STRUCTURED PACKING PROVEN SUPERIOR FOR TEG GAS DRYING

James A. Kean, Harry M. Turner, Brian C. Price
ARCO Oil & Gas Co.
Dallas

Structured packing for natural gas dehydration has proven superior to trayed columns in tests by ARCO Oil & Gas Co.

Structured packing provides roughly twice the capacity and up to 50% greater efficiency.

The combination of high gas capacity and reduced height of an equilibrium stage (HETP), compared with trayed contactors, makes the application of structured packing desirable for both new contactor designs and existing trayed-contactor capacity upgrades.

ESSENTIAL ELEMENT

In terms of capital and energy, gas dehydration is often a minor part of the total plant. But proper dehydration is nevertheless essential to the operation of the facility as a whole.

On offshore platforms, the dehydration unit is often the largest single processing unit and strongly affects weight and layout requirements.

Triethylene glycol (TEG) absorption has been the workhorse of the industry for applications ranging from small field units to large-scale, low dew point units. Traditionally, the glycol absorber contains 6-12 trays to absorb water.

Bubble-cap trays are used because of the extremely low liquid rates inherent in this process. Designs based on data developed over the past four decades have produced reliable and flexible facilities for moderate dew point requirements.

Recent facilities have pushed the dew point depression needs beyond traditional applications. Facilities with water dew points of -40 F. and colder require more contact stages and extra care in design for proper operation.

One option to the trayed TEG contactor is the use of structured packing.

Before 1986, this packing had not been used in TEG dehydration and had not been utilized in applications with extremely low liquid loadings.

The potential of a shorter contactor and a dramatic reduction in contactor diameter led ARCO Oil & Gas in 1985 to conduct pilot-scale testing of structured packing.

Subsequently, the company has successfully used this packing worldwide in several applications.

Structured packing was developed as an alternative to random packing to improve mass-transfer control by the use of a fixed orientation of the transfer surfaces. Fig. 1 shows an example packing.

The material is a series of closely spaced corrugated sheet-metal plates stacked vertically on edge with corrugations at 45 from horizontal. Flow passages direct the gas and liquid flows counter currently to each other.

Packing sections, usually less than 12 in. high, are stacked so that the corrugations are oriented 90 to the layers above and below.

Structured packing is manufactured and sold by several vendors in the U.S. and Europe. Packings are generally of the same geometry but each vendor has a different type of surface treatment intended to enhance vapor-liquid contact.

The packing comes in different surface areas per unit of volume; process conditions and contactor diameter determine selection. Structured packing has been used in a variety of fractionator applications but has seen use in TEG dehydration only recently.

Based on sizing from packing vendors, the packing provides roughly twice the throughput of bubble-cap trays for dehydration. Additionally, the expected HETP results in as much as 40% reduction in contactor height.

PILOT TESTING

With so little data on structured packing in TEG service, ARCO Oil & Gas set up a test facility for this application in 1985. The pilot plant was built at ARCO's Block 31 facility near Crane, Tex., to measure efficiency and capacity of various structured packings.

The test program covered two different dehydration applications.

The first was a low dew point application using 17 ft of packing and gas-stripped glycol (99.95 wt %) to achieve a gas-outlet water content of 0.1 lb/MMscf.

The second series of tests used 75, 94.5, and 108 in. of packing and 99-99.4 wt % TEG to simulate operation of a typical offshore TEG dehydration system requiring 7 lb/MMscf pipeline specification.

The purpose of the offshore test program was to verify that structured packing would perform well at reduced bed heights, which would permit design of offshore TEG contactors to fit between decks while meeting pipeline water-content specifications.

A TEG dehydration system at Block 31 was modified to include the following elements:

A 14.312 in. ID contactor designed for easy removal and replacement of structured packing (Fig. 2)
An absolute coalescing filter separator to measure glycol carryover
A 4 in. ID x 14.5 ft tall stripping column packed with 5/8 in. Pall rings enabling 99.98 wt % TEG with only 3.5 standard cu ft (scf) stripping gas/1 gal TEG circulated. Gas temperature was controlled by a water cooler.

