Coal, O&M

Inspecting for Corrosion Fatigue

Issue 6 and Volume 112.

Recent catastrophic failures worldwide are prompting a second look at the inspection tools and methods for mitigating corrosion fatigue.

By Eric Thurston, Corporate Senior Metallurgical Engineer, EON-US Generation Engineering Group

Power boiler startup and shutdown cycling and proper water chemistry are paramount in extending the life and reliability of the unit. Loss of control of either of these processes can cause cracking of the protective magnetite scale on the inside surface of the boiler tube so that the unprotected surfaces create corrosion initiation sites. Damage to, or failure of, a metal due to corrosion combined with fluctuating fatigue stresses—also know as corrosion fatigue—can then ensue.

The bulk of corrosion fatigue failures are usually “pinhole” in nature, however some have been catastrophic, with injury to personnel. The recent catastrophic failures worldwide have prompted a second look at the inspection tools and methods for mitigating corrosion fatigue.

Using video probe technology, along with existing tolerable flaw diagrams for the diameter of boiler tubes inspected, this case study illustrates and validates the inspection tools and method used to mitigate corrosion fatigue in a 49-year-old boiler.

The boiler is a pulverized coal-fired radiant model with balanced draft furnace. The design maximum continuous steam flow is 825,000 lb/hr main steam at a secondary superheat outlet pressure of 1,770 psig. and 730,000 lb/hr reheat steam flow at a reheater superheater outlet pressure of 675 psig. Main steam and reheat steam temperatures are controlled by excess air and attemperation at 1,005 F. The design steam temperatures are based upon a normal feedwater temperature of 470 F. The unit has run approximately 340,000 hours year to date and has logged 762 starts. The boiler went into commercial operation in 1959.

The unit has a past history of running with severe condenser leaks. It was normal practice at this facility 30 to 40 years ago to treat the boiler with a coordinated phosphate treatment system, with up to 20 ppm phosphate to offset the pH adjustment due to condenser leaks. Mono-phosphate, di-phosphate and tri-phosphates were used in this system. In the early 1990’s, the coordinated phosphate system was switched to a congruent phosphate to follow EPRI guidelines. In 1997 the plant switched to a low phosphate system, which is typical in today’s European boiler systems. The phosphates are currently controlled at 0.5 ppm. The unit also has a past history of hydrogen damage around the burner regions.

Another reason for inspecting this unit was a failure that occurred in the penthouse in 2005. The failure’s location was determined to be at the high crown seal where flat bar was welded to the crown of the division waterwall tubing (Figure 1). As a result of radiographic testing (RT), 23 dutchmen were removed from the division and side walls. The metallurgical results of this failure identified corrosion fatigue as the damage mechanism.

Figure 1
Corrosion fatigue failure in penthouse at division wall of the investigated boiler in 2005
Click here to enlarge image

The boiler inspection phases were prioritized to focus the scope for the four-week outage. Phase 1 consisted of inspecting all tubing outside of the firebox, in particular down comers, supply tubes and risers. Phase 2 consisted of inspecting all tubing within the firebox, in particular around buck stays, wind box and rigid attachments. Phase 3 consisted of thick walled locations such as drums and headers. The priority of inspection for the unit was heavily based on historical failures documented by metallurgical analysis. In all phases, visual inspection was performed and 84 percent of the water-touched boiler was viewed. Sections of the boiler that were not inspected were not in personnel space or had recently been replaced.

Three criteria were used for characterizing the corrosion fatigue indications found. Priority 1 was designated as corrosion fatigue that met or exceeded the maximum tolerable flaw dimension for the size of boiler tubing inspected and was slated to be removed during the current outage. Priority 2 signified that there was corrosion fatigue present; however, it did not meet the rejection criteria of a Priority 1 and either will be re-inspected or removed during the next outage in two years. A Priority 3 designation was used to document areas where other damage mechanisms, either active or dormant, were located. Three hundred and thirty-five Priority 3 indications were observed in the boiler.

The original equipment manufacturer (OEM) was used in the inspection and repair of this boiler. Having access to the assembly drawings and technical expertise of the OEM regarding tear down and reassembly was the winning combination needed to perform this corrosion fatigue outage in the time allotted. It was also helpful working with the OEM to determine the best type of inspection technology to use through an engineering study report. A videoprobe was used to inspect 84 pecent of the boiler water-touched internals. Using videoprobe technology in the 2-31/32” diameter tubes allowed for minimal cost of scaffolding in the boiler. Figure 2 (page 46) shows a view of the videoprobe in the 2-31/32” waterwall tube. The corrosion fatigue cracking highlighted at the 10 o’clock position in the figure is the 32nd tube from the rear wall.

