Peter M. Ricca
Alyeska Pipeline Service Co.
Anchorage
Trans-Alaska Pipeline (TAPS) has initiated a large-scale corrosion mitigation program using chemical inhibitors.
Ultrasonic inspection in 1988 revealed internal corrosion predominantly in low-velocity piping where emulsified water had precipitated out of the crude oil and supported microbial corrosion.
TAPS developed a system of manual batch treatment to treat the numerous dead legs in the pump-station facilities along with procedures for monitoring inhibitor effectiveness.
This inhibitor treatment system was implemented in 1990 at TAPS' stations.
This first article about the program focuses on the corrosion sites and project design. The conclusion deals with treatment and monitoring procedures and presents some preliminary results.
THE SYSTEM
The 800-mile Trans-Alaska Pipeline transports 2 million b/d of crude oil from the North Slope to tidewater at Valdez, Alas. The pipeline has 12 pump stations, 10 of which contain turbine-driven centrifugal pumps, and 2 bypass stations containing no main line pumps.
The 27 API gravity crude oil transported by TAPS contains 1-2% sulfur and less than 0.35% water. Pipeline oil temperatures range from highs of 140 F. (60 C.) at Prudhoe Bay to 95 F. (35 C.) at the Valdez terminal.
After waterflood pressure maintenance began, the produced water contained increased concentrations of sulfates as the result of seawater breakthrough. In 1988, ultrasonic inspections on TAPS revealed internal corrosion in pump-station piping.
In September 1988, engineering work was initiated to quantify the source and extent of the problem and plan a mitigation program of inhibitor treatment. Evaluating the extent of corrosion took place in 1989, and mitigation treatment began in February 1990.
Inhibitor-treatment systems were first installed at the four northern pump stations which showed the most severe internal corrosion.
These northern installations served as prototypes for installations at the eight other pump stations. After inhibitor treatment began, extensive corrosion monitoring was undertaken to determine the effectiveness.
A typical TAPS pump station consists of a main pump building, reinjection pump building, manifold building, control building, shop-warehouse building, a crude-oil relief tank, a fuel-storage tank, and crew living quarters. Pump Station 4 (Fig. 1) is typical.
The five major functional pipe groups are: pump suction, discharge, and bypass piping; pig traps and pig passing piping; pressure surge-relief piping; oil reinjection piping; and small-diameter piping and building drains. Each station also has a large surge-relief tank and several smaller building sump tanks.
The first four pipe groups contain pipe sizes ranging between 12 and 48 in. (30.48-121.92 cm) in diameter. This large-bore pipe is designed to ANSI B31.4 (Class 600) specifications and made from low-temperature steel.
Small drain lines range in size from 2 to 6 in. (5.08-15.24 cm). These lines are designed to ANSI (Class 150) specifications.
DEAD-LEG CORROSION
Engineering investigations revealed the extent and location of corrosion present.
Ultrasonic inspection (UI) of more than 760 pipe locations revealed internal corrosion in most no-flow, low-flow, and intermittent-flow crude piping as well as in gravity drains and sumps.
All five functional groups of piping were affected to greater or lesser degrees. The only pipe segments found free of internal corrosion were manifold and pump piping in continuous service where oil velocities exceeded 3.5-4.0 fps (1.07-1.22 m/sec).
UI maps were constructed on a 1-in. (2.54-cm) grid and identified a predominance of pitting corrosion. Where corrosion occurred, pitting corrosion rates in the 3/8-9/16 in. (0.95-1.43 cm) W.T. piping averaged 30-40 mils/year (0-076-0.102 cm/year) with some pitting rates observed as high as 75 mils/year (0.191 cm/year).
The corrosion was most prevalent and severe in the bottom of stagnant or low-flow pipe where process water could precipitate out and support colonies of sulfate-reducing bacteria (SRB).
Severity appeared related to pipe geometry and the ability of dead legs to accumulate water. Corrosion was very random in distribution and intensity, which is characteristic of microbial-induced corrosion (MIC). Water samples taken from valve-body drains and sumps confirmed the presence of high concentrations (i.e., 106 SRB/ml) of planktonic SRB and high hydrogenase activity.
In 1990, several pieces of severely damaged pipe were removed from service. Active MIC inside the pipe was confirmed when fresh surface-deposit samples revealed a sludge-like mass containing high levels of sulfate-reducing and other anaerobic bacteria (i.e., slime formers).
Dissolution products of corrosion, that is, FeS and Fe(OH)2, were also detected with the bacteria-rich sludge.
The MIC damage could be classified into two broad categories: areas showing isolated discrete pits 0.5-1.0 in. (1.27-2.54 cm) in diameter covered with a dense barnacle-like cap of iron sulfide; and areas showing broad slime-covered troughs of connected pits, 1-2 in. (2.54-5.08 cm) wide and up to several feet (1 m) long.