Pilot plant site elevation was 3,000 ft at 13.2 psia. TEG (1.5 gpm) was stripped and pumped to the contactor where the flow was split depending on the desired TEG-contact test rate.

The TEG contactor was designed to hold up to 17 ft of structured packing and was equipped with inlet vapor distributor and a drip-tube liquid distributor with 12 drip points. A trough liquid distributor was also tested for comparative performance.

A process diagram of this facility is shown in Fig. 3.

TEST PROCEDURES

The test proceeded in the following order:

A gas rate was selected and performance data were taken at varying TEG rates before a new gas rate was tested.
In some cases not all process conditions were tested when it was obvious the packing was unable to meet performance requirements. Also, data were not taken where it appeared equilibrium had been achieved or turndown was not in question.
Water injection along with visual inspection of water dumped from the inlet separator ensured inlet-gas saturation. Fig. 4 shows the water content data used for dew point evaluations.
Gas flow to the contactor is indicated by the variable Fs which is the product of the gas superficial velocity (Vs) multiplied by the square root of the gas density (pg), as in Fs = Vs /'pg.
For our study, values for Fs were varied from 0.5 to 3.0.
Flood point was determined at a TEG rate of 0.7 gpm/sq ft. Flood-point data were determined by raising the gas flow at 0.7 gpm/sq ft TEG until the contactor's pressure drop started increasing without limit.
Each type packing had a characteristic flood point with the exception of Koch/Sulzer plastic atop Flexipac II which had excessive carryover at Fs = 3.9, at which point testing was stopped.
Gas flow during flood testing was metered with a 23/4-in. orifice in a 4.026-in. meter run with a 100-in. range mete r.
Aging tests were conducted to determine if efficiency changed with time. All packing had been solvent washed before shipping to eliminate effects on wettability of manufacturers' preservative compounds.
In most cases, three dry-gas samples and three corresponding lean-TEG samples were taken for each process condition. One rich-TEG sample was taken at the end of each process condition.
Carryover was gauged from two 6-in. schedule-160 boots at the bottom of the absolute separator.
Low dew point test conditions were 650 psig and 8090 F. TEG circulation rates were 0.3, 0.5, and 0.7 gpm/sq ft.

Offshore test conditions were 1,050-1,100 psig (Nutter), 600 psig (Koch), and 90110 F. TEG circulation rates were 0.33, 0.67, and 1.0 gpm/sq ft.

Test data measurements were made with the following major equipment:

Outlet water content was measured with a LockwoodMcLorie analyzer with a 74-ml sample loop volume. The 74-mi sample loop allowed good resolution down to 0.03 lb H2O/MMSCf without risk of sample column break-through.
Lean and rich TEG was measured with a Mitsubishi CA-10 portable moisture meter (Carl Fisher method).
TEG flow rate was metered with a Max Machinery flow meter, 0-2 gpm range, with factory-tested accuracy better than 0.001 gpm.
Gas flow was measured with a Daniel meter tube (4-026 in. ID) and senior fitting along with a Barton meter having a 100-in. range.
Contactor pressure drop was measured with a barton meter with a 50-in. range for the first five packings tested and a 100-in. range for the final six.

EFFICIENCY

Eleven structured packings were tested for efficiency at low dew point. Typical results for three of the packings are summarized in Table 1 and Figs. 5 and 6.

Data analysis was aided by a computer program developed by ARCO for McCabe-Thiele theoretical tray calculations.

Efficiency varied considerably among the packings tested, ranging for example from 3.7 to 5.5 ft per theoretical stage at Fs = 3 and 0.3 gpm/sq ft.

Another measure of efficiency can be seen in Fig. 6 where one can compare the circulation ratio required for a given dew point depression. Notice that two of the more efficient packings appear to be approaching equilibrium at 7 gal TEG/1 lb H2O.

Five of the 11 were capable of meeting the target outlet dew point specification of 0.1 lb/MMscf.