Figure 2
Videoprobe view of corrosion fatigue in a straight run tube on the rear wall.
Click here to enlarge image

After the indications were found and sized, 3”-diameter tube tolerable flaw diagrams were used to determine the critical flaw length. This diameter of tubing was the closest to the tubing used in this vintage boiler. The tubing used in the waterwalls for this boiler was 2-31/32” x 0.203 min. wall SA-178-C material. The critical flaw length was determined to be 3” in length. Several matrices were assembled to give an overall 3-D map of the boiler and the location of the corrosion fatigue indications. Twenty-one tubes met the rejectable criteria, while many other tubes with smaller cracks were found. Those with smaller cracks will be either inspected in two years for crack growth rates or removed. Fifty-eight tubes in all were removed and sectioned. Many of these removed samples met the Priority 2 criteria. Since repair equipment was readily available, the tubes were also extracted. These were used in a study to determine validation of the videoprobe analysis.

The validation study showed that in the hands of a skilled operator, accurate assessments of crack lengths can be made with a videoprobe. More importantly, corrosion fatigue was mitigated in high traffic areas, such as the tube in Figure 3. This rear waterwall tube ruptured when it was clamped with minimal force into the horizontal band saw for metallurgical sectioning. Subsequent measurements confirmed a minimum wall thickness of 0.036” at the rupture site.

Figure 3
Rear waterwall tube that was connected to a buckstay above the lower slope at the main floor elevation which split open during sectioning for metallurgical analysis. Corrosion fatigue cracking was measured at 3.25” in length. The measurement of the remaining wall thickness after overload was 0.036”.
Click here to enlarge image


Other Targeted Locations

Other locations were targeted. Wall tubes where the upper crown seal and side wall casing came within five feet of each other were found to contain instances of corrosion fatigue. Waterwall tubes where wall restoration was performed many years ago were found to have corrosion fatigue. Figure 4 (page 50) illustrates a wall restoration where no blow-through was observed. However, the residual stresses created when the weld repair was made, coupled with past boiler chemistries, initiated a corrosion fatigue site. Lower side wall header drains were also targeted due to the differences in thermal expansion from one header to the next. Furthermore, load-bearing tubes, such as the water-cooled hanger tubes, were inspected and found to exhibit corrosion fatigue at hanger attachment lugs in the penthouse.

Figure 4
Wall restoration of waterwall tube led to increased residual stresses creating corrosion fatigue
Click here to enlarge image

The boiler repair plan was to replace all the supply and riser tubes that exhibited indications since these were the highest safety priority in the OEM engineering report. All Priority 1 tubes with corrosion fatigue defects on every wall were replaced, since these have a high probability of failure. Also, the repair locations were heavily biased toward personnel high-traffic areas. The left and right wall Priority 2 indications were completely addressed due to the earlier timing of repairs in the project. The penthouse water cooled hanger tubes were entirely addressed to complete all penthouse repairs. The front wall at the lower windbox corners near the main floor was also addressed due to high traffic.

From a safety perspective, many of these damage mechanisms were found in locations on the boiler in high-traffic areas. Rejectable Priority 1 corrosion fatigue indications were found in the corners of the windbox located on the main floor level where personnel walk every day. The west corner of the windbox is around 15 feet from the plant’s most-used elevator. Also, the transition from the lower slope to the rear wall also showed rejectable indications that were mitigated. This is also in a high-traffic area on the main floor.

In conclusion, the use of current videoprobe technology in capable operator hands can accurately and efficiently map corrosion fatigue in older-vintage boilers on the ID tube surface. Furthermore, boilers of this design and vintage will most likely have active corrosion fatigue due to hard attachments, which was typical of boiler construction at the time. When two or more attachments are in close proximity (five feet or less), the potential for corrosion fatigue greatly increases. Load bearing tubes, such as water-cooled hanger tubes, are much more susceptible to corrosion fatigue than non-loaded tubing. As a result of the inspection and repair of the unit, corrosion fatigue has been mitigated and the safety and reliability of the boiler has been improved.

Author: Eric Thurston is the corporate Senior Metallurgical Engineer with the EON-US Generation Engineering Group. He has previous experience as a boiler tube failure consultant and quality engineer. Thurston has authored two papers for NACE on boiler inspection techniques and mechanical cleaning of deposit weight density samples, and co-authored “Review of Two Case Histories Involving Boiler Tube Failures” presented at the Electric Utility Chemistry Workshop. Mr. Thurston holds a BSE degree from the University of Cincinnati in Materials Engineering and is a Certified Welding Inspector. He is a member of ASM and AWS. Other contributors to this project and article are Tom Troost, plant general manager and Russell Baker, mechanical engineer.