Back to Basics

[ click here for the Back to Basics Archive ]

An Assessment of Remote Visual Testing
System Capabilities for the Detection of
Service Induced Cracking

by M.T. Anderson,* S.E. Cumblidge⁺  and S.R. Doctor+

 

While reviewing this article, I adjusted my glasses and thought of the days when I had the potential to resolve very fine features without optical aids. Remote visual and optical testing has advanced in recent years, allowing even those of us needing assistance with our vision to locate and document extremely small cracks. The authors of this article present a well written discussion addressing variables, advantages and limitations of current remote visual and optical testing systems. Bruce Crouse Contributing Editor

Figure 1-3
Figure 4-6
Table 1

 

INTRODUCTION
Visual and optical testing (VT) is the most fundamental, and arguably, the simplest of all nondestructive testing (NDT) methods. It is widely used as a primary test method or may serve only to provide complementary information for other, more indirect NDT methods. The human eye is highly adept at detecting small features or irregularities on the surfaces of materials and direct VT, if applied under specific parameters with appropriate optical tools and lighting, can exhibit highly reliable test results. However, many VT applications use remotely operated video camera systems due to factors such as the location, size and geometry of the parts being tested, or the adverse environment of the required test surfaces. While remote visual testing systems are convenient to deploy for periodic tests, one should understand the variables introduced by these systems in order to determine their overall capabilities and limitations in detecting the targeted degradation.

---

The primary feature of a crack to be visually detected is its width, or crack opening dimension.

---

In the commercial nuclear power industry, remote visual testing is used to examine components in the primary coolant system, including internal surfaces of the reactor pressure vessel and core support structures. This testing is conducted as part of an established inservice testing program, required by Title 10, Part 50 of the Code of Federal Regulations (US Nuclear Regulatory Commission, 1956). Implementation variables, such as the frequency of testing, visual resolution, lighting parameters, acceptance criteria and inspector qualifications are discussed in the American Society of Mechanical Engineers (ASME) Boiler and Pressure Vessel Code, Sections V and XI (ASME, 2001). The US Nuclear Regulatory Commission (NRC) reviews and approves applicable ASME code editions for use at operating nuclear plants and ensures that inservice testing programs comply with federal regulations. Current inservice testing programs at operating reactors rely heavily on NDT to detect the presence of service induced degradation that may lead to a loss of pressure boundary or structural integrity; remote visual testing is one in a suite of NDT techniques that are deployed for this purpose. Since 1977, the NRC Office of Nuclear Regulatory Research has funded a multiple year program at the Pacific Northwest National Laboratory (PNNL) to evaluate the reliability and accuracy of NDT techniques employed for inservice testing at commercial nuclear power plants. Through this program, research is performed to establish the technical bases for improved tests in reactor systems and components, to evaluate the effect of inservice testing reliability on reactor system integrity and to make recommendations to the ASME code to improve the effectiveness and adequacy of inservice testing techniques and programs.

The materials used in the design of safety related components at commercial operating reactors are fairly robust; they were selected to provide good corrosion resistance, high strength and fracture toughness. The most advanced forming and welding processes for their time were used under the auspices of rigorous quality control programs during the fabrication of important components. Operating and residual tensile stresses are usually highest on the surfaces of the components near structural discontinuities such as welds or thickness transitions. For these reasons, it is commonly believed that service degradation will be initiated on the surfaces (not from embedded fabrication discontinuity growth) and will most likely occur first on the internal surfaces exposed to environments that could accelerate discontinuity initiation and growth. Therefore, in many cases, known degradation processes are expected to result in crack initiation on the internal surface of safety related system components. For remote visual testing to be considered effective, test systems and implementation practices must possess a capability for detecting small cracks before they grow to a size that could challenge the leak tightness of the pressure boundary.

In recent years, the US nuclear industry has tried to minimize the burden associated with inservice testing by reducing current volumetric testing (for example, ultrasonic or radiographic testing) and/or surface testing (including electromagnetic, liquid penetrant and magnetic particle testing) of certain components in commercial nuclear power plants. Substitution of remote visual testing is one method that could potentially reduce this burden. The advantages of using remote visual testing are that these tests generally involve much less radiation exposure and testing time than do the current volumetric and surface techniques. In some cases, the industry has proposed to perform "enhanced" remote visual tests as alternatives to existing volumetric or surface tests (Dorman, 2000). This enhancement is based on the ability of the system to resolve a 12 µm (5 x 10^-4 in.) diameter wire, intended as a baseline system calibration. In order to reduce the inservice test burden by supplanting established volumetric or surface methods with remote visual testing, an analysis of all the pertinent issues is needed to support the reliability of remote visual testing in determining the structural integrity of reactor components.

This paper is intended to provide a basis for describing the technical issues that must be addressed when applying remote visual testing to detect cracking phenomena by highlighting the inherent capabilities and limitations associated with current system deployment. The work has been aimed at commercial nuclear reactor components; however, these issues exist for all industries where remote visual testing is expected to reliably detect service induced cracks in similar materials and structures.

 

VISUAL TESTING PARAMETERS

Context
In its most basic terms, summarized as "look at the test subject and try to detect a discontinuity," VT appears to be rather simple. However, when the size of a discontinuity is close to the resolution limit of a test system (where this system may include such things as a human eye or a video camera), the tests can become deceptively complex. Factors influencing the reliability of VT include, but are not limited to, light levels, lighting angles, surface conditions, resolution limits of the equipment used, magnification, contrast between the discontinuity and the surface, the amount of time spent testing the sample and a host of human factors (ASNT, 1993; Moore et al., 2001).

