Guided Wave Testing - Maximizing Buried Pipe Corrosion Knowledge from each Excavation

Transcription

1 Proceedings of the ASME 2012 Pressure Vessels & Piping Conference PVP2012 July 15-19, 2012, Toronto, Ontario, CANADA PVP Guided Wave Testing - Maximizing Buried Pipe Corrosion Knowledge from each Excavation Andy Crompton E. Easter Ave., Ste. 100 Centennial, CO Structural Integrity Associates, Inc. acrompton@structint.com Roger Royer 301 Science Park Road, Ste. 121 State College, PA Structural Integrity Associates, Inc. rroyer@structint.com Mark Tallon Progress Energy Brunswick Nuclear Plant 8470 River Road SE Southport, North Carolina Stephen F. Biagiotti, Jr. P.E E. Easter Ave., Ste. 100 Centennial, CO Structural Integrity Associates, Inc. sbiagiotti@structint.com ABSTRACT Excavation and Direct Examination of buried piping using conventional non-destructive examination (NDE) has been the traditional inspection approach for decades and remains the only quantitative method for piping evaluations in plants when internal in-line inspection tools cannot be used due to geometry or other constraints. This difficult to assess piping presents many challenges, including limited effectiveness of traditional indirect inspection tools, high cost and operational concerns associated with excavations, and the ability to evaluate only a small sampling of a piping system. Many inspection technologies exist for buried pipe assessments; however, no indirect techniques provide the ability to yield quantitative wall loss values suitable for ASME fitness for service calculations beyond what s exposed in the excavation. An evolving technology, guided wave testing (GWT), has many applications including the ability to provide assessment information beyond the excavation. In this paper, the application of GWT for buried piping inspection will be discussed. We will review: principles behind its operation; the competitive technologies on the market; challenges for the technology; management of data within the Electric Power Research Institute (EPRI) industry standard buried pipe database (BPWorks 2.0); trending; case histories showing how GWT can be used to extend the knowledge gained during an excavation by screening adjacent areas for more significant corrosion than observed in the excavated and exposed area; coupling GWT results with other inspection technologies to gain an enhanced interpretation of the overall condition of the line; and how to incorporate this data into an effective structural and/or leakage integrity program as part of the reasonable assurance process. PRINCIPLES OF GWT Guided waves are acoustic, mechanical, or electromagnetic waves with propagation characteristics dictated by the properties and boundaries of the medium in which they travel. Ultrasonic 1 Copyright 2012 by ASME

2 guided waves, used for screening to detect changes in crosssection, are elastic mechanical stress waves introduced into the pipe wall and constrained to propagate along the axis of the pipe segment within the pipe wall. The waves are of relatively low frequency (15-85 khz) and are scattered and/or reflected by changes in the pipe cross section or acoustic impedance. Reflections can originate from corrosion, welds, supports, coating interfaces or transitions, earth entrances, etc. One main difference between using guided waves for screening and traditional ultrasonic testing or bulk waves is their ability to size and locate defects. Further comparison of these techniques is provided in Table 1. Table 1. Differences between Bulk and Guided Waves Bulk Wave Time consuming for large areas Point by point scan directly beneath sensor Partial coverage Defect must be directly accessible to be detected Accurate defect sizing Guided Wave Rapid screening Single sensor position propagates waves on either side of the sensor Volumetric coverage Detected defects may not be directly accessible Approximate circumferential and axial sizing For piping applications, the types of guided waves used for excitation are either longitudinal or torsional. Longitudinal waves are characterized by the compression motion they create along the axis of the pipe. A drawback to longitudinal waves is they are not ideal when considering the inspection of liquid filled pipes as they can propagate into that medium causing distortion of the data. A torsional wave is most commonly used for pipe inspection and results in the twisting of the pipe preventing the energy from transferring into the pipe media. Upon impinging on a reflector, each of these excitation methods result in the reflection of symmetric and mode converted flexural waves. Flexural waves are generated from reflectors which are not symmetric about the pipe s circumference such as an isolated corrosion patch. Therefore, the symmetric and flexural responses can be used to characterize a given reflector. As an example, a uniform weld will produce a large symmetric reflection with little flexural response. In contrast, isolated corrosion or pitting will produce a large flexural reflection with a lower symmetric component. Torsional modes may be more or less sensitive to certain types of defects than longitudinal modes due to the different in/out of plane displacement components each mode is comprised. Penetration power of both types of waves can vary significantly with frequency and boundary