Offshore test results are summarized in Figs. 7 and 8.

The HETP exhibited at the reduced bed heights tested was very similar to the low dew point 17-ft bed height test results, verifying that structured packing will perform well at bed heights as low as 6 ft.

These tests also showed that HETP is pressure dependent, with HETP increasing with pressure. This is attributed to lower gas diffusivity as a result of higher gas density and also increased liquid surface tension at higher pressure.

HETP generally increased with higher liquid rates. As the circulation rate was increased, however, fewer theoretical trays were required because the increased concentration gradient of water between the gas and liquid phases in the contactor offsets the need for vapor-liquid contact. In addition, HETP was generally lower with increased gas velocity.

High gas rates spread more liquid across the structured packing surface and cause more liquid hold-up and residence time. Both spreading and hold-up facilitate better mass transfer.

CAPACITY, ENTRAINMENT

Nine of the 11 packings were tested for capacity at 0.7 gpm/sq ft. Flow was increased in small increments greater than Fs = 3, allowing the contactor differential pressure to stabilize before flow was increased to the next higher rate.

When the differential would no longer stabilize, the flood point was reached. Floodpoint results are summarized in Table 2.

In general, packings with higher capacities had lower efficiencies. All but one of the packings demonstrated operating capacities of Fs = 3 or greater with a comfortable margin from flood.

The five most efficient packings were also tested for carryover at Fs = 3 and 0.7 gpm/sq ft. Entrained TEG was collected in an absolute coalescing filter separator downstream of the contactor. The top four in carryover ranged from 0.16 to 0.35 gal/MMscf.

Because the test contactor had no internals to reduce entrainment, most of the carryover could be eliminated with an internal vane unit.

Carryover was also found to be a function of distributor height above the packing. With drip tubes 6 in. above the packing, there was a fivefold increase in carryover compared with drip tubes resting on the packing.

The farther drops must fall, the more they are exposed to shearing and the more likely they are to reach terminal velocity which causes greater splashing on impact into smaller more entrainable droplets.

As suggested by the dew point depression data in Table 1, lower outlet-water content was achieved at 3:1 turndown in gas rate from Fs = 3.

Conversely, at constant gas rates, a reduction in glycol circulation led to increased outlet-water content, as expected.

At a circulation ratio of only 1 gal TEG/1 lb H2O removed, however, the five most efficient packings achieved an outlet-water dew point of 0.1 lb/MMscf or less.

Offshore testing which demonstrated a 5:1 turndown of gas rate from Fs = 2.5 also led to reduced outlet-water content at the lower gas rate. Recent testing at ARCO's Wilburton, Okla., facility where structured packing is installed, demonstrated that even higher turndown in gas rates is possible.

When gas flow was reduced by a ratio of 12.5:1 from Fs = 3, outlet-water content rose from 1.4 lb/MMscf to only 1.8 lb/MMscf.

DESIGN COMPARISONS

Bubble-cap contactors have been designed based on a C factor:

Gm = C /pg(pl - pg)

where:

Gm = Mass velocity, lb/(hr-ft2)

C = Empirical factor

pg = Gas density, lb/cu ft

pl = Liquid density, lb/cu ft

A design C factor of 650 is typically used, and contactors have been pushed to an equivalent C factor of 800 in many applications.

Structured packing contactors are usually sized based on the variable Fs, as defined earlier.

Pilot testing has shown that an Fs of 3 is a reasonable value for use for sizing which represents loading of about 75-80% of flood. This equates to an equivalent C factor of about 1,300 for this application. Bubble-cap contactors are generally sized with a tray efficiency of 25% (although some operators have used 33%). For most moderate dew point applications, this results in using 6-8 trays in the contactor. For low dew point applications, this results in 12-16 actual trays required.

Test results show that for structured packing, the HETPs vary from about 3 ft for moderate dew point applications to 6 ft for low dew point applications. Outlet dew points of 7 lb H2O/1 MMSCF can be achieved with only 8 ft of packing. Low dew points require twice that amount.

