Acoustic emission testing of seam-welded high energy piping systems in fossil power plants.
Ever since the catastrophic failures of seam-welded, hot reheat (HRH) piping at Southern California Edison's Mohave plant in 1985 and Detroit Edison's Monroe plant in 1986, utility companies have been carefully considering the need for periodic inspections of critical piping to guard against creep-induced failures. Figure 1 illustrates the creep-damage mechanisms associated with seam-welded, high-energy piping. A number of serious defects in seamed piping were removed after inspections in the late 1980's, and for a number of years there were no more catastrophic failures . Beginning in 1992, however, there have been six known failures of seam-welded superheat (SH) link piping supplied with Combustion-Engineering boilers, as well as two failures in hot reheat long seamed bends. Two of these have been catastrophic: Virginia Power's Mt. Storm Unit 1 in June 1996, and Kansas City Power & Light's Hawthorne Unit 5 in August 1998. No loss of life occurred in either of those two failures, but the cost of repairs and loss of power generation is of critical concern to utility companies in this age of growing competition. All failures of SH link piping have occurred on units with accumulated service time of 125,000 to 225,000 hrs. Figure 2 shows micrographs of cavitation damage and advanced damage of microcracks from a failed long seam bend [4, 5]. These are truly microscopic defects. Compounding the problem of inspection is the inaccuracy of supplied documentation, which may not reflect the true alloy content and method of fabrication. The Hawthorne SH link piping was not known to be seam-welded. The general aging of fossil plants will continue to raise concerns about the safety of operating seamed high-energy piping systems. Even seamless piping systems have had problems, including creep-related failures of circumferential welds, and the through-wall creep failure of a seamless SH bend that had been improperly fabricated. Current strategies for effectively managing the safety and life of seam-welded piping are based upon periodic inspection of the weld area for evidence of in service damage.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
Equipment and AE Testing Set-up
The process of AE monitoring applied to piping systems starts with installation of AE transducers on welded waveguides (WG) along the length of the piping system. Spacing intervals for the WG are typically 4.6-6.1 m (15-20 ft), and installation of the sensors does not require full removal of piping insulation. There are several unique requirements for successful monitoring of high-energy piping with AE [1, 2]:
* The use of high-frequency sensors (300-400 kHz) and high frequency filtering (>200 kHz) to mitigate the effects of the steam-flow background noise. This noise is predominant below 300 kHz, and would obscure detection and accurate source location if lower-frequency, or broadband, sensors were used.
* The employment of a "floating" or automatic threshold that can control the sensitivity of detection by keeping the voltage threshold of detection above the average background noise.
* The use of active linear source location to determine the position of emitting sources on the line. The accuracy of location is dependent on the distance between sensor/waveguide positions, the pipe diameter, and the position of the emitting source. In the middle of the array between two sensors, accuracy has been demonstrated at [+ or -]2.4 cm ([+ or -]0.6"). Near the sensors accuracy may degrade to [+ or -]60 cm ([+ or -]24"). Still, this limits areas for follow-up inspection.
* The use of active AE-signal-feature filtering to further refine the data and eliminate obvious noise sources, such as flow turbulence. Rise time, duration, and average frequency have proven most valuable.
* Simultaneous recording of piping temperature and pressure are required to provide correlation between active AE sources and the likely source mechanisms.
Data Evaluation and Correlation
The primary characteristics of seam-weld creep-related sources are behavioral in nature--they respond to the pressure in the piping (hoop stress) and other mechanical sources of stress (geometry, hanger supports, etc.). During online conditions with normal peak load cycling, creep-related sources reveal themselves by repeated behavior with each peak load cycle :
* The sources are sensitive to pressure, and may show a pronounced effect of emission rate with pressure (Figs. 3, 4).
* During load cycling, emission rates will typically peak near the start of the peak-load period.
* There is periodic emission activity during steady-state pressure and temperature conditions.