conditions. Therefore, depending on the application and requirements, one mode may be better suited than the other Guided wave transducer arrays are typically comprised of a comb configuration which can be modified either during construction or in the field to alter the frequency band used for excitation and analysis. One main factor that influences the frequency is the transducer spacing whereas a smaller spacing results in higher analysis frequencies and greater resolution (due to the shorter wave lengths) which can result in shorter screening distances. The frequencies available are also governed by the boundary conditions, or more specifically the pipe size and thickness, which influence the guided wave dispersion curves. The guided wave dispersion curves for a given structure represent all wave modes available for excitation [1]. Each reflector along a pipe segment is characterized by its amplitude, flexural, and frequency responses. A symmetric amplitude response is present for all reflectors and is proportional to positive or negative cross-sectional change (CSC) when compared to the total surface area exposed by a transverse cut of the pipe. A non-symmetric amplitude response is created if the feature is non-symmetric and is used to estimate the feature s length around the circumference of the pipe, also known as circumferential extent (CE). For example, a weld has 100% CE; however, a defect with a given cross-sectional wall loss is more critical for a smaller CE as shown in Figure 1. Figure 1. Wall loss as a function of CSC and CE The attenuation or decay of the signal is logarithmic over the inspection distance down the pipe in either direction from the sensor location. The attenuation can be caused by pipe appurtenances, wrappings, linings, contents, soil loading, and/or general corrosion. Reflections from a series of defects and the associated energy loss are shown in Figure 2. Screening is generally conducted by comparing the known CSC of girth welds to unknown features such as corrosion. Welds are generally assumed as 22.5% CSC (positive) and vary from weld to weld; however, weld measurements can be taken to improve the accuracy of the screening. A flange or threaded coupling represents 100% CSC and therefore no or little energy is transmitted beyond this feature, marking the end of the diagnostic range for all examinations. A unique feature such as a 45 degree bend in most cases distorts the wave such that data 2 Copyright 2012 by ASME

3 may become unreliable for interpretation beyond the feature and generally marks the end of the test. Guided Wave 100% Energy Estimated 20% CSC for a Defect as Figure 3. This figure displays two responses of similar amplitude prior to applying the DAC curves. The first peak occurs in an area where the decay is relatively small providing an excellent SNR and the ability to distinguish small CSC; however, further down the pipe this same characterization is not possible and the sensitivity towards small CSC has decreased. 20% 80% 16% 64% 13% 51% Centerline of Pipe 1% CSC at 28 feet 7% CSC at 136 feet Figure 2. Reflections from Guided Waves Due to logarithmic decay, a distance amplitude correction (DAC) curve is applied to the data as a method to compensate for loss of amplitude over distance and assign a constant amplitude (%CSC) over a distance. Some of the curves used during analysis are: the weld DAC, used as the first means of proper amplitude correction and reference; call DAC, used for a visual threshold for feature detection; and a noise DAC, which can be used to determine the signal-to-noise ratio (SNR). Some common types of DAC curves placed over guided wave data are shown in Figure 3. Weld DAC (22% CSC) Call DAC (6%) Figure 3. DAC Curves Noise DAC The sensitivity achieved for a test represents the percent CSC that can be detected at a given SNR. One element contributing towards the confidence of identifying and classifying features is the SNR. It should be noted that CSC is one measure for reflections and many other factors influence the reflection of guided waves at a defect. For this reason, during the analysis the technician may also characterize indications in the testing range based on the reflector s flexural, frequency, and focused response. The inspection range is typically a function of the required sensitivity. For example, if the sensitivity is increased for GWT from 5% to 10% at a constant SNR the distance for the inspection will also increase. However, smaller features less than the specified 10% CSC sensitivity may not be discernible from the noise in the signal. Further illustration of sensitivity over distance is shown in Figure 4 using the same DAC Curves Figure 4. Sensitivity over Distance Typically, guided wave inspections are conducted by first performing an axisymmetric inspection at multiple frequencies over the full bandwidth of the transducer arrangement. The use of multiple frequencies is beneficial because reflectors are often frequency dependent in terms of their response such that some frequencies may be more or less sensitive to a given reflector or defect than others. Upon completing the axisymmetric inspection and analysis, focusing techniques are often employed to further characterize the circumferential location and extent of a defect or reflector. Two of the most common techniques available for doing so are (1) synthetic focusing, or passive focusing and (2) phased array focusing, or active focusing. Both of these techniques require the guided wave transducer array to be segmented circumferentially. Descriptions of these techniques are as follows: 1. Synthetic Focusing Axisymmetric or partial loading excitation and reception where the received signals are postprocessed to produce an image of the reflector. This approach enhances circumferential sizing capability for a reflector. 