NGL recovery influences Louisiana processing

• The Paradis fractionation plant, west of New Orleans, has a standard four-tower train (Fig. 2)[74,789 bytes].

• Phase I: A 350-MMscfd dew point control unit at the Larose gas processing plant that was placed in operation April 1998.
• Phase II: A 600-MMscfd cryogenic NGL-recovery facility at Larose that started up in November 1998. Much of the equipment from Phase I was integrated into the two-train cryogenic plant in Phase II.
• The demethanized product from Larose is fractionated at an expanded fractionation facility in Paradis, La. This fractionator produces ethane, propane, iso and normal butanes, and natural-gasoline products and is designed to handle a range of feed compositions.

Discovery Producer Services LLC is a Texaco/Williams/British Borneo joint venture. Engineered, procured, and constructed by Black & Veatch Pritchard Inc. (BVPI), Kansas City, and operated by Bridgeline Gas Distribution LLC, the Discovery plants will produce and fractionate more than 42,000 b/d of NGL product.

Figs. 1 and 2 show the Larose gas-processing plant and the Paradis fractionation plant.

Gas gathering

The 105-mile, 30-in. main line pipeline system was designed by Project Consulting Services Inc., New Orleans, and connects multiple production facilities (OGJ, Oct. 19, 1998, p. 73). Pipeline capacity, certified by the U.S. Federal Energy Regulatory Commission (FERC), is approximately 600 MMscfd of raw natural gas. In addition, the pipeline system can accept up to 7,500 b/d of produced condensate.

The pipeline system includes a large slug catcher designed by M&H Enterprises, Houston, at the end of the pipeline to facilitate pigging operations. The 7,500-bbl slug catcher consists of eight 48-in. pipes, each approximately 460 ft long. Fig. 3 [93,274 bytes] shows the approximate location of the Larose and Paradis plant sites.

Condensate recovered in the slug catcher is stabilized by a series of two pressure reductions. Vapor generated by flashing the condensate is recompressed and combined with the vapor exiting the slug catcher. This combined stream is then processed in the NGL-recovery portion of the facility. Stabilized condensate is pumped into an 18-in. Equilon pipeline.

Larose phases

Phase I would provide a means for early production and allow additional third-party gas supply to be sought before completion of the cryogenic plant. When the Phase II plant was completed, Phase I would be shutdown and converted to Phase II operation.

With a Phase I operating life of less than 12 months, as much equipment as possible would be reused in Phase II to minimize cost.

The Phase I facility was designed to process 350 MMscfd of inlet gas with a glycol-injection/refrigeration scheme. Approximately 16 gpm of ethylene glycol is injected into the inlet gas, and the resulting two-phase stream chilled using propane refrigerant (Fig. 4) [71,010 bytes].

A three-phase separator separates the ethylene glycol and condensed hydrocarbon streams.

The gas stream exiting the separator exchanges heat with the inlet gas before entering the residue-gas pipeline. Recovered hydrocarbon liquid is stabilized to meet the NGL vapor-pressure specification and routed to the NGL pipeline. Ethylene glycol is regenerated with conventional regeneration before being reinjected.

Three 1,200-hp gas engine-driven reciprocating compressors provide the required refrigeration.

The Phase II facility is designed to recover approximately 90% of the ethane and essentially 100% of the propane and heavier components from 600 MMscfd of 3.2 gal/Mcf inlet gas.

Approximately 42,000-b/d of demethanized product mix moves via pipeline to Paradis fractionation. As economic conditions warrant, the plant can also be easily converted to ethane rejection.

Inlet gas from the slug catcher is metered and routed to an inlet separator to remove any remaining free liquid. The gas is then dehydrated in a four-bed molecular sieve system.

A slipstream of the inlet gas is used to regenerate the sieve, then returned to the process upstream of the dehydrators.

The cryogenic plant process (Fig. 5) [71,074 bytes] was selected based on the desired recovery level, efficiency, capital cost, and simple flow scheme. In order to increase the plant operating range and minimize the impact of downtime on producers, two 300-MMscfd cryogenic trains were installed.

Where possible, equipment from Phase I was incorporated into the Phase II plant. For example, the propane refrigeration, heating medium, condensate handling, and fuel-gas systems were integrated into the Phase II facility.