Fig. 9 shows a comparison of trayed vs. structured packing contactors with the same design conditions.

Table 3 and Fig. 9 provide a detailed comparison of trayed vs. structured packing contactors in the same service. Use of structured packing in this application results in a 30% cost reduction, 49% weight reduction, and a 28% height reduction.

FIELD APPLICATIONS

Table 4 contains performance data for several dehydration trains at ARCO-operated facilities where structured packing is installed. During 1991, five more installations are being started up with capacities ranging from 60 to 110 MMscfd. Most of these are offshore.

ARCO's study suggests the following conclusions and recommendations:

Tests with structured packing show it to be superior to trayed columns providing roughly twice the capacity and as much as 50% greater efficiency.
The combination of high gas capacity and reduced HETP in comparison with trayed contactors makes the application of structured packing desirable for both new contactor designs and existing trayed-contactor capacity upgrades.
Packings from each of the three packing vendors performed well and have useful application in dehydration service.
Test results indicate a design point of Fs = 3 for the more efficient packings. For moderate dew point requirements, such as 7 lb/MMscf, a higher capacity packing could be used with design rates up to Fs = 3.5.
A vane-type mist eliminator with 3-4 in. mesh pad face should be used with structured packing in TEG contactors. This will reduce carryover under normal operating conditions.
The design gas velocity in a tower which uses structured packing is approximately twice the normal design velocity in a trayed contactor with a wire mesh mist eliminator. Glycol carryover is likely if a wire mesh mist eliminator is used in conjunction with structured packing.
A high-efficiency drip point distributor should be used with structured packing in glycol dehydration service. The flow pattern exhibited by the drip-point distributor is far superior to the pattern exhibited by either a notched trough distributor or a spray-nozzle distributor.
The number of drip points should be equal in all quadrants of the distributor with the peripheral drip points positioned no more than 2 in. from the vessel wall.
A typical drip-point layout is a 4 in. x 4 in. matrix with the points nearest the wall oriented to accomplish the best possible distribution in that region. It is also imperative to install the distributor on a horizontal plane as close to the structured packing as possible in order to ensure proper liquid distribution and reduce carryover.
The distributor should also be designed with vapor space capable of functioning at Fs = 3. This requirement should be specified to the distributor's manufacturer to avoid poor liquid distribution at high gas velocity.
At low gas rates the packing efficiency decreases because of a gas bypassing effect, with even 1-5% gas bypass significantly reducing structured-packing efficiency.
The bypassing is caused by low liquid holdup, leading to insufficient contacting between the gas and liquid because there is less liquid surface available for gas contact. In dehydration service, however, the number of theoretical contact stages required tends to decrease with increasing glycol flow rate. This offsets efficiency loss and provides liberal turndown.
Turndown has been remarkable with this type packing. Field tests down to 12:1 have shown that dehydration specifications are still met. No lower capacity limit has yet been determined.
HETPs are a function of dew point depression (approach to equilibrium) as well as tower loading and operating pressure.
In general, packing with higher surface area will achieve higher efficiency but have lower capacity.

BIBLIOGRAPHY

Bonilla, J. A., Shieh, J., and Wang, P., "High Performance Packing in Gas Treating," presented at AlChE Meeting, Denver, August 1988.

Bravo, J. L., Roche, J. A., and Fair, J. R., "Mass Transfer in Gauze Packing," Hydrocarbon Processing, January 1985, pp. 91-95.

Kunesh, J. G., Lahm, L. L., and Yenagi, T., "Liquid Distribution Studies in Packed Beds," presented at AlChE Meeting, Chicago, November 1985.

Perry, David, Nutter, Dale E., and Hale, Andy, "Liquid Distribution for Optimum Packing Performance," Chemical Engineering Progress, January 1990.

Spiegel, L., "Applications of Structured Packings in the Process Industries," presented at AlChE Meeting, Chicago, November 1990.

Stoter, F., Olujic, Z., and de Graauw, J., "Measurement and Modeling of Liquid Distribution in Structured Packings," presented at AlChE Meeting, Chicago, November 1990.