Debottlenecking removes Auger production constraints

Debottlenecking of Shell's Auger facilities increased oil production capacity to 75,000 bo/d from the initial design of 40,000 b/d. Further expansion will allow Auger to produce 100,000 b/d.

Planned modifications will further raise platform capacity to 100,000 bo/d and 300 MMscfd.

SOI installed Auger, its first tension-leg platform (TLP), in 1994 in 2,860 ft of water. The platform was designed to process 40,000 b/d of oil, 100 MMscfd of gas, and 25,000 b/d of water. Since starting production, the surface fluid handling facilities have been the primary constraint limiting production.

The original platform design did not anticipate the higher well flow rates and surface temperatures. Problems encountered include:

• Poor gas/liquid separation efficiency in high-pressure separators

• Excessive natural gas liquid (NGL) condensation in the gas export pipeline

• Gas dehydration system limitations because of high production rates and surface temperatures

• Other facilities engineering-related problems.

Auger

SOI completed the Auger discovery well in May 1987 and initial production commenced in 1994. SOI has 100% interest in the leases that were acquired in 1984 and 1985. The central Gulf of Mexico field is in Garden Banks Blocks 426, 427, 470, and 471, about 214 miles southwest of New Orleans and 255 miles southeast of Houston.

Estimated field reserves are 220 million bbl of oil equivalent. Total project cost for Auger is about $1.2 billion including drilling.

The Auger project was Shell's first TLP. Since then, in May 1996 Shell has installed the Mars TLP. The water depth records of Auger and Mars will be eclipsed during the next several years as other SOI deepwater developments including Mensa, Ram/Powell, and Ursa come onstream.

Topsides facilities

The Auger topsides include:

• An integral drilling rig

• 32 well slots (24 oil/gas, 8 waterflood) around a 74 ft by 194 ft rectangular well bay

• 140 bed quarters building

• Processing facilities for oil, gas, and produced water including separation; dehydration; treatment; compression; and utility and safety systems, such as flare, firewater, etc.

The original Auger design called for producing rates of 40,000 bo/d, 100 MMscfd gas, and 25,000 bw/d, and planned injection of 60,000 bw/d for a waterflood.

The TLP is a floating structure and is subject to considerably more movement than fixed platforms. Critical process equipment with gas/liquid or liquid/liquid interfaces was, therefore, designed to accommodate the motion.

Additionally, because the TLP is held in place by tendons, buoyancy must be provided. This places a premium on reducing topside facility weight. Fortunately, alternative technologies not common on fixed platforms can combine reduced weight and sensitivity to motion.

Two examples include hydrocyclones rather than induced gas flotation for removing suspended oil from produced water and structured packing rather than valve or bubble cap trays in the glycol contactor for gas dehydration.

Fig. 1 [31030 bytes] is a very simplified schematic of the original Auger oil and gas processing facilities. Auger oil wells produce 37° API gravity oil and 0.68 gravity gas, while the gas wells produce 42° API gravity oil and 0.65 gravity gas.

The original concept included separate high-pressure separators for oil and gas wells designed to operate at 1,550 psig. Liquid from the two high-pressure vessels and a parallel test separator, not shown, flow to an intermediate-pressure (IP) separator, operating at 600-700 psig. Oil from the IP separator next flows to a free-water knockout (FWKO) at 125 psig. Currently, no water is produced. After the FWKO the oil flows to a bulk-oil treater (BOT) operating at 40 psig and finally to an oil tank from which it is pumped to a 75 mile, 12-in. pipeline to Eugene Island Block 331.

Oil from the FWKO passes through two heat exchangers to pick up heat from oil exiting the BOT and from the heat media system.

Gas from the high-pressure oil, gas, and test separators is commingled and flows to a filter separator and then to a triethylene glycol (TEG) contact tower where the gas is dehydrated. Although the sales gas/water specification is 7 lb/MMscf, the gas must be dried to below 4 lb/MMscf to avoid hydrates forming at the low Gulf of Mexico deepwater temperatures, about 40° F.

Gas is then delivered into a 37 mile, 12-in. Shell pipeline to a platform in Vermilion Block 397. At the platform, liquid and gas are separated, metered, and recombined and delivered into the gas transportation company's pipeline to shore.

Gas from the intermediate pressure and all lower pressure separators is compressed to pipeline pressure and combined with high-pressure gas upstream of the filter separator and dehydrator. A 2,000 hp dc motor-driven, four-stage reciprocating, flash-gas compressor (FGC) compresses gas from the stock tank, BOT, and FWKO. A centrifugal booster gas compressor (BGC) driven by a 5,000-hp gas turbine provides the final compression stages.

