Dec. 2, 1991
K. R. Ferguson Chevron U.S.A. Inc. Midland, Tex. A. G. Stutheit Chevron U.S.A. Inc. Lafayette, La. Use of complete in situ weld overlay to repair large process vessels in corrosive service has proven cost effective and technically feasible. The dual units are located in a hot potassium carbonate CO2 treating plant located on the Sacroc unit near Snyder, Tex. The unit is operated by Chevron U.S.A. Inc.
K. R. Ferguson
Chevron U.S.A. Inc.
Midland, Tex.
A. G. Stutheit
Chevron U.S.A. Inc.
Lafayette, La.

Use of complete in situ weld overlay to repair large process vessels in corrosive service has proven cost effective and technically feasible.

The dual units are located in a hot potassium carbonate CO2 treating plant located on the Sacroc unit near Snyder, Tex. The unit is operated by Chevron U.S.A. Inc.

This procedure was performed in lieu of other corrosion solutions or repair alternatives that offered less security, worse economics, or both. Both vessels were reentered after at least 1 year of operation after repair and were found to have excellent overlay integrity.

Past industry experience in related repair methods (such as strip lining) suggests ultimate failure is probable. Many refineries and gas treating and petrochemical facilities may be plagued by similar problems calling for relatively large and thorough retrofitting of such pressure vessels.

This project is presented in a series of three articles which discusses in detail the various planning, execution, and safety considerations required successfully to carry out the field activities of this 6-month long project.

This initial article will present a background description and a brief summary of the various alternatives considered. The second article will discuss project evaluation and preparation for the on site work.

The final article will provide a description of the actual project execution and results.


The Canyon Reef formation was originally tapped in 1949 and had produced its first 1 billion bbl of oil by 1973.

After this producing formation was unitized in 1953, efforts to slow the production decline of this reservoir with water injection began in 1954, followed by CO2 injection in 1972.

In order to separate the resulting CO2 being produced from the reservoir, the Sacroc CO2 plant was constructed to coincide with the start Of CO2 injection in 1972. Sun Exploration & Production, now a unit of Oryx Energy, Dallas, operated the plant.

Two other plants, the North Snyder gas plant (NSGP) and the BHP (formerly Monsanto) plant were later constructed to separate CO2 produced in the north and south areas of the unit, respectively.

All three plants were constructed immediately upstream of Sun-operated hydrocarbon gas plants existing when CO2 injection began. Fig. 1 shows the location of the CO2 and hydrocarbon plants in the Sacroc unit.

The Sacroc-Sun plant was built in front of the hydrocarbon processing plant which was constructed in 1954. This plant is located 4 miles northwest of Snyder in Scurry County, Tex., and throughout its operational history has handled the bulk of the gas (70-95%) produced from the Sacroc unit.

Because the Sacroc/Sun CO2 facility "front-ended" the Sun hydrocarbon plant from both physical and process standpoints, Sun was selected to operate this "unit-owned" facility to reduce manpower and increase processing effectiveness.

The plant consists of two identical and independent process systems referred to as the 100 and 200 Trains (Fig. 2).

Over the past several years, the amount of H2S in the produced gas has increased to levels which destroyed the inherent process inhibition of carbon-steel equipment, causing severe corrosion in the plant and posing both safety and operational hazards.

The overlay project sought to eliminate the corrosion-related safety and operational hazards at the Sacroc-Sun CO2 plant by upgrading the metallurgy of the absorber columns and carbonate piping.

(Upgrading the metallurgy refers to overlaying the inside of the absorber columns with stainless steel and replacing the rich-carbonate piping with stainless steel pipe.)

This project can be considered the final phase in a complete corrosion proofing of this plant; the other major components of the plant, the stripper columns, were replaced with stainless-clad vessels in 1986.


The primary mechanism for CO2 separation at the Sacroc-Sun facility is hot potassium-carbonate absorption based on the Benfield process, which is inherently corrosive.