Corrosion was generally located in a 15 arc along the bottom of pipe.
Fig. 2 shows examples of damaged pipe immediately after removal.
Fig. 2a shows two parallel tracks of in situ sludge and corrosion debris on the bottom of freshly drained pipe.
Fig. 2b shows fresh slime-covered SRB colonies which developed along the corrosion tracks. The colonies were typically 1-2 in. wide and of variable length.
Fig. 2c shows a broader, well-defined, slime-covered SRB colony, where water congregated at a weld dam. In this instance, the colony spanned both corrosion tracks along with the bottom of the pipe.
Fig. 3 shows examples of damaged pitted pipe after cleaning.
Fig. 3a shows large areas of deep, 125+ mil (0.318+ cm) pitting concentrated along the bottom of the pipe. Fig. 3b shows how pits become smaller and more isolated further up the pipe wall above the 6 o'clock position.
Fig. 3c shows examples of two deep connected troughs of pitting along the bottom of the pipe.
Troughs were observed in low points and syphons where MIC colonies remained constantly wetted with water. This phenomenon was also frequently observed where water became dammed between protruding interior weld beads.
Direct observation of damaged pipe revealed no carbonate scale formation on the pipe interior. Away from MIC-affected areas, the pipe wall was generally close to original nominal thickness. This indicated low general corrosion rates and confirmed that the corrosion problems were primarily microbial in nature.
The high potential future repair costs either by pipe repair or replacement dictated rapid implementation of inhibitor mitigation work.
DESIGN CRITERIA
Inhibitor systems were required to meet the following design criteria:
- The system must be effective in terms of corrosion control and include provisions to monitor effectiveness.
- The system must be rapidly installed at all 12 TAPS pump stations, approximately 65 miles (105 km) apart.
- The system must have minimum impact on existing pump-station operations and equipment.
- The system must be economical in terms of capital and operating costs.
To indicate the best inhibitor-application methods, the stagnant areas of the crude-oil system were categorized as large-bore, intermittent-flow legs; small-bore, intermittent-flow legs; true "dead legs" (or natural pipe syphons); and tanks and sumps.
Typical configurations of these dead-leg groups are shown in Fig. 4.
For effective control of the internal corrosion problems identified, each pump station required treatment of approximately 12 discrete pipe segments, one large 55,000-210,000 bbl (8,800-33,500 cu m) relief tank, and three smaller 49 bbl (7.8 cu m) sump tanks. At the 12 pump stations, a total of 129 potential injection points and 184 potential monitoring points were identified.
INHIBITORS, EQUIPMENT
The feasibility of using various batch or continuous injection schemes was evaluated. Major variable costs were inhibitor chemicals, injection labor, and equipment maintenance.
The largest component of variable cost, which influenced system selection, was annual inhibitor consumption. This depended on treatment concentration and frequency.
After extensive discussions with manufacturers, inhibitors with good film persistence and incidental biocidal properties were selected.
With these products, a protective film could be established and maintained in the intermittent-flow lines if batch treated at 1,000 ppm (mg/l.) biweekly, or at 50 ppm (mg/l.) continuously.
Manual batch treatment could be accomplished with a multipoint mobile system, where each dead leg could be individually dosed at 1,000 ppm (mg/l.) while oil flow was throttled through process valves at low flow rates. Biweekly dosing at 1,000 ppm (mg/l.) would require 6-8 hr to accomplish.
Automatic batch treatment could be accomplished with a single point, fixed system by injection of inhibitor at 1,000 ppm (mg/l.) into the main line pump suction and sequential diversion of inhibited oil through the dead legs.
Biweekly dosing at 1,000 ppm (mg/l.) would require 1 hr to accomplish with an automatic valve-sequencing controller. Automatic continuous treatment could be accomplished with a single-point, fixed system by periodic injection of inhibitor at 50 ppm (mg/l.) into the main line pump suction and the bypassing of inhibited oil through the dead legs.
Daily dosing, four times per day, at 50 ppm (mg/l.) would require 4 hr to accomplish with an automatic valve-sequencing controller.
The following systems were evaluated:
- Manual batch treatment (multipoint)
- Automatic batch treatment (single point)
- Automatic continuous treatment (single point).
Table 1 shows a comparison of these treatment systems.
The manual batch system was selected for treatment of the intermittent-flow dead legs because of low inhibitor consumption and low capital costs. A multipoint, mobile system also provided more long-term flexibility in application of different combinations of inhibitors to the dead legs.
Mobile manual batch-treatment equipment could also be used in true dead leg and tank applications. The box above shows the equipment required for a manual batch-treatment system.