FGC discharge gas is routed to the first stage BGC suction with intermediate pressure separator vapors feeding the BGC interstage. Liquids condensed in compressor discharge coolers and separated in interstage scrubbers are recycled to the appropriate vessel in the oil separation train.

Production during this phase peaked at 55,000 bo/d and 105 MMscfd.

Stage 1 debottlenecking

Stage 1 debottlenecking activities took place during late 1994 to early 1995 (Fig. 2 [24972 bytes]).

Debottlenecking began in the summer of 1994 when it became apparent that the production potential of high-rate Auger wells far exceeded the surface facility handling capability.

Production from individual oil wells reached 12,000 bo/d and 25 MMscfd.

Several critical problems surfaced shortly after first production in April 1994. These problems mainly were with wells flowing at much higher flow rates than had been anticipated. Initially, 24 producing wells were planned with a design processing temperature of 90° F. Well temperatures were to be kept high by nitrogen in the production tubing/casing annulus to insulate the fluids from the cooling effect of seawater, thus minimizing paraffin and hydrate problems.

By July 1994, the production from just three wells was 27,000 bo/d and 75 MMscfd. The surface temperature of the wells ranged from 120 to 140° F. Early problems experienced included:

• Poor separation efficiency in high-pressure separators and filter separator (collapsed filter elements, no coalescing) resulting in carryover of oil into the glycol system

• High gas temperature and related water-carrying potential

• Dehydration system glycol carryover into the gas export line

• Difficulties in testing wells

• Excessive liquid hydrocarbon condensation in the sales gas pipeline

• Inability to accurately predict facility performance with process simulation model because of foaming in separators

• Inability of the gas pipeline company to handle condensed liquids and oil carryover from the high-pressure separators, which caused fouling and foaming of the glycol system.

To minimize surges, flow from all oil wells has been split between both high-pressure vessels. Gas wells have mainly been shut in because the oil wells are capable of keeping the facilities and pipelines full. Also, because of the elevated production temperatures, the heat exchangers upstream of the BOT were bypassed.

The debottlenecking study primarily addressed facility-related problems at Auger that limited production rates. Operations and engineering staff assigned to the Auger project, along with Shell head office and research engineers were involved in the study.

By August/September 1994, a short-term debottlenecking plan had been formulated. Gamma scan of high-pressure separators and fluid sample tests indicated that foaming was likely occurring in the separation vessels, and the vane-type demisters in these vessels were largely ineffective on foam. Tests on the high-pressure gas filter separator located upstream of the TEG contactor indicated that it was not completely removing entrained oil carried over from the high-pressure separators. This resulted in contamination of, and resultant foaming in, the glycol system.

The original design sales gas dew point of 2 lb/MMscfd was to be achieved by reducing the gas inlet temperature to 90° F. The high well flow rates were producing higher temperatures in the process facilities and thus there was the potential for the gas stream to carry more water into the TEG Contactor.

To reduce the temperature of the gas entering the TEG system, an air cooler was installed in the high-pressure gas stream. Also, the production tubing/casing annulus nitrogen was displaced with completion fluid, leading to a drop of about 10° F. in the well fluid temperature.

Several options were considered to improve separation efficiency in the high-pressure vessels, including a retrofit with vortex tube inlet devices, installation of internal spray nozzles to alleviate the foam problem, and other modifications to the vessel internals and inlet piping.

The option selected was to install a cyclone separator between the new air cooler and the filter separator. The cyclone would recover any liquids condensed from the gas by the air cooler plus liquids entrained from the high-pressure separators. This option was judged optimum because of the minimal downtime required and the high cost of deferred production.

Major progress was made during late 1994/early 1995. In addition to installation of the cyclone vessel and high-pressure gas coolers, the BGC and oil pipeline pumps were rewheeled to provide higher discharge pressures and flow rates.

The piping and heat exchanger of the final compression stage of the BGC were revised for higher discharge pressures. Spare pipeline and lease automated custody transfer (LACT) charge pumps were ordered because standby pumps in these services had been pressed into continuous service. A back-pressure valve was added on the high-pressure separators. FGC natural gas liquids (NGL's) were piped to the flare scrubber to minimize recycle load on this compressor.