The original plant was designed on the basis of 12.5% CO2 in the produced gas. In 1974, 2 years later, the plant was expanded to accommodate 24% CO2 in the produced gas, but this expansion addressed only plant capacity, not the basic plant metallurgy.

CO2 and H2S levels in the feed gas had been rising to approximately 52% and 1,200 ppm, respectively, since the plant was constructed. This condition increased corrosion potential to levels which challenged the integrity of the primarily carbon-steel metallurgy and inhibition process in this facility.

The onset of H2S production from the reservoir was not originally foreseen as a design constraint. Yet it has an extremely important effect on the dependability of the process.

Vanadium pentoxide is introduced into the Benfield process as a corrosion inhibitor, forming a tightly bonded iron-oxide protective scale over the surfaces of the carbon-steel process components. This inhibition process is quite effective at low levels of H2S.

But when the concentration of H2S in the feed gas approaches higher levels (1001,000 ppm, for example), the tendency of the H2S to form an iron sulfide-iron pyrite scale competes with and effectively negates the protective scale-forming abilities of the vanadium pentoxide inhibitor.

The sulfide scales formed by the presence of H2S are substantially less protective than the oxide scale and therefore subject the entire carbon-steel process system to a pitting corrosive attack in areas where a premium scale cannot be established.

Of particular concern was the mechanical condition of the two absorber columns at the Sacroc-Sun CO2 plant which had exhibited significant localized deterioration because of pitting corrosion since 1985.

These two identical columns (Fig. 3) are 115 ft high, 12 ft in diameter, have a wall thickness of approximately 3 in., and comprise the first stage in the "hot pot" process in which high-pressure produced gas is physically contacted with lean, hot potassium carbonate.

Absorption of the CO2 into the potassium-carbonate stream occurs in the absorbers and is later released from the carbonate in the low-pressure stripper columns. The failure of a similar absorber column operating under nearly identical process conditions at the NSGP in 1984 and 1985 pointed out the need for remedial action at the Sacroc-Sun facility.

The latter column experienced a localized-pitting corrosion rate of at least 4,000 mils/year (mpy), which resulted in perforation of the 2 in. thick vessel wall over a period of 6 months or less. Although this failure did not result in any injury or loss of life, it demonstrated the need for attention to the potential problem which existed at Sacroc-Sun.


Several possible solutions were considered (Table 1):

  • Conversion of the process to a less corrosive solvent, namely methyl diethanolamine (MDEA)

  • Removal of H2S from the feed-gas stream by addition of a separate H2S-removal facility

  • Replacement of the existing carbon-steel absorber vessels with stainless steel-clad vessels

  • Covering the internal surfaces of the absorber columns with an array of stainless-steel sheets ("strip-lining")

  • Weld overlay of the existing absorber vessels with a continuously fused layer of stainless steel.


Conversion of the process to a different, less corrosive solvent, namely MDEA, was evaluated as the primary option because it had the additional potential of increasing the operating efficiency of the plant. A detailed design study of this option, however, indicated that it was infeasible for three reasons:

  1. The high cost of retrofitting an existing plant with strict space and operational constraints

  2. Selection of a generic, untested solvent with known operational complexities (for example, foaming)

  3. Difficulty in quantifying energy-saving characteristics due to the retrofit nature.


At first glance, removal of the H2S from the feed gas upstream of the hot-pot facility seemed like an attractive process alternative because elimination of the H2S from the gas stream entering the CO2-removal facility would reestablish the effectiveness of the vanadium pentoxide corrosion inhibitor in the hot-pot facility and essentially resolve the corrosion problem.

Further investigation of this alternative, however, revealed the cost of an H2S-removal facility to be prohibitive when compared to all other options mentioned here.

Complete H2S-removal facilities designed to handle the volume of gas processed by the Sacroc-Sun plant were found to have price tags as high as $80 million, compared to a solvent-conversion cost of approximately $11 million.