A significant portion of the capital and operating cost of the facility involved the selection of the residue compression: three turbine-driven Elliott centrifugal compressors, with each Solar Mars turbine providing 15,000 hp (ISO rated).

While the Solar Mars units were selected based primarily on capital cost and fuel efficiency, three 33% units also provided a wider operating range. The design residue-gas delivery pressure was 1,000 psig, with available residue-gas markets consisting of both Bridgeline Gas Distribution and Texas Eastern Transmission Co. pipelines.

Two 800-hp cooling water pumps provide 33,000 gpm of bayou water for the facility. In the once-through system, water is removed from an adjacent bayou and filtered before being routed to the exchangers. Because the temperature rise across each user is limited to 15° F., the water can then be discharged directly back into the bayou.

During ethane rejection, hot residue gas leaving recompression provides the required reboiler duty. This heat integration reduces the cooling-water requirement in the residue-gas coolers and eliminates the need for hot oil in the cryogenic plant.

Use of hot residue gas to provide reboiler duty also allowed the duty of the hot oil system to be reduced by 66%.

Paradis expansion

The Discovery Project scope consisted of expanding the existing four-tower fractionation train to an ultimate capacity of 42,000 b/d. The Paradis fractionator uses a conventional process configuration (Fig. 6) [71,974 bytes].

Originally designed for 25,000 b/d, the fractionation train had been shutdown for several years as production had declined. Because the train was idle, existing equipment could be revamped and reused where possible.

As a result of the higher plant throughput, increased reboiler and condenser duties required complete replacement of the heating medium and refrigeration systems and major modification to the existing cooling tower.

Thorough mechanical inspection further revealed that only major equipment items should be salvaged. Tower loadings for the new conditions required replacement of all trays and internals. Three new 1,665-hp gas engine-driven reciprocating compressors were included to meet the refrigeration demand.

The purity ethane product is treated with 50% diglycolamine (DGA) to remove CO2 and dehydrated with triethylene glycol (TEG). The ethane product is then condensed prior to delivery into the pipeline system. Although only trace sulfur compounds are expected, provisions were also made for installation of future propane treating.

Supplemental product storage was added to accommodate planned shipping schedules. In total, more than 57,000 bbl of aboveground product storage is available. Fractionated products are shipped via Texaco's Expanded NGL Distribution System, truck, or rail.

The existing fractionator control was upgraded to a Foxboro distributed control system (DCS). The Foxboro system also provides advanced process control (APC) to ensure quick response to process changes.

The APC system minimizes reflux rates while meeting product specifications. In addition, tower pressures are operated as low as possible to minimize reboiler and condenser duties.

Challenges

Equipment design

One example was the selection of the refrigerant compressor cylinders. While both phases utilized the available horsepower, the compressor suction pressure during Phase I was approximately 5 psig vs. 30 psig during Phase II operation. If the selection were based solely on the Phase I condition, the resulting cylinders would be oversized for Phase II and exceed the available horsepower.

Conversely, if the selection were based solely on Phase II conditions, the resulting cylinders would be undersized for Phase I and result in an unacceptable residue-gas dew point.

Given the limited operating life of the Phase I plant, the capital expense for additional compressor cylinders could not be justified. Also, long-term operation with valves partially unloaded was viewed as a maintenance concern.

Rather, an intermediate cylinder size was proposed that would meet the Phase II requirements and yet have minimal impact on the Phase I residue-gas dew point.

The selected 22.5-in., first-stage cylinders provide approximately 90% of the specified Phase I refrigerant flow. This resulted in a residue-gas dew point of 0° F. (vs. a target of -10° F.), which remained sufficient to prevent condensation in the downstream pipeline system.

During Phase II operation, variable volume pockets are adjusted to provide the required refrigerant flow without exceeding the available horsepower.

Similar analyses were required for other equipment items to ensure proper operation in all operating modes.

Equipment selection included careful consideration of capital costs, required operating ranges, operating life, and operating and maintenance costs.

Surplus equipment

Several existing equipment items from the Texaco Henry, La., gas plant were made available for use. Although designed for similar applications, some modifications were often required to meet the specified Phase I and Phase II operating conditions.