* The AE location profile is typically spread out over 1 m (40") or more of piping length, and shows intermittent high-density locations of activity (Fig. 3).
* The amplitude range of emission sources broadens to higher values with higher activity rates (Fig. 2).
* Emission rates are much higher during startup conditions, even before substantial pressure loading. This demonstrates that the damaged area is responsive to stresses even when the piping is not in the creep regime (>510[degrees]C, 950[degrees]F).
[FIGURE 3 OMITTED]
The amount of emission generated by the creep mechanism, the repetitive nature with each peak load cycle, and the extensive dynamic range (45-90 dB amplitude) of signals, is extraordinarily different from normal ductile fracture mechanisms, such as fatigue crack growth in mild steels. Many thousands of locatable signals are sometimes accumulated over 1 m or so of weld length and several cycles (days) of steam line operation at peak load. The sheer numbers of the sources is inconsistent with a ductile crack growth mechanism, which produces infrequent emission of more limited dynamic range with repeated load cycles. The acknowledged mechanism of creep in seam welds is the development of cavities (cavitation) around nonmetallic inclusions and carbides on the grain boundaries in the fine-grained heat affected zone (HAZ) or fusion zone of the seam weld (Fig. 1). Isolated cavities soon give way to aligned cavitation along grain boundaries, then coalescing into scattered microcracks. Final consolidation and linking into macrocracks along the seam-weld direction occurs in the last stage of growth, which can be very rapid depending on a host of factors (wall thickness, annealing state, inclusion densities, thermal and localized mechanical stresses, etc).
[FIGURE 4 OMITTED]
The early stage of this process involves the degradation of the bonding between particles and the metal matrix. These are load-carrying interfaces, and their eventual failure (decohesion) is the most plausible explanation for the amount and dynamic range of emission detected in the creep process. From the viewpoint of classifying AE behavior, this bears similarity to the experience of monitoring an organic-based composite material that has incurred extensive matrix damage. This also explains the emission that has been noted during the thermal excursion on startups, even without pressure in the system, when the piping is clearly not operating in the creep range. Damaged particle-to-matrix interfaces are prone to disbonding under high strain conditions, and startups are known to produce an even higher axial strain than at full load operation. Indeed, the results of the extensive EPRI field testing program to date has yielded detection of cavitation damaged seam welds that have not developed to the stage of micro- or macrocracking.
A separate test program conducted in collaboration with a UK utility demonstrated that controlled creep-crack growth in small specimens produced increasing emission with increased crack growth rate. The emission rate was orders of magnitude higher for the increment of crack growth than would have been expected at lower temperatures and growth under fatigue conditions. The decohesion mechanism remains active throughout the creep regime, regardless of whether induced by directed stress at the tip of an active crack or in a volume of weld without visible cracking. This leads to high probability of detection of the creep-related failure process from very early stages, well before the damage represents a significant threat of structural failure.
Correlation of AE findings on seam-welded lines with other NDE methods and metallograpy were an important part of the EPRI studies and field tests from 1991 to 2001. Double blind testing was performed on Pacific Gas & Electric's Potrero #3 line in 1994, American Electric Power's Gavin #1 line in 1996, and Sierra Pacific Power's Valmy #2 line over 1997-1999. In these tests good correlation was established between conventional automated and manual multi-angle UT methods and AE cluster locations in seam and girth welds. But metallography was not used extensively in these tests to confirm the nature of the indications. Later testing would provide more extensive correlation between AE and advanced UT methods (TOFD, Phased Array, Focused Array), and more sensitive metallographic analysis (cryo-cracking with SEM examination). These included programs on Kentucky Utilities Ghent #1 HRH line and Brown #3 SH link piping (1997-98), Central Power & Light's Joslin #1 HRH line (1997), Illinois Power's Baldwin #1 HRH (1998), Salt River Project's Navajo #2 HRH line (1998), Southwestern Public Service Co's Harrington #2 HRH line (1999), and Portland General Electric's Boardman HRH line (2001).