2. Phased Array Focusing An active focusing technique where time delay/amplitude phasing is applied to the transmitting transducer segments in order to cause constructive interference of the waves produced from each segment to occur at some particular axial and circumferential position along the pipe. This approach also enhances circumferential sizing capability and can provide improvements in penetration power and detection sensitivity as more energy is being delivered to the focusing positions due to the constructive interference of all the transducers. 3 Copyright 2012 by ASME

4 Benefits of using focusing techniques are as follows: Improved defect probability of detection Decreased defect false alarm rate Increased inspection confidence Improved penetration power potential (phased array) Defect circumferential length and depth measurement potential COMPETITIVE GUIDED WAVE TECHNOLOGIES ON THE MARKET There are multiple manufacturers of competitive technologies on the market including Guided Ultrasonics Limited TM (GUL), Plant Integrity Limited TM (Pi), FBS Incorporated TM (FBS), and Southwest Research Institute TM (SwRI). Depending on the manufacturer, either piezoelectric or magnetostrictive sensor technologies are utilized for generating and receiving the guided waves. This discussion will focus on the piezoelectric approach where an array of piezoelectric transducers are coupled to perform these functions. These transducers are aligned in a module for either longitudinal waves (aligned with the pipe s central axis) or torsional (aligned with pipe circumference) and held in bracelets often referred to as collars. The collar is typically clamped around the pipe and coupled directly to the pipe utilizing an inflatable bladder to hold the transducers in contact with the pipe. A different collar or combination of collars is needed for each pipe diameter and the number of modules in each collar increases with pipe diameter. Circumferential clearance varies from approximately four inches to a little more than an inch. The axial length of the collar is generally less than 10 inches. An example of a collar and its components is shown in Figure 5. The generation of the waves, receiving, filtering, and post-processing is conducted with a unit that is usually connected to a laptop computer and analyzed using the manufacture s software. The systems are typically battery powered. Module Figure 5. Guided Wave Collar Transducer Long term guided wave monitoring systems are available for permanent installation and can be environmentally sealed. These systems are becoming common-place for use in excavations and installed prior to backfilling. These systems allow for an above ground or remote connection to monitor piping conditions over time. Permanently installed collars are available using both piezoelectric and magnetostrictive sensing approaches. The main difference between these transducers is the magnetostrictive sensor is composed of a coil and magnetostrictive strip that is bonded to the pipe surface to generate the desired wave modes and frequencies. APPLICATIONS OF GWT AND CHALLENGES WHEN APPLYING THE TECHNOLOGY Guided wave technology was originally developed for testing for corrosion under insulation but has since been applied in many other applications. Successful applications include: vertical risers into buried sections, wall penetrations into buried sections, adjacent testing of excavations, corrosion under supports typically above ground, road crossings, river crossings, platform riser inspections of splash zone, and weld locating for insulated above ground piping. Each of these scenarios present unique challenges and limitations. For this paper, the challenges and limitations for GWT of buried sections will be discussed along with some suggested techniques to enhance the overall effectiveness of applying GWT. In order to inspect the pipe segment of interest using GWT, access needs to permit the placement of the collar. This usually includes complete access to the entire circumference of the pipe substrate with a circumferential clearance of at least 3 inches and a preferred 12 inches. The pipe may require coating removal if the coating is a tape wrap or coal tar enamel applied coating. Fusion-bonded epoxy (FBE) generally does not require removal. However, enhanced coupling may be achieved thus increasing the SNR if the coating, including FBE is removed. The collar itself requires approximately 10 inches of coating removal axially. The technology however has some limitations in the area adjacent to the collar. For example, the region under the collar and extending typically 1.5 feet on both sides of the collar is referred to as the dead zone. No reflected responses are detected in this region because the transducers are incapable of transmitting and receiving at the same time. Thus, tradition UT is used in this region