Longitudinal seam-welded hot-reheat steam piping operating in the creep regime is a continuing life-management challenge for many older fossil-fired power plants. In response to catastrophic seam-welded piping failures in the 1980’s, the Electric Power Research Institute (EPRI) developed a comprehensive inspection protocol to insure continued safe operation of these piping systems . The protocol requires full inspection of seam-welded hot-reheat pipe once a threshold of service exposure (calculated creep life consumption) has been reached, and re-inspection at intervals after the initial inspection depending on the inspection results. Inspection for sub-surface cracking using ultrasonic testing (conventional or advanced) is strongly recommended, in combination with checking for surface cracking using wet fluorescent magnetic particle testing (WFMT). Initial inspection and re-inspection of these piping systems represents a large maintenance cost for utilities, especially as older plants remain in service due to the changing economics of power generation.
SI’s inspection protocol for hot-reheat seam-welded piping systems is comprised of a combination of visual, WFMT, and advanced ultrasonic techniques for detection of surface cracking, sub-surface cracking, and incipient creep damage. Ultrasonic inspection is currently performed using time-of-flight-diffraction (TOFD) testing on 100% of the seam weld length and annular phased array (APA) testing at selected locations along the seam weld length. In some cases, linear phased array (LPA) ultrasonic testing may be used in place of TOFD. Both TOFD and LPA have been shown to be able to detect creep damage at the microcracking stage, which is estimated to correspond to at most 85% of the creep-rupture life being consumed. APA has been shown to be able to detect creep damage at the aligned cavitation level, which is estimated to correspond to at most 70% of the creep-rupture life being consumed.
If no damage is found in an inspection, SI’s recommended re-inspection interval is defined based on the above-mentioned detection limits, so that a seam-welded pipe that has been inspected using APA with no indications of service damage will have a longer recommended reinspection interval than a seam-welded pipe that was only inspected using TOFD or LPA. However, in some cases, high densities of non-metallic inclusions in the pipe base metal act as ultrasonic reflectors, which mask the presence of creep cavities, so that APA cannot definitively determine that service damage is not present. For these cases the recommended reinspection interval will be relatively short, defined on the TOFD or LPA inspection findings alone.
Given the very high cost of inspections (driven mainly by scaffolding and insulation removal costs) and increasingly shorter outage time windows in which to perform them, SI has recently developed and applied a sampling, examination, and creep-rupture testing methodology to generate information and data that can be used to justify longer reinspection intervals for seam-welded pipes in which, due to the presence of high inclusion densities or other fabrication flaws, early-stage creep damage cannot be detected using APA ultrasonic testing. These inclusions and fabrication flaws may even limit the applicability of TOFD or LPA ultrasonic testing. This article presents a typical application example as a case study.
The power plant in this example is a supercritical coal-fired unit generating 1100 MW of power. Since starting operation in the early 1970’s, the unit has accumulated approximately 300,000 service hours with approximately 500 starts. Sections of the hot-reheat piping system are specified as 36” outer diameter by 1.984” minimum wall thickness ASTM A-155, Class 1 pipe, which is seam-welded 2¼ Cr – 1 Mo material (Grade 22). In a recent APA ultrasonic inspection, several pipe spools in this system were found to have high densities of fabrication-type indications along each weld fusion line and in the cusp region of the double-V groove type welds present. Typical TOFD and APA ultrasonic scans of these areas are shown in Figure 1. Because of these indications, the TOFD detection threshold was used in the re-inspection interval recommendation, which was 27,000 operating hours in this case. The reduced inspection interval was not acceptable to the client due to outage schedules and high cost of inspection.
Two cylindrical core samples were extracted from the seam weld of the pipe spool which showed the high fabrication flaw density. The three-inch diameter cores were removed and the resulting holes plugged in a single day. In this case a threaded plug with an outer seal weld was used, but a plugged sock-o-let type fitting is another option. Figure 2 shows photos from the core removal process and the plug repair.
The cores were sent to SI’s Materials Science Center in Austin, Texas for metallurgical characterization. One of the as-received cores is shown in Figure 3. The location and orientation of the seam weld was marked on outer pipe surface; the weld reinforcement was visible on the oxidized inner pipe surface. The cores were sectioned transverse to the seam weld and prepared for metallographic examination to confirm the presence of inclusions and/or creep damage in the form of cavitation. Initial sectioning was performed by electro-discharge machining (EDM) to preserve as much material as possible for creep-rupture testing.
Figure 4 shows a macroscopic view of one of the prepared cross-sections; the weld had a double-V geometry typical of hot-reheat seam welds and in agreement with profiles found in ultrasonic tests (Figure 1). Prominent weld heat-affected zones (HAZs) were present in the base metal adjacent to the weld, indicating that the welds had received a subcritical post-weld heat treatment (PWHT) rather than a full renormalize-and-temper PWHT. There was no macro- or micro-cracking observed on either core. Examination at higher magnifications showed that the weld metal, HAZ, and base metal microstructures were normal for the specified Grade 22 material, and that the carbide precipitates were not excessively coarsened or spheroidized. There was no significant creep cavitation visible in either sample, but there were numerous bands of inclusion stringers in the base metals, as shown in Figure 5. The presence of these bands is again consistent with the ultrasonic findings.