Of particular importance were the gas/gas exchangers. Two large shell-and-tube exchangers, each with more than 115,000 sq ft of surface area in four shells, were to be used in this service.

Similar to Phase I, the exchangers were designed for ethylene glycol injection into the inlet gas. After separation of liquids, the residue gas was cross-exchanged at high pressure with the inlet-gas stream.

Although designed for larger flow rates, the design temperatures and pressures for the exchangers were very near the Phase I requirement, thus making the exchangers an ideal fit.

During Phase II operation, however, the residue-gas stream operates at approximately 300 psig rather than 1,100 psig as in the original design. At this lower pressure, increased residue-gas velocity was a concern. Detailed exchanger calculations confirmed that high velocity near the baffles could cause tube vibration.

To reduce shell-side velocity, the inlet-gas stream was changed from the tube side to the shell side. This eliminated vibration concerns and increased the overall heat-transfer coefficient by 20%.

Because the inlet gas is condensing, however, the shell-side inlet was modified so that the inlet gas would enter the top of the exchanger and condense as it flowed down through the shell side. In effect, the original outlet nozzles would become the new inlet nozzles. To prevent erosion damage of the tubes near the inlet, impingement plates were added to each shell.

The design temperature of the gas/gas exchangers also represented a limitation. With a design temperature of -50° F., the exchangers were unsuitable for the residue-gas temperature of -65° F. Therefore, a small brazed aluminum exchanger was included upstream to preheat the residue gas and avoid metallurgy concerns.

Although additional design effort was required to adapt the equipment to new conditions, BVPI was able to use the majority of the surplus equipment available. Along with additional engineering requirements, minor refurbishment and modifications were often needed.

At the Larose facility alone, 14 surplus equipment items were used and resulted in an estimated $1.8 million cost advantage to the project.

Operating permit

Predictions indicated that the operating permit would likely be issued in early 1997, 2-4 months after the scheduled site opening. This ruling created an unforeseen delay in construction and jeopardized the scheduled start-up.

To minimize impact of the delay, equipment procurement was allowed to proceed as planned, which allowed vendor drawing submittals and fabrication to continue, and BVPI's detailed engineering effort could proceed.

Although this represented an element of risk, any delay in the procurement of the long-lead brazed aluminum heat exchangers, turbine-driven compressors, and turboexpander packages would have likely resulted in an extended schedule for Phase II.

When the FERC operating permit was granted in March 1997, BVPI detailed engineering was proceeding as originally scheduled. Because of the 4-month delay in site work, however, equipment deliveries were scheduled to begin before preparation of the construction staging area.

Through coordination with vendors and construction personnel, BVPI arranged off-site equipment storage until needed in the field. Construction personnel used this to their advantage by, in many cases, unloading equipment items and placing them directly on their foundations.

Archaeological discovery

Further investigation confirmed that an archaeological site-apparently a 300-year-old fishing village-was near the proposed administration/control building. Excavation by a team of experts was scheduled to begin once the FERC permit was granted.

During the remediation period, no construction activity could occur within the mitigation area until cleared by the archaeologist. Preliminary estimates suggested that 2-3 months would be required.

Because the administration/control building was designed to house the electrical and DCS equipment, a lengthy construction delay would also result in a start-up delay. A joint BVPI/Discovery team was formed to develop options to minimize the schedule impact. Based on the team's recommendations, a separate power building was proposed to contain the motor control center and electrical switchgear, adjacent to the administration/control building. This allowed construction to begin independent of the archaeological work.

A temporary trailer was used to set up the DCS equipment prior to completion of the control room.

With remote input/output, only a few fiber optic cables would need to be reinstalled after the DCS equipment had been moved into the control room. These cables were cut long by construction personnel and then rerouted to the permanent control room facility.

In the end, approximately 5 months were required to complete the archaeological mitigation.

Phase conversion

As construction approached completion, however, inlet-gas volumes were insufficient to require both trains. To reduce the disruption for producers, a new conversion plan was developed.

While the majority of Phase I equipment was to be reused in the Phase II 200 train, the 300 train consisted of new equipment that could be placed in operation without requiring a shutdown of Phase I.