to evaluate this zone. Another region of limited interpretation is the near field which occurs where the guided waves are forming. Features are detectible in this region but the amplitude may not be accurate. Directional control is not perfect in this region and mirrors (false reflections) from real reflections can arise. This region can extend 1.5 feet to a few feet in addition to the dead zone. These regions can be directly examined using traditional ultrasonics and visual inspection, by overlapping tests such that these zones are inspected by another adjacent guided wave test (including pitch/catch methods), or utilizing magnetostrictive transducers which typically have a combined near field/dead zone of less than two feet. The easiest way to eliminate the effects of the near field/dead zone in the region of interest is to place the collar such that the near/dead zone is outside the region of interest; however, many times the configuration of the piping does not permit an alternate location. One means to calibrate the test is through the detection of known features with an approximate size such as welds (at bends or 4 Copyright 2012 by ASME

5 ends of pipe spool pieces) and to adjust the DAC curves for relative sizing and categorization of unknown features (i.e. defects). If there is no weld present or it cannot be confidently distinguished from the noise in the data, the overall confidence and sizing of features is adversely affected. The presence of well-adhered internal linings, external coatings, compacted soil, and/or general corrosion, can increase signal attenuation thus limiting inspection distance. The attenuation caused by general corrosion and a bitumastic coating, as an example, may look similar, and in some cases cannot be distinguished between each other. Bends can distort the waves beyond the bend leaving areas of lower confidence for screening due to the blind spots created from wave distortion. Testing experience has indicated that for a buried bitumastic coated pipeline or testing through multiple bends is unlikely to produce data for which features below 10% CSC can be detected with high confidence. Attenuation or test distance is also affected by the thickness and temperature of the coating and somastic coatings thicker than 0.5 inches may render the guided wave data inconclusive. Coatings that have been left in the sun for an appreciable amount of time may also severely attenuate the signal immediately beyond the coating. Coatings should be removed to the extent possible to determine the condition of the piping and it is recommended up to 10 feet be removed to aid in the effectiveness of GWT. Other large reflectors may be encountered during testing such as equal diameter branches or tees (when compared to the tested pipe), elbows and wall penetrations. GWT inside a tunnel, screening buried regions outside the tunnel are generally affected by the type of wall penetration and elbows or bends encountered from inside the tunnel or in the buried section beyond the wall penetration. These configurations generally limit the inspection distance to the first elbow in the buried section (for bitumastic coated pipe). Adjacent large reflectors directly on either side of the collar, such as a bend on one side and a wall penetration with a welded anchor on the other can disrupt the guided wave formation thus elevating the noise in the data. This also causes false echoes due to the large reflectors present in the near field, further limiting the overall effectiveness of the screening for the buried section. It should also be noted that the testing configuration at a penetration may not allow an alternate location to eliminate possible sources of false echoes and prevent the near field/dead zone from extending into the buried section of interest. For example, Figure 6 shows a collar location that may extend the near zone into the buried region of interest; however, the alternate location places two bends between the collar and region of interest likely providing no data for the region of interest. The first location, although not optimal, presents the only location to collect GWT data for the region of interest. This testing configuration between large reflectors is similar to that of risers or vertical sections where above ground access is provided between a bend and the earth entrance and similar inspection distances can be expected due to the placement of the collar between large reflectors. Ideally, penetrations or risers should be tested greater than 6 feet from the wall penetration/earth entrance on straight sections. The likelihood of achieving a sensitivity of 10% or less is generally low beyond the first buried bend for bitumastic coated pipes. Cased segments are also a good application for GWT due to their straight geometry and the lack of soil loading along the pipe lessening the overall attenuation. Alternate Collar Location Area of Interest Collar Location Figure 6. Wall Penetration Testing Locations The technology cannot distinguish between a positive and negative wall loss; therefore, testing of a single support inside a wall penetration may indicate a response and associated trailing echoes from a positive CSC; however, the support to pipe interface may include corrosion product and/or wall loss which would be additive to the support response. Without previous