Another early problem involved higher-than-predicted flash gas generated in liquid dump lines. This two-phase flow through the piping reduced the available pressure drop for the dump valves. Also, where chokes were installed downstream of level control valves to minimize pressure drop across the valve, two-phase flow was limiting flow across the choke. The chokes were removed from the dump lines from high pressure and intermediate-pressure separators to increase flow capacity.

Larger dump valves were installed on the FWKO and BOT. Also, a second dump line from the BOT to the dry oil tank allowed lowering of the treater pressure, thus reducing the amount of gas flashing off at the dry oil tank and reducing the back pressure on the level control valves. High flash gas rates had greatly exceeded the capacity of the first stage FGC.

Several operational changes also improved facility performance. The most difficult well was diverted from the high pressure and intermediate-pressure separator. The operating pressure of the high-pressure system was increased from about 1,500 to 1,800 psig.

Flow from the wells was split between the high-pressure oil separator and the high-pressure gas separator, rather than being segregated into one vessel or the other. Trials of antifoam chemicals in the high-pressure systems and friction-reducing chemicals in the oil export pipeline had some success.

Finally, fluid samples from several wells allowed laboratory analyses and pressure-volume-temperature (PVT) studies to be incorporated into the process simulation model. This then provided satisfactory agreement with actual Auger operations.

As a result of the Stage 1 debottlenecking, Auger's production rates increased to 62,000 bo/d and 125 MMscfd. Production still was limited by liquids in the gas pipeline, foaming in the high-pressure separators, and by TEG contactor capacity.

Attempts to reduce liquid drop out in the gas pipeline by changing facility and pipeline operating conditions, varying pigging frequency, and use of Joules-Thomson recycle from high pressure to intermediate-pressure systems have achieved only limited improvement.

Stage 2 debottlenecking

Stage 2 debottlenecking took place in late 1995 (Fig. 3 [30542 bytes]).

In September 1995, the gas pipeline company had a curtailment scheduled, which precipitated a complete production shutdown on Auger. During this time, SOI installed several new pieces of equipment.

Installed were a new vertical high-pressure gas separator-incorporating a vane-type inlet spreader, internal spray nozzles for foam abatement, and outlet gas vane-type mist extractor-and a new coalescing filter vessel. Other late 1995 debottlenecking included:

• 20-in. pipeline tie-in to the Auger oil pipeline to Eugene Island 331

• New manifold header for the high-pressure vertical separator and installation of a vent boom and vent scrubber for the new separator PSVs

• New glycol contactor and increased circulation capacity from 750 to 1,200 gal/hr with flow split between the two contactors

• Installation of a larger glycol/glycol heat exchanger.

• Installation of higher capacity packing in the glycol still and reboiler sparger columns

• Installation of a permanent vapor recovery unit (VRU) to supplement the existing FGC, but tie-ins were not completed.

By mid-1995, the reservoir engineers had determined that the waterflood system would not be needed. Removing it would provide the necessary weight savings for further expansion of the production facilities.

As a result of these equipment additions, flow rates increased to 73,000 bo/d and 154 MMscfd, but the high-pressure vertical separator did not perform up to its 170 MMscfd design rate. The problem appeared to be inefficient oil and gas separation, which manifested itself as increased gas rates at the intermediate-pressure separator and oil carryover into the outlet gas stream.

Stage 3 debottlenecking

Stage 3 debottlenecking took place during 1996 (Fig. 4 [30497 bytes]).

In April 1996, the gas transportation company executed a major shutdown to increase its liquid-handling capacity. During this shutdown, vortex tube inlet devices were installed in the high-pressure vertical separator, and several gas pipeline control valves were retrimmed. Also during the shutdown, miscellaneous tie-ins were made in preparation for further expansion and to finalize the VRU installation.

After these changes, Auger rates climbed to over 73,000 bo/d and 163 MMscfd. Effectiveness of the VRU, however, was hampered by recycling of compressor-generated NGLs.

After installation of the vortex tubes, the need for injection of defoamer chemicals decreased significantly and the separator efficiency improved, specifically in reducing the amount of gas carry-under in the oil feeding the intermediate system.

Conversion of the existing filter separator to a coalescer vessel and revision of tower internals in the original TEG contactor are planned for later in 1996. An upgrade to the produced water-handling system will also be required because water production is anticipated by the end of 1996.

A continuing problem that has not yet been resolved is the excessive condensation of hydrocarbon liquids in compressor interstage coolers. Condensed liquids are dumped to lower-pressure vessels where they reflash and the vapors are then recompressed. Considerable compressor capacity is required to handle the recycle vapors.