By the late 1990's it was becoming better understood that creep damage in seam welds did not initiate as distinctive crack-like flaws, but rather as an accumulation of microstructural damage evidenced by "cavitation" development around inclusions and carbides along the grain boundaries in the heat affected zone or fusion zone of the seam weld (depending on whether the structure was normalized and tempered or subcritically annealed after welding). Several high-profile failures of seam-welded piping after missed or misinterpreted UT findings (Sabine #2 HRH bend in 1992, Mt. Storm #1 SH link piping in 1996, Gaston #4 HRH bend in 2001) led to a greater sense of urgency in the fossil utility industry to find earlier stage damage more reliably in seam welds. The AE studies mentioned provided proof of early stage detection of creep damage at the cavitation stage, often well before UT methods could reliably indicate a developing problem. Only the advanced metallographic method involving the use of cryo-cracking and SEM examination at 2000-5000X was able to confirm this damage that AE was detecting at an early stage.
The standardization of the EPRI AE methodology for seam-welded piping began in earnest in January 2002 at the ASTM E07.04 acoustic emission subcommittee meeting. EPRI gave approval to the use of its documents and database as a necessary background for the development of the standard. The proposed standard WK 658 "Standard Test Method for Acoustic Emission Examination of Seam-Welded High Energy Piping" is nearing final balloting, and is expected to be approved in 2008. It is one of the most comprehensive and specific ever undertaken by an ASTM committee on an AE application.
Acoustic emission has proven its worth in online testing programs. Approximately 30% of lines tested have shown no significant findings of creep damage, and most others have shown only minor activity at suspect locations. The majority of seam-weld findings has been in elbows and bends, followed by hanger locations on horizontal line segments. These are known to be higher stressed areas, and offer further validation of the AE methodology. The correlation with follow-on nondestructive inspection has been very good, but the lesser sensitivity of UT inspection methods will generally not confirm early stage creep damage at the isolated cavitation stage. The economics of inspection and relative certainty of detection at an early stage of creep damage should be increasingly attractive to companies attempting to manage their piping systems in a climate of reduced capital and operations-maintenance spending.
[1.] B. Morgan, C. Foster, Acoustic Emission Monitoring of High-Energy Piping, Volume 1: Acoustic Emission Monitoring Guidelines for Hot Reheat Piping, Electric Power Research Institute, TR-105265-V1. 1995.
[2.] J. Rodgers, R. Tilley, Standardization of Acoustic Emission Testing of Fossil Power Plant Seam-Welded High Energy Piping, ASME Pressure Vessels & Piping Conference, San Diego, July 25-29, 2004, PVP-Vol. 471, pp. 113-131.
[3.] J. Foulds, R. Viswanathan, J. Landrum, "Guidelines for the Evaluation of Seam-Welded High Energy Piping, Electric Power Research Institute, TR-104631, 1996.
[4.] R. Munson, J. Rodgers, R. Tilley, The Utilization of Advanced Metallographic Techniques to Verify In-service Damage in Long-seam Welded, High Energy Piping, EPRI Fossil Plant Inspection Conference, Atlanta, 1999.
[5.] J. Foulds, R. Carnahan, Examination of Sabine 2 Hot Reheat Pipe Seam Weld Cracking, Electric Power Research Institute, TR-107141, 1997.
JOHN M. RODGERS
Acoustic Emission Consulting, Inc., Fair Oaks, CA 95628
|Printer friendly Cite/link Email Feedback|
|Author:||Rodgers, John M.|
|Publication:||Journal of Acoustic Emission|
|Date:||Jan 1, 2007|
|Previous Article:||Thirty years experience of industrial applications of acoustic emission testing at TUV Austria.|
|Next Article:||Acoustic emission and X-ray tomography imaging of shear fracture formation in concrete.|