In addition to reducing downtime for producers, this plan involved less risk because the Phase I plant was available if unforeseen problems delayed the 300-train start-up.

Because Phase II piping interconnections were installed with spectacle blinds to facilitate conversion, few additional piping modifications were required with the new start-up sequence. In the original plan, however, the DCS was to be shutdown and restarted with the Phase II program.

To start the 300 train, additional DCS programming was required. Foxboro was contacted and the necessary loops were added to the Phase I DCS program.

With thorough planning, the necessary modifications were completed in 12 hr. During this period, gas was allowed to pack the inlet-pipeline system so that production was not interrupted.

Paradis debottlenecking

With engineering at approximately 45% complete, Discovery asked BVPI to study the modifications necessary to achieve a suggested 50,000-b/d capacity.²

A joint BVPI/Discovery team evaluated both the existing and new equipment and concluded that it would not be economically feasible to increase the capacity to 50,000 b/d. Essentially all of the major equipment, including the existing towers, would require significant modifications or replacement.

The next step was to determine if opportunities existed to increase capacity without a large increase in either capital cost or the project schedule.

Further investigation revealed that a capacity of 42,000 b/d could be achieved, with the refrigeration compressors representing the ultimate bottleneck. Because the compressors had been purchased and permitted, increasing the compression requirements was ruled out due to significant schedule impact.

Once the new capacity was determined, major effort still remained to incorporate the change. Consisting of project members from multiple disciplines, the debottlenecking team identified critical activities and developed a work plan.

An aggressive 6-week schedule was proposed to complete the exercise.

Working closely with Discovery and vendor engineers, BVPI determined whether to modify or replace equipment items. Required modifications to the tower internals were done so that existing support rings could be reused.

Similarly, heat-exchanger arrangement and dimensional changes were minimized to reduce civil and piping rework.

As a result of a 30% increase in the required hot-oil-heater duty, burner replacement was necessary to provide a higher heat release. At the conclusion of the detailed equipment evaluations, new internals were required for three towers, and several heat exchangers were retubed or increased in size.

Cooperation led to the fractionator capacity being increased by 25%, with only a 5% increase in project cost.

Challenging goals set by the team also allowed development of a schedule recovery plan. Engineering and construction opportunities were identified which virtually eliminated any schedule impact.

References

1. Spaulding, S.R., and Stevens, T.S., "Louisiana NGL-A New Era," presented to the 77th GPA Annual Convention, Mar. 16-18, 1998, Dallas.
2. Talib, J.H., Germinder, B., and Hitchcock, M.P., "Team work increases fractionation capacity cost-effectively," Hydrocarbon Processing, October 1997, p. 121.

The Authors

Kevin L. Currence is a senior process engineer for Black & Veatch Pritchard Inc. in Overland Park, Kan. He is a 1990 graduate of the University of Missouri-Columbia with BS degrees in chemical engineering and biochemistry. Prior to joining BVPI, Currence spent 6 years with Conoco Inc., where he supported various gas-processing projects both as a process and project engineer. Currence joined Black & Veatch Pritchard in 1996 as a process engineer in Trinidad. He is a registered professional engineer in Oklahoma.

Brian C. Price is technology manager for gas processing and cryogenics for Black & Veatch Pritchard Inc. in Overland Park, Kan. He is in charge of technology development and process design for gas processing, NGL recovery, and LNG production facilities for Pritchard. Price has over 25 years' experience in gas processing and related technology. Prior to joining BVPI, he worked for ARCO Oil & Gas Co. in various positions, including manager of process engineering and projects manager. Price is a member of AIChE and is active in the Gas Processors Association and API. He currently serves on the Editorial Review Board for the GPSA Engineering Data Book and is past Chairman of the Technical Committee for GPA. Price has BS and MS degrees in chemical engineering from Oklahoma State University.

Bill Coons is an engineer for Texaco, currently assigned to Texaco's worldwide upstream, commercial development. He specializes in gas plant design, PSM implementation, and operation of gas-processing plants and related facilities, including offshore gas and condensate production from the Danish North Sea. Coons joined Texaco in 1978 after receiving a chemical engineering degree from Oklahoma State University.

Copyright 1999 Oil & Gas Journal. All Rights Reserved.