Page 180 - Gas Purification 5E

..llkanoIamines for Hydrogen Sulfide and carbon Dioxide Removal 167

the packing should be wetted by the amine phase fist when the plant is initially commissioned
or after a turnaround (Tse and Santos, 1993: DUPart and Marchant, 1989; Strigle, 1994.j.
According to Honerkamp (1975), typical random packing sizes for LPG treating are 1 to
1!4 in. with 1 in. being the most common size. Tse and Santos (1993) and Perry (1977Aj
state that the most common packing sizes for LPG treating are l!< in. to 2 in. Smgle (1994)
states that for all liquid-liquid treating applications the maximum packing size should be 2
in. with 1% in. the more common choice. Tse and Santos (1993) and DuPart and Marchant
(1989) present an equation which indicates that random packing used for LPG treating
should be X in. or ,pater. Also, according to Tse and Santos (1993) and Strigle (1994), the
maximum packing size should be no greater than one-eighth the contactor diameter. All of
these criteria are normally satisfied by the use of 1- to 2-in. packing.
As the LPG flows up through a bed of random packing, the LPG droplets tend to coalesce
and mass transfer efficiency is lost. Accordingly, there is a maximum effective bed height. If
the packing height requirement exceeds the maximum effective bed height, it is necessary to
divide the bed into two or more separate packed sections. The. LPG flowing into the upper
bed is collected and redispersed to obtain good mass transfer. Honerkamp (1975) noted that
LPG contactors with 16 ft of random packing performed as well as contactors with 36 ft of
packing. While this suggests that individual bed heights should be limited to less than 16 ft,
Honerkamp recommended that individual beds of random packing be restricted to a maxi-
mum height of 8 ft for LPG treating service. Smgle (1994) recommends a maximum random
packing bed height of 12 ft for all liquid-liquid contacting applications. For LPG treaters,
Wart and Marchant (1989) also recommend that the maximum packing height per bed be
limited to no more than 12 ft, since deeper beds result in a loss of efficiency. While the rec-
ommendations differ somewhat, 12 ft is probably a reasonable estimate of the maximum
effective bed height for LPG treating with typical 1H to ?-in. random packing.
Each 8 to 12 ft bed of packing is approximately equivalent to one theoretical stage. When
morc than one theoretical stage is required, it is necessary to use multiple beds with effective
dispenser/support plates for each bed. Most liquid hydrocarbon treating requirements can be
accomplished with three or less theoretical stages; therefore. one to three beds of packing are
commonly used in commercial installations.

Cantactor Diunzem. It has been common practice in the industry to base the diameter of
packed LPG/amine contactors on the superficial velocity of the two liquid phases combined
together. There is considerable variation in design superficial velocity recommendations.
Perry (1977A, B) recommends sizing LPG contactors based on 15 gpm/ft2. Perry's recom-
mendation, which is based on natural gas plant liquids, is supported by DuPart and Marchant
(1989). Russell (1980) suggests a superficial velocity of 12 gpmlft', while Changela and
Root (1986) and Veldman (1989) recommend 10 gpm/ft2. The GPSA Engineering Data Book
(1987) recommends 20 gpm/ft'. A design superficial velocity of 15 gpm/ft2 is probably
appropriate for natural gas plants because there is usually a large difference between the den-
sity of the LPG and amine phases, the amine and the LPG are usually clean, and hydrocar-
bon liquid production typically declines with time. In a refinery, 10 gpm/ft' would seem to
be the proper choice because refinery capacity usually increases through a progressive series
of revamps, the amine is often contaminated with particulates and surface active compounds.
and the LPG usually has a higher gravip than gas plant liquids. In any case. design of a new
LPG treater should be based on a design superficial velocity of 15 gpm/ft2 or less.
The flooding correlation of Crawford and Wilke i1951j should be used with caution when
emhating LPG treaters. Strigle (1994) suggests that LPG treater design be based on 12% of