knowledge of typical support responses or direct comparison of multiple supports, it is difficult to distinguish between corrosion and the support. Testing through multiple supports at high SNR allows comparison of supports, both in amplitude and frequency response, to locate anomalous supports for further examination. Focusing techniques can also be employed in these situations to assist in detecting corrosion. For buried segments, the support example is analogous to coating transitions at joints, such as field applied tape wrap over girth welds or variation in coating thickness along segments causing reflections that cannot be distinguished from corrosion. Knowledge of the type of coating and field applied techniques can help provide more information for GWT responses and aid in the decision making process for additional follow-up. Permanently installed collars provide the opportunity to trend beyond bends (which normally mark the end of buried tests) and inside the near zone by accounting for the amplitude distortion, which is theoretically constant over time. The capability of trending from baseline data improves sensitivity for defects at structural features and remote defect-like features. Although structural features are additive in the presence of defects, the growth of the defect at or remote from another feature can be detected by the change from baseline data, achieving sensitivities of less than 1% CSC where typically buried tests are limited to 5 and 10% CSC sensitivity. For these reasons, GWT is a screening technology that can identify areas of interest, worthy of subsequent follow-up and 5 Copyright 2012 by ASME

6 provide a qualitative ranking of indication severity. Limitations and challenges should be reviewed during the pre-assessment inspection phase to determine the feasibility for GWT and other indirect inspection tools. The remaining wall thickness cannot be measured directly using this technique and quantitative techniques (i.e. pit depth measurements or ultrasonic thickness testing) should be supplemented as appropriate. USING AS PART OF A UPTI PROGRAM In April 2011, the Buried Pipe and Tank Integrity Task Force released the document, Industry Guidance for the Development of Inspection Plan for Buried Piping [2]. This document provides guidance to US nuclear plants with the implementation of the NEI Underground Pipe and Tank Integrity (UPTI) initiative. Within this document, the concept of reasonable assurance of integrity is introduced for buried pipe systems; optimizing inspection scope, while not requiring 100% inspection. The process is designed to increase the confidence in a structure, system, or component (SSC) to perform its intended function. excavation was used to repair the location. Direct Current Voltage Gradient (DCVG) surveys over this and similar lines were used to select further areas for excavation and determine the extent of degradation on all lines of similar construction at the site. In order to maximize the amount of data collected for the excavation, GWT was chosen to screen the piping adjacent to the excavations and identify any areas that may require further remediation. GWT was conducted over the coating and again after it was removed to assess the specific benefits of coating removal for this screening. A significant increase in the SNR was observed after the coating was removed and therefore all subsequent testing locations had the coating removed to improve the sensitivity and screening coverage. The confidence level for the testing was above average with buried screening at 5% sensitivity achieved upwards of 90 feet in one direction. In one excavation, a severe GWT indication was detected 3 feet beyond the excavation resulting from a gouge (mechanical damage) with an average remaining wall thickness of inches and a through wall pin hole as shown in Figure 7. The process relies on a series of activities that refine the inspection scope to discrete locations with the greatest potential for degradation. Indirect inspection techniques, methods that gauge the performance of a corrosion control technique with direct contact with the buried structure, are key indicators used in this process. These indicators ultimately yield prioritized excavation recommendations. However, they just narrow the highest likelihood degradation site to a region ±20 ft. The cost of excavations is high at most plants. To increase the value of any excavation, GWT can be paired with conventional UT to validate that an indication more severe than those measured on the exposed pipe do not exist just beyond the walls of the excavation (statistical Type II error). The indirect inspection methods used to prioritize excavation locations are made at the surface of the soil and represent a weighted average cone of influence. The deeper the source of the degradation, the wider the influence cone. Thus, as long as GWT indications are not observed that are more severe than those detected in the exposed pipe, the utility gains additional confidence that they did evaluate the worst case location on the line. Additionally, GWT is considered an indirect inspection technique and the additional length of pipe interrogated can be used in the determination of the total pipe length inspected for Reasonable Assurance purposes. This has a significant impact on the UPTI inspection process because