VLD) Superficial Capacity Factor
m3/s kg/h m3/s m/s
Fs, Factor Vapor Superficial Velocity Liquid Mass Flow Rate
m/s(kg/m3)^0.5
Liquid Volumetric Flux Liquid Superficial Velocity
m3/h/m2 m/s m3/s
Liquid Load (LLoad) Flow Parameter
Random NR1.0 Stainless Steel Metric Metric
m/s kg/h
0.120 3,364.0 0.014 0.0242
System Limit Jet Flood Pressure Drop Efficiency
TR 136 TR 147 TR 147 TR 152
IDs
0.590 0.211 17,175.0
Process Hardware
Bottom Pack_E
50.29 0.0140
Output Model
Bottom RPMetal
0.0079 0.581
No. of Warnings and Messages: Calculation Messages Input Warnings
0 1
Tray Efficiency: HTUOV HTUV HDUV HTUL
mm mm mm mm
166.8 95.3 33.2 317.1
System Limit Capacity Factor: V Load Capacity Design as % Capacity
m3/s m/s %
0.081 0.1422 17.0%
Pressure Drops: Static Pressure Drop
mm H2O/m
Dry Pressure Drop Dynamic Pressure Drop
mm H2O/m mm H2O/m
Dynamic Press. Drop/Stage Bed Dynamic Pressure Drop Liquid Holdup (Parameter)
mm H2O mm H2O
0.1 6.1 2.0 18.3 0.087
Constant L/V Flood Capacity: V Load Capacity Design as % Capacity
ft3/s m/s
0.029 0.0505
%
48.0%
Constant Liquid Flood Capacity: Flooding Safety Factors & Margins: VLD (Sys Limit)/VLD(Design) VLD (Const L/V Flood)/VLD (Design) VLD (Const Liq Flood)/VLD (Design)
5.87 2.08 3.02
Page 8 of 9
V Load Capacity Design as % Capacity
m3/s m/s %
0.041 0.073 33.1%