IMPLEMENTATION
Final design of the inhibitor application system was initiated in the fall of 1989 and inhibitor introduced into first dead leg in February 1990.
In most cases, the dead legs and tanks were not previously equipped with existing small-diameter fittings which could be easily modified for inhibitor injection or coupon insertion.
Because of the large number of fittings required (129 injection points and 184 monitoring points), we selected Rohrback-Cosaco 2-in., 6,000-psi (41,370-kPa) access fittings for installing pipe penetrations. These fittings were welded in place, the welds dye checked and pressure tested, and a 1 3/8-in. (3.49-cm) diameter hole cut in the pressurized pipe.
With these fittings, the inside of the pipe can be accessed under pressure with a retriever tool. The fittings were also equipped with a side tee and valve to facilitate injecting inhibitor or withdrawing fluid samples.
Where fittings were used for injection, they were generally placed on the top of pipe and stainless-steel quills inserted midway into the flow stream. Wherever possible, multiple injection fittings were tied together by manifolds made from 3/4 in. (1.91 cm) stainless-steel pipe. Injection points were fitted with quick-disconnect "dry break" fittings.
A typical dual injection and monitoring fitting location is shown in Fig. 5.
Where fittings were used for monitoring, coupons or electric resistance probes were installed. Where located outside, a temporary protective enclosure was provided around the fitting. These enclosures could be warmed with a portable heater when temperatures dropped below -20 F. (-29 C.).
The corrosion treatment typically required injection of 5-20 gal (19-76 l.) of concentrated inhibitor into the intermittent-flow leg during low flow.
True dead legs were treated by displacement of crude with a water, methanol, and inhibitor mixture.
These treatments required inhibitor, premixed in bulk, in 300-5,000 gal batches. The concentrated and premixed bulk inhibitors are shipped, staged, and applied in the following manner.
The concentrated inhibitor was shipped by road in reusable 200 and 400 gal stainless-steel containers using a licensed commercial carrier. The inhibitor is freeze-thaw stable, with a freezing point of -50 F. (-46 C.).
The material is held in outside storage depots at each pump station. Because the materials are classified as flammable hazardous liquids, outdoor storage sites must have linings and dikes and be located away from ignition sources.
The containers are brought indoors to warm up 24 hr before use. The concentrate injection truck contains a 3 gpm (11.4 l./min), high-pressure electric pump, and 30-kw diesel generator set.
On the day of treatment, the concentrate tanks are loaded onto a pump truck and moved around the site to various injection manifolds. Because several different inhibitors are used, the truck was sized to hold two portable containers inside the insulated heated van.
When finished at a pump station, the tanks are removed from the injection truck and returned to the pump station's storage depot. The empty pump truck then proceeds to the next station and loads inhibitor from the next storage depot.
Pump station depots minimized the risk of continuous movement of hazardous liquids between stations on rough arctic gravel roads in the injection truck. The injection truck, therefore, did not have to be licensed and certified to haul flammable hazardous-material liquids on public roads.
The bulk inhibitor used for displacement was premixed in 5,000 gal (19,375 l.) batches with 50:50 methanol-water solvent at a bulk yard in Anchorage. The premix was transferred to a 6,300-gal tank trailer and hauled to the remote pump stations.
The bulk trailer contained its own 15 gpm (58.1 l./min), high-pressure pump and 50-kw diesel generator set. The trailer was rated to carry flammable, hazardous liquids, and the tractor and driver were provided by a licensed, insured, commercial trucking firm.
ACKNOWLEDGMENT
The author wishes to acknowledge the following individuals who contributed in the design and implementation of this project: R. W. Ader and G. W. Pomeroy, Alyeska Pipeline Service Co., provided project engineering services; H. G. Byars, ARCO and L. A. Disbrow, CTI Inc., provided corrosion engineering methodology; D. Crawford, Nalco Chemical, provided inhibitor technology; D. Thorpe, TTS Inc., provided inspection services; I. Vance, D. Brink, B. T. Waterman, and H. A. Cash, BP America Research, provided microbiological technology.
REFERENCES
- Byars, H.G., and Gallop, B.R., An Approach to the Reporting and Evaluation of Corrosion Coupon Exposure Results, Materials Performance, Vol. 27, No. 11, November 1975, pp. 9-16.
- Injectech Inc., Lactate/acetate SRB Test, Ochelata, Okla.
- Gen-Probe Inc., Use of RNA Probes in Detection of Sulfate Reducing Bacteria, Poster 16, Hogan, J. et al., San Diego.
- Ramsco Inc., API Sand Medium Test, Soldotna, Alaska.
- Caproco Ltd., Hydrogenase Test, Edmonton, Alta.
Copyright 1991 Oil & Gas Journal. All Rights Reserved.
