Page 183 - Gas Purification 5E

170 Gas PuriJcation

gravity (Rogers, 1994). The structured packing is provided as a single bed multiple beds are
required for random packing. A single bed is possible because disperser plates are located
between sections of structured packing in alternating stacks. By providing the packing as a
single bed, the required bed height and contactor tangent to tangent distance can be reduced
and more theoretical stages can be achieved for a given vessel height. Structured packing also
has significantly higher capacity than random packing. Recent designs have achieved superfi-
cial velocities of 55-60 gpdff based on the combined flow of both phases (Rogers, 1994).
These designs had the interface controller located in the bottom of the treater and amine was
the dispersed phase. At superficial velocities this high with LPG as the continuous phase, it
may be necessary to install coalescing pads on both the rich amine and treated LPG streams to
lit amine and LPG entrainment. Units revamped with structured packing are reported to
have achieved highly efficient H2S removal (Copper Strip test results: #1A) and to have met
pipeline C02 specifications (Rogers, 1994). Although it is claimed that structured packing is
more resistant to plugging than random packing, structured packing is more difficult to clean,
and full flow filtration of the lean amine entering the LPG contactor is recommended.
Sieve Tmys

While the vast majority of LPG treaters utilize random packing, sieve trays are occasional-
ly used. However, Honerkamp (1975), Fleming et al. (1988), and DuPart and Marchant
(1989) report that sieve trays are less efficient than random packing for LPG treating. Sieve
tray operating data for LPG treating are given in Tables 2-30 and 2-33. Table 2-33 provides
extensive data for three different treating units.
Design criteria for sieve trays in LPG treating service are reviewed by Tse and Santos
(1993). Average LPG velocities through the sieve tray holes should be from 0.5 to 1.0 ft/sec.
The holes are set on either square or triangular pitch and are usually !4 to K in. diameter. Tray
spacing varies from 0.5 to 2 ft with 1.5 to 2 ft being more typical. Tse and Santos (1993) cite
Laddha and Degaleesan (1978) for examples of detailed sieve tray rating calculations. Sieve
trays have been reported to have limited turndown in LPG treating service (Honerkamp,
1975; Bacon, 1972). If LPG flows vary significantly, it may be desirable to provide for LPG
recirculation as suggested by Changela and Root (1986).

Cocurrent Confactors

Honerkamp (1975) and Bacon (1972) reported on the design and operation of a single
stage contactor using an eductor. Figure 2-97 gives the flow diagram for this plant with a
material balance and operating conditions. In this design, the amine was used as the driving
fluid and experienced a 120-psi pressure drop across the eductor; the LPG experienced a
slight pressure increase. Operating conditions were 445 psig at 55°F. The singlestage educ-
tor-mixer treated LPG with 15 wt% MEA. The treated LFG attained a purity of 10 ppmw
CO? when the rich MEA loading was 0.22 moles C02/mole MEA. The settler downstream of
the eductor was designed for 30 minutes hydrocarbon liquid residence time, and the interfa-
cial area provided for amine-hydrocarbon liquid settling was about 0.58 ft2/gpm (equivalent
to 2,500 gal hydrocarbon/day/ft’). Honerkamp (1975) notes that the mass transfer efficiency
of single stage contacting devices is relatively insensitive to varying LPG flows, while both
sieve trays and random packing are affected by turndown.