The Vickers hardness of the base metal and weld metal were measured and found to be normal for service-exposed Grade 22 material (approximately 170 HV10). The alloy type was checked using handheld x-ray fluorescence spectroscopy (PMI testing); the nominal compositions of the base and weld metals were both consistent with Grade 22 material (2.25Cr-1Mo).
FIGURE 6. Location of creep-rupture test blanks within a core sample.
Cross-weld creep-rupture tests were performed on specimens manufactured from the cores. Because the cores were very small it was impossible to machine meaningful specimens directly from them. Instead, rectangular pieces were machined from the cores onto which grip extensions were electron-beam welded. In this way the rectangular core sample pieces became the gauge lengths of the final test specimens. The overall specimen manufacturing process involves etching cross-sections to define the best locations for cross-weld specimens (Figure 6), machining test blanks centered on the cusp of the seam weld where damage typically develops, and then electron-beam welding grip-end pieces (also Grade 22 material) onto the test blanks (Figure 7). There was no PWHT performed after welding, so the microstructural condition of the weld was not altered in the specimen manufacturing process. Figure 8 shows the final machined specimens with the approximate location of the weld fusion lines marked.
Constant-load creep-rupture tests were performed in air at 1150°F (621°C) at several stress levels consistent with service loads. Some tests were run to failure, while other tests were terminated after a few months of test time. The test data are compared to the population of base metal and cross-weld literature data for Grade 22 material in Figure 9, which plots the Larson-Miller parameter (LMP) for the tests as a function of stress.
The Larson-Miller parameter combines temperature and time in the following form:
LMP = T ´ [20 + log(tr)]
where T is the temperature and tr is the rupture life or test duration. The results for the specimens tested at the three highest stresses are well within the scatterband for new cross-weld material, while the duration of the un-failed test at the lowest stress is approaching the scatterband, indicating that seam welds have very little creep damage, consistent with the microstructural observations of no cavitation, no microstructural coarsening, and no marked hardness decrease.
Overviews of a tested specimen interrupted prior to failure are shown in Figure 10. The specimen was oxidized, as expected; the two ruptured specimens had a ductile morphology with some associated necking. Longitudinal metallographic sections were prepared through these specimens to confirm the location of deformation. Final rupture was in the weld HAZ, consistent with the expected behavior for cross-weld loading of welds that have received a sub-critical PWHT.
In this case, because the examination findings and creep-rupture test data indicated that the weld had very little creep damage, it was possible to significantly extend the re-inspection interval for the pipe spool from which the cores were extracted. Although the core sample creep test results are only directly applicable to the pipe spool from which the cores were extracted, the other pipe spools that also showed high base metal inclusion densities were confirmed to have been from the same heat and manufacturing process (same size, weld process), therefore the inspection interval was also extended for those spools.
It is important to recognize that the creep test results cannot generally be directly applied to confidently predict the remaining life of the pipe. The issues with using cross-weld creep rupture data in predicting remaining lives of seam-welded components have been described in Refs.  and . A cross-weld test specimen extracted from a seam weld and loaded in uniaxial tension clearly does not fully represent the overall stress state for the full weld in the pressurized pipe, and there are additional complications created by using accelerated conditions (temperature and/or stress) in the creep-rupture test. However, as noted in Ref. , uniaxial cross-weld tests tend to underestimate remaining lives, so the results are expected to be conservative. Ref.  notes the additional complicating factors of weld metal to base metal strength mismatch; there also can be global stress-concentrating features associated with the seam weld such as pipe ovality and weld peaking angle. Hence cross-weld creep-rupture test results must be carefully considered in conjunction with other inspection findings, operating history and current operating conditions, and other possible pipe loads. These factors should all be included in a comprehensive seam-weld life assessment.
This example shows how examination and testing of core samples removed from “un-inspectable” (from an advanced ultrasonics perspective) seam-welded piping can greatly increase the confidence in the condition of the seam weld, and therefore provide an engineering-justified extended re-inspection interval, at a relatively modest cost relative to that of the overall inspection. Beyond seam welds, this core sample creep test methodology can also be used to characterize other discrepant material conditions, such as low hardness regions in Grade 91, or can be used to obtain creep data specific to a particular service-exposed material to obtain more accurate remaining life assessments.
 EPRI, Guidelines for the Evaluation of Seam-Welded High-Energy Piping, Fourth Edition, EPRI Report 1004329, 2003.
 Viswanathan, R. and J. Foulds, “Accelerated Stress Rupture Testing for Creep Life Prediction—Its Value and Limitations.” Journal of Pressure Vessel Technology, 1998. 120(5): p. 105-115.
 Segle, P., et al., “Some Issues in Life Assessment of Longitudinal Seam Welds Based on Creep Tests with Cross-Weld Specimens.” International Journal of Pressure Vessels and Piping, 1996. 66: p. 199-222.