K. R. Ferguson
Chevron U.S.A. Inc.
Midland, Tex.
A. G. Stutheit
Chevron U.S.A. Inc.
Lafayette, La.
Chevron U.S.A. Inc. has used a unique in situ weld-overlay procedure to repair large process vessels in a hot potassium carbonate CO2 treating plant. The plant is located on the Sacroc unit near Snyder, Tex.
Planning and preparation for the unique application of the process have been the subjects of two earlier articles (OGJ, Dec. 2, 1991, p. 54; Jan. 6, p. 69). This final one details the project's procedures and execution.
VESSEL PREPARATION
With all instrumentation removed from the absorber vessel, the first step in the overlay process was the removal of all internal hardware (that is, bubble-cap trays, tray rings, downcomer bars, distributor piping, ladder rungs, and vortex breakers).
At the same time, any nozzles or couplings too small to be overlaid were replaced with solid stainless-steel nozzles. All such nozzles were replaced according to National Board Inspection Code (NBIC) repair procedures, which required local hydrotesting and dye-penetrant testing (PT) of each nozzle.
Hydrotesting and dye-penetrant testing were witnessed and approved by a certified vessel inspector. The internal projections of the larger nozzles had to be contoured to a relatively smooth surface by arc gouging.
Fig. 1 depicts the final overlaid surface adjacent to one of the larger (10 in.) nozzles. This nozzle had an original internal projection of about 1 in., which had to be contoured as illustrated prior to overlay.
The internal surface of the nozzle itself was manually overlaid. In addition to the inside surface of the larger nozzles being contoured, all corrosion pitting and damage caused by tray-ring removal had to be gouged out, then built back up with mild steel.
Any such buildup was conducted according to NBIC requirements and subjected to dye-penetrant and magnetic-particle testing (MT).
Following the completion of this surface-preparation process, the entire internal surface of the vessel was sandblasted to a National Association of Corrosion Engineers (NACE) white-metal surface (SSPC-SP-5) to accommodate the weld-overlay process.
INTEGRITY ANALYSIS
Part of the preparation of the vessels for the application of weld overlay was a thorough examination of the vessel to assess its mechanical integrity.
Early in the project, Chevron determined the need for random analysis of the original fabricated joints in the vessels. Local records had indicated minimal examination, and overlay stresses were going to be imparted to the vessels via the welding process (albeit quite small).
In addition to the replacement of all smaller nozzles, each connection in the vessel was inspected with ultrasonic (UT), dye-penetrant, and magnetic-particle techniques to ensure their mechanical integrity prior to overlay. Any flaws which might cause problems were removed and repaired.
Before actual overlay operations, the vessel wall had to be smoothed and contoured to a surface which could be handled by the automatic welding machines.
The removal of heavy scale in certain areas, primarily in the lower part of the columns, suggested the vessel needed to be examined for any condition that might undermine overlay quality.
Initial visual inspection of the internal surfaces of the original fabrication welds indicated a general condition inferior to Chevron's current standards for vessels of this class.
Some porosity was indicated, and weld dimensions generally lacked uniformity. Surface inspection (MT and PT) of one of the 10-in. nozzle attachment welds indicated cracking which required at least cosmetic repair prior to overlay. This initial inspection led Chevron to pursue more thorough examination of all nozzle welds.
Because of the specimens' geometries, nozzle examination primarily took the form of UT analysis. Some radiography (RT) was performed on one nozzle to attempt further to define UT indications.
Finally, a proprietary method of UT analysis was applied to a small number of specified nozzles in a final effort to determine cracking potential of certain fabrication welds. Evolution of the methods used and criteria applied will be detailed in a later section.
Nozzle integrity analysis could involve as many as four steps:
- Initial surface evaluation by NIT
- Scoping of internal evaluation by UT according to ASME Boiler and Pressure Vessel (BPV) Code UW-53 and Appendix 12
- Specific evaluation by RT with increased exposure time and cobalt or iridium source material
- Evaluation by a proprietary UT process to verify the presence of planar (crack propagating) or volumetric (nonpropogating) flaw characteristics.
Step 3 was only utilized in one instance on initial evaluation of the first column. Steps 1 and 2 were performed on all existing nozzles (approximately 25) and Step 4 was performed as needed (on 8 nozzles).
TESTING
Magnetic particle testing (MT) was performed primarily to identify surface cracks which might preclude the successful application of weld overlay, In addition, to ensure that cracks indicated by this method did not affect overall nozzle integrity, dye-penetrant testing was used after initial identification and during grinding and gouging to remove cracks.
At the onset of examination work, an uncalibrated high sensitive gain was set on the ultrasonic instrumentation to discover the presence and general location of any weld defects.
Both straight beam and shear-beam analysis techniques were utilized by an American Society of Nondestructive Testing Level 2 technician to assist in exact location. After initial scoping evaluation indicated several questionable areas--determined by comparison to a standard "distance amplitude correction" (DAC) curve-- it was apparent that the severity of indications could not be adequately determined with the inordinately high gain used for identification.
Chevron decided to apply criteria as though the work were new construction; thus an angle-beam calibration block was constructed according to BPV Code Section V, Article 5. This would enable the relative flaw orientation and size to be more accurately measured.
Reexamination against the new DAC curve generated by the calibration block indicated that most of the nozzles were free of critical flaws. One nozzle, however, continued to be of concern.
At this time radiographic testing (RT) was utilized to learn more about the UT flaw indication (whether planar or volumetric). Because of nozzle geometry, however, the RT could not be interpreted according to normal code criteria, and the results were inconclusive.
An additional UT examination was performed with a proprietary UT technique developed by Southwest Research Inc. (SwRI) of San Antonio.