decision points exist in the process that dictate the number of excavations required as a function of percentage of pipe length inspected via an indirect inspection technique. If more than 50% of a system is evaluated, the potential for only one excavation may be required. CASE STUDIES A combined cycle generating plant began operation in A leak caused by external corrosion was discovered in 2011 on a 10-inch FBE coated buried natural gas delivery line. An Figure 7. Case Study #1 Brunswick Nuclear Plant installed the off gas 8 carbon steel pipe in The pipe was coated with coal tar enamel approximately 3/16 of an inch thick. During the pre-assessment phase of the excavation, GWT was chosen to screen the areas adjacent from the exposed excavated area and identify any areas for potential degradation. The primary threat identified for this pipe segment was external corrosion. The screening distance for the piping was 19 and 22 feet for 10% sensitivity and 14 and 14 feet for 5% sensitivity in the negative and positive directions, respectively. For this case, no indications in the buried section were detected; however, a weld was located outside the near zone/dead zone enhancing the overall confidence in the test and severity categorization if needed. The near zone also extended partially into the coated sections where amplitudes may not be reliable. In this case, the coating was visually inspected for any disbondment or holidays reducing the likelihood of any external degradation in the area. During the visual inspection, a small coating holiday approximately 1.5x1.5 inches was identified. GWT results for this location are provided in Figure 8. 6 Copyright 2012 by ASME

7 4 Foot Follow Up Area GWT Estimation UT Prove-up 0 0 Figure 8. Case Study # A GWT exam was conducted to assess the buried section of a 24-inch service water line in a nuclear power plant from inside a building. The transducer collar was placed on a straight section of pipe ~7 ft. from the inside wall penetration to look out into the buried section. A T-piece was present in the buried section of the pipe at a distance of ~19 ft. from the inside wall penetration (26 ft. from the collar). Two welds were detected in the buried section in addition to the T-piece which marked the end of the test. The presence of multiple welds in the data allowed for proper setting of the DAC curves thus improving confidence in the data. Multiple corrosion-like indications were present within the 4 ft. section of pipe directly before the T- piece (15-19 ft. from the inside wall penetration). Based on the DAC settings, these responses represented CSC s ranging from ~10 18%. Phased array focusing was used to estimate the circumferential profile of the most severe response. The 4 ft. section of piping was excavated to confirm the results of the GWT findings. Conventional ultrasonic thickness (UT) measurements showed that the pipe was experiencing internal corrosion throughout the 4 ft. section of pipe identified with GWT. The axial and circumferential positions of the most severe corrosion agreed with the guided wave data. The total cross sectional loss at this location was calculated to be 16.5%. GWT focusing estimated that the corrosion was concentrated between the 9:00 and 10:30 positions which was confirmed during the UT prove-up. Indications called out in the GWT axisymmetric inspection as well as the GWT estimation and UT confirmation of the circumferential position of corrosion damage are shown in Figure Figure 9. Case Study #3 MANAGING GWT DATA IN UPTI PROGRAMS 90 A database solution is available through the Electric Power Research Institute (EPRI) for nuclear power plants to store all assessment data including indirect survey and guided wave data. Companion tools allow all the data to be visually represented using a geographic information system (GIS) data model MAPProView. These tools allow for record keeping and retrieval of risk ranking results per segment/group and inspection results to demonstrate reasonable assurance for the integrity of the piping at the site. The EPRI BPWorks TM data management tool is shown in Figure 10, along with GWT inspection results Brunswick Nuclear Plant shown in Figure Figure 10. BPWorks Data Management Tool 7 Copyright 2012 by ASME

8 Figure 11. GWT Inspection Data for Brunswick Nuclear Plant CONCLUSIONS This paper reviewed the application of GWT for buried piping inspection to increase the confidence of structural and leakage integrity as part of a NEI UPTI program [3]. GWT can be used to extend the knowledge gained during an excavation by screening adjacent areas for more significant corrosion than observed in the excavated and exposed area. Coupling GWT results with other inspection technologies further enhances the understanding of the overall condition of the line. Incorporating this data into excavation inspection plans will increase the effectiveness of structural and/or leakage integrity programs. REFERENCES 1. J. L. Rose. Ultrasonic Waves in Solid Media. Cambridge University Press, New York, Industry Guidance for the Development of Inspection Plans for Buried Piping, NEI Buried Pipe Integrity Task Force, Final Draft Approved for Use, April Guideline for the Management of Underground Piping and Tank Integrity, NEI 09-14, Revision 1, NEI Buried Pipe Integrity Task Force, December Copyright 2012 by ASME