Peak rates are expected to reach 75,000 bo/d and 165 MMscfd.

Future expansion

Fig. 5 [33303 bytes] shows the future expansion plans. Work has begun to further increase the capacity of Auger to 100,000 bo/d and 300 MMscfd of gas.

New facilities on Auger will include a 14-in. riser, a new dry oil train, a fourth pipeline pump and LACT train, a new parallel BGC, some flare train work, and additional glycol-system capacity. These expansions are possible because of the weight savings gained by demolition of the waterflood system.

Some type of condensate weathering or stabilization may be provided to reduce compressor NGL recycle. The existing 12-in. oil pipeline will be converted to gas service and a new 16-in. oil pipeline will be installed from Auger to pipeline infrastructure in shallower water.

Lessons learned

When undertaking a new facility design, the best fluid characterization is usually used, but relying on that to be exact can give too tight of a design for the facilities. Some flexibility should be considered if actual production data do not exactly match the design conditions.

Also, for specific component designs, some sort of sensitivity analysis may be beneficial. A case in point is the Auger gas pipeline. The gas rates have been somewhat higher, but the real surprise has been the liquids generated in the pipeline.

Small differences in the concentration of mid-level hydrocarbons, (C4, C5, C6, etc.), in the gas stream can give huge differences in the volumes of liquids generated in the gas pipeline. Some sensitivity analysis might have dictated a more flexible design for the liquids-handling system going toward shore.

Another thing learned is that design engineers should plan for facility expansion from the start. This would include flanged breakout spool piping to minimize on site welding and downtime. Also, if in the original design, when staff time and resources were more available, higher rates were anticipated, accounted for, and documented, then the job of the debottlenecking engineer would be made much easier.

Various contingencies could be documented and preliminary expansion designs completed. This would mean that debottlenecking efforts could be executed directly.

Many facility project engineers have developed their skills by working major projects and facility designs. In these designs for new production facilities, several key parameters are understood and fixed.

Generally, a design basis is developed, leading to a cost estimate, and finally a schedule is developed. All three components are then largely fixed and the engineers can execute the project with these basic boundary conditions unchanged. An engineer's job then is to hold the cost, meet the schedule, and build the agreed design.

In the case of debottlenecking an operating facility, all normal constraints are turned upside down. The schedule is usually as soon as possible, and may change based on plant turnarounds and shutdowns dictated by outside forces. Schedules, therefore, are always changing, trying to take advantage of every opportunity for acceleration.

The capital cost, while not unlimited, is usually minor compared to the profit potential provided by either minimizing shut-in time or maximizing production rates for a major production facility.

Finally, the design basis is not fixed because new data and production information continue to come to light. The debottlenecking design must be flexible to account for changes in production and new information. It also must be flexible to take advantage of other opportunities such as unexpected shutdowns.

In short, everything you learned building the project the first time is out the window for the debottlenecking effort. All your normal constraints are backwards.

Finally, when working with an operating facility, it is vitally important to communicate with the field operations staff. The differences in location and rotating schedules of the field people make this a huge time burden, but a necessary one.

Field personnel will have to operate any change or addition that the engineer wants to make. If they are not on board with the designs from the beginning, odds for success are limited. Also, most engineers do not want to admit this, but the operations personnel can have some very good ideas for process improvements.

These field-generated ideas should be seriously considered along with those generated from the ivory towers.

Based on a paper presented at the 1996 SPE Annual Technical Conference and Exhibition, Denver, Oct. 6-9.

The Authors

Tom Judd is a staff facility engineer for Shell Offshore Inc. in New Orleans. He is part of a group of engineers supporting the Auger project and participated in the original design and construction of Auger.
Prior to joining Shell 6 years ago, he was a design consultant for 10 years. Judd has a BS in civil engineering from the University of Illinois at Urbana. He holds a Professional Engineering license in civil engineering in Louisiana.

C.B. Wallace is a chemical engineering advisor in the plant and facilities support department of Shell E&P Technology Co. in Houston. He primarily is responsible for process design of natural gas process plants, and oil and gas production facilities.
Wallace joined Shell 32 years ago. He has a BS in chemical engineering from Texas A&M and an MS in chemical engineering from the University of Texas. He is a member of AIChE and is a registered Professional Engineer in Texas.

Copyright 1996 Oil & Gas Journal. All Rights Reserved.