At this point (December 1987), the next best alternative to solvent changeout and H2S removal appeared to be replacement of the absorber columns with stainless steel-clad vessels. An appropriation was written to perform this changeout.

But by the time bids were received for fabrication of the new vessels, radical conditions in the strategic metals market (particularly with respect to nickel and chromium) had caused the cost and delivery of stainless steel to rise to levels which prompted a reevaluation of this option.

The estimated cost of the clad columns had nearly doubled during the 6-month design and bid period, and the delivery had been extended to more than 1 year for a single vessel.

Because of this volatility in the strategic metals market (as well as the stainless-steel market), vessel manufacturers would quote neither firm prices nor deliveries. Since the rate of corrosion in the Sacroc-Sun absorber vessels could approach 4,000 mpy, perforation of one or both of these vessels could occur before the first new vessel was received. These factors eliminated vessel replacement as a viable option.


In late 1984 and twice in 1985, Sacroc was forced to make repairs to a similar absorber column at another CO2 plant in the field because of corrosion failure. Late in 1985, a final repair was made to a column by strip-lining an area of 150 sq ft with 316L stainless-steel sheets measuring 10 in. x 30 in.

Such strip-lining is accomplished by continuous welding of panel joints to the base material (Fig. 4).

In addition to absorber column repairs, Sacroc had utilized strip lining for repairing low pressure "stripper" columns which release the absorbed CO2 from the hot potassium carbonate solution.

Sacroc's experience with such techniques had been mixed over the years; until 1984, however, it was felt to be an adequate repair in some services (that is, local areas of stripper columns and reboilers).

But about this time, it became increasingly evident to Chevron that permanent repair could not be accomplished by the strip-lining technique.

Both Chevron's experience and that of the industry in general' indicated susceptibility to lining leakage and the phenomenon known as stress corrosion cracking (SCC) resulting from concentrated process fluid leaking behind the lining panels.

Recent cracks had been discovered in samples from a Chevron refinery vessel in similar service.

For this reason, the strip lining was applied at Sacroc's NSGP with the express intent of its being temporary (less than or equal to 2 years' duration) to a continuing corrosion problem.

Later tests from both NSGP and the small panels installed in one of the Sacroc-Sun absorbers in early 1988 confirmed the beginnings of detectable cracking of the base material.

The latter case (Sacroc-Sun plant) was determined at the time of preparation for the first column overlay, thus confirming the need to alter methods of repair.


At this point, reexamination of the alternatives indicated that in situ weld overlay of the existing vessels with a continuously fused layer of stainless steel was the most attractive option both in terms of cost and schedule.

Several major obstacles had to be overcome, however, before attempting this project because vessels of this size (including all nozzles) had never been entirely overlaid in the field before, particularly under conditions such as those which would exist during this project.

As mentioned previously, the Sacroc-Sun CO2 plant consists of two identical process trains which operate in parallel to remove 52 MMscfd Of CO2. These process trains are physically located adjacent to each other and share a common main pipe rack.

The absorber column vessels themselves are separated by a distance of only 7 ft. The proximity of the active absorber vessel processing high-pressure (500 psig) natural gas to the inactive absorber in which arc-gouging and welding would be taking place for approximately 2 months posed some very serious safety considerations.

However, after investigation of the weld-overlay process, supported by technical presentations by several contractors qualified to perform the operation, it was determined that such an endeavor would be feasible from both technical and safety standpoints. In addition, engineering estimates from several overlay contractors indicated that the cost of weld overlay was the most economically viable option of all considered.

At this point (February 1988), an appropriation was written to request a scope change and adjustment to funds required for weld overlay of the absorber columns.


  1. Kloff, S.W. "Corrosion of a CO2 Absorber Tower," Plant/Operation Progress, Vol. 5, No. 2 (April 1986), pp. 65-72.

Copyright 1991 Oil & Gas Journal. All Rights Reserved.