SwRI employs a unique combination of enhanced equipment and operator skill as a means of differentiating between volumetric (nonpropogating) and planar (propagating) flaws.
This technology, which may be commonly referred to as "satellite-pulse" technology, distinguishes secondary reflective waves (tipdiffracted waves) which may be found in conjunction with a planar flaw (crack), from those reflected pulses normally associated with volumetric flaws (such as slag and porosity).
TIME-RELATED PULSES
SwRI discovered that tipdiffracted, or secondary reflected waves, revolving around the flaw in satellite fashion, indicate a unique set of time-related pulses which are also captured on the oscilloscope as amplitude-re-lated information.
The amplitude is not as important as the time relationship indicating the presence of a tip or crack-related reflection. The satellite-pulse techniques used in conjunction with standard 45 shear-wave UT analysis and varying frequencies of transmitted UT pulses proved valuable.
A generally high level of training and skill are required for such a technique. A Level 3 certified NDE inspector was used for the analysis.
The application of the SwRI technology aided in interpreting that the flaws were volumetric and nonpropagating. This was confirmed by relative analysis of X rays showing the same two indications at slightly different orientations.
Special UT and RT indicated that the nozzle flaws were indeed volumetric and as such required no repair.
Similar analysis using Steps 1 and 2 described previously were performed on the second column on all nozzles, with Step 4 performed on only four questionable nozzles.
All but one of the nozzles met the criteria for the pulse-satellite technology and were determined to be nonpropogating. One of the nozzles in the second vessel indicated a flaw of questionable size but which appeared to be volumetric as well.
Supported by the long and acceptable operational history of this vessel, Sacroc decided to restore it to service without repair.
As a precaution, Chevron has performed UT analysis with a second proprietary SwRI system which makes a permanent record of the indication. The precise application of the permanent recording of the nozzle condition will allow comparison in 18-24 months with the same technology to determine whether any propagating tendency is apparent over time.
Although no adverse indication is expected, this action is being conducted as a precaution as part of a long-term pressure vessel compliance and inspection program.
This description portrays the use of three separate techniques which Chevron found valuable for determining that original nozzle and vessel fabrication had survived years of service, should readily accept weld overlay, and thereafter should provide many years of service under harsh operating conditions.
Before the internal surface of the vessel could be overlaid, it had to be preheated to approximately 93 C. (200 F.).
ASME Section VIII, Division 1, permits the use of preheat in lieu of postweld heat treatment. Preheating was accomplished with an oxygen-acetylene torch focused immediately in front of the automatic arc by the welding technician.
OVERLAY PROCESS
The overlay contractor used a total of 10 fully automatic gas-metal-arc welding (GMAW) machines operating simultaneously within the vessel. The welding power supplies were housed in a protective enclosure at the base of the column.
Two extra power supplies were maintained as standby units. The wire feeders and automatic tracking controls were located inside the column.
Two of these units were hung from the upright on each umbrella scaffold. (Umbrella scaffolding will be described later).
As shown in Fig. 2, the automatic welding heads were mounted on horizontal tracks (tack-welded to the shell) that allowed them to traverse the shell horizontally and vertically (with a rack).
In addition to moving up and down the shell, the welding heads moved in and out and automatically tracked the wall contour. Wire feed accelerates to 381 mm/sec (900 in./min) and travel speeds of about 14 mm/sec (34 in./min) were utilized.
All of the cables and hoses which supply the welding machines with power and shielding gas had to enter through existing openings in the vessel walls.
One piece of equipment which made this entire project possible within a reasonable time frame is called an umbrella scaffold. As the name implies, it can be inserted through a manway and then unfolded like an upside-down umbrella.
Up to four of these units were installed in the column and hung from electric hoists, which allowed the welders to move quickly from working level to working level.
Because of the heat and fumes caused by 10 welding machines simultaneously operating in such a confined space, air conditioning and ventilation were extremely important.
The overlay contractor addressed this situation with two 5-ton air conditioners, one of which was installed on top of the vessel while the other unit was mounted on temporary brackets at the middle of the vessel.
INSPECTION; REFITTING
Because of the localized and high penetration rate of pitting corrosion, it was imperative that the finished stainless-steel surface be free of any defects such as cracks or pinholes which might allow contact of the process stream with the underlying carbon-steel base metal.
Although the strict quality-assurance program employed during the application of the overlay (described in Part 2 of this series; OGJ, Jan. 6) reduced the possibility of such flaws, Chevron felt it necessary that some type of nondestructive examination (NDE) be performed absolutely to ensure the mechanical integrity of the finished surface.
The procedure selected for this inspection was the flourescent liquid-dye penetrant process (WFPT).
The principle behind the WFPT process involves applying a flourescent dye to the surface to be inspected and then washing the dye off the surface with water after it has had a sufficient time to penetrate into any surface imperfections.
After time for the dye to leach back out of these imperfections, the column was darkened and an ultraviolet light was used to locate any flaws or imperfections where dye was evident. Any such indications were marked for later repair and reinspection.
After completion of the overlay, new stainless-steel tray rings, downcomer bars, trays, distribution piping, ladder rungs, and vortex breakers were installed. Fig. 3 presents a top-down view of the new stainless-steel valve trays in the 200 Train absorber column.
At the same time, Sun plant personnel were reconnecting instrumentation, which was accomplished during the final overlay work, allowing expedient recommissioning of the process trains.
Chevron feels that, in the case of the Sacroc-Sun CO2 plant absorber columns, weld overlay was a successful cost-effective approach to major vessel repair.
Copyright 1992 Oil & Gas Journal. All Rights Reserved.