 

Acuity Factors
A human eye with 20/20 vision is able to resolve features as small as 75 µm (3 x 10^-3 in.) in size at a 250 mm (10 in.) distance (ASNT, 1993). This limit is based on the density of rods in the retina of the eye and on the diffraction limit imposed by the size of the eye. The eye is also, however, able to detect features too small to be accurately resolved. It is possible under perfect conditions to detect a crack with a surface width or crack opening dimension as small as 10 µm (4 x 10^-4 in.) on a mirror polished surface (ASNT, 1993). The minimum detectable crack opening dimension becomes much larger if the surface is rough or not perfectly clean. These limits do not account for factors such as scratches, machining marks or any camouflaging effects caused by macroscopic features such as weld roots or crowns.

Any system used in VT (ranging from the naked eye to a digital closed circuit television system) will have a measurable visual acuity. The visual acuity of a system has four pseudoindependent measures (De Petris and Macro, 2000):

• visible minimum - the smallest dot the system can detect

• separable minimum (resolution) - the smallest separation between two lines the system can detect

• visual acuity by Vernier - the ability to perceive spatial variation between two objects

• readable minimum (recognition capability) - the ability to recognize complex shapes such as letters or numbers.

These visual acuity parameters describe what a system can detect and discern. A system with a detection limit of 10 µm (4 x 10^-4 in.) and a resolution limit of 30 µm (1.2 x 10^-3 in.) at a given distance, can "see" a 10 µm (4 x 10^-4 in.) wide line on a sheet of paper but cannot resolve a 10 µm (4 x 10-⁴ in.) gap between two 10 µm (4 x 10^-4 in.) lines. A letter or number will appear to be a dot if it is larger than the visible minimum of a system, but below the readable minimum of the system. The letter will be identifiable only when it is above the recognition capability of the system.

 

Modulation Transfer Function
The image sharpness produced by mechanical visual systems, such as still and video cameras, can be described in terms of their modulation transfer function. The modulation transfer function is a measure of the detected versus the actual contrast ratio as a function of the spatial frequency of the indications. For example, a camera will generally show nearly 100% contrast with two black lines on a white background when the lines are far apart (low spatial frequency); however, if the lines are very close together (high spatial frequency), the system can blur the lines and the spaces between the lines together, reducing the contrast and thus reducing the modulation transfer function. Measuring the modulation transfer function of a system as a function of spatial frequency is a very reproducible and objective way to measure the visual sharpness of a camera system.

 

Resolution Tests
A resolution test is another common technique used to characterize the visual acuity of a system. A resolution test determines the smallest distance between two lines that can be discerned by the system. A resolution target generally has several sets of parallel or converging lines with notations on how many lines per millimeter are present at each point. Performing a resolution test consists of making an image of a standard resolution target and determining the point at which the system can no longer separate the lines. The main problem with resolution tests is that they rely on the observer to determine which lines are separable and which are not, adding an element of subjectivity to this measurement. However, a resolution test has the advantage of being faster and easier to administer than a test of system modulation transfer function. Examples of resolution targets include the IEEE Resolution Target, which conforms to STD 208-1995, Measurement of Resolution of Camera Systems (IEEE, 1995), the 1951 US Air Force Resolving Power Target and the ISO Camera Resolution Chart, the details of which targets conform to a standard found in Sine Patterns (2004).

 

Video System Acuity
The maximum visual acuity of an analog or digital video system can be described by comparing the native resolution of the system to the size of the area on which the system is focused. This measure of acuity assumes perfect optics and a perfect electronic capture of the image. Using this technique, a 1200 by 800 pixel camera that can focus on an area 75 by 50 mm (3 by 2 in.) would have a pixel size of 62.5 mm/pixel (2.5 x 10^-3 in./pixel). Any indications that fall below this size would be pixilated and recorded as a lower contrast shadow in the larger pixel, as the contrast from the indication is averaged with the background in the pixel. The color and shading of the pixel is dependent on the contrast between the indication and the background and on the modulation transfer function of the camera. An example of a linear indication in which the width of the line is significantly less than the pixel size is shown in Figure 1. With analog and digital video systems, one needs, as a theoretical minimum, at least one pixel (or line) width between two lines to resolve them, assuming the lines are perfectly aligned with the camera (Jensen, 1968). In practice, this corresponds to at least 1.4 pixels between two lines to always be able to resolve them, regardless of the angle and orientation of the lines.

Kinetic Visual Acuity
Another important variable in visual acuity is the speed at which the imaging detector is moving over the tested area. The term kinetic vision acuity is used for the acuity of a given system when scanning a moving target. The loss of visual acuity as a function of scan speed is highly dependent on the technology used to capture the images. A high speed film camera can produce sharp images of a bullet in flight, while a poor video system can show noticeable blur at slow scan speeds. In general, high quality image capture with video systems requires that the camera be halted over the area to be tested for a few seconds. An example of the loss of visual acuity when a camera is scanning is shown in Figure 2. The picture is significantly sharper when the camera is almost stationary (system normal visual acuity) over a zone as opposed to scanning (system kinetic visual acuity).

Indication Identification
Once an image is captured by a camera system, two important parameters are used to quantify how well an indication can be identified from the image; these parameters are the size of the indication on the image and the signal to noise ratio. The size of the indication is best measured in this context in terms of the number of resolution lines. A historical rule of thumb from aerial reconnaissance photography is that three resolution lines on a side are needed for detection; five resolution lines along a side are needed for identification (Jensen, 1968). However, these rules were developed for round objects and are difficult to apply to crack detection, as cracks are very long (many resolution lines with a typical system) relative to their width (possibly less than one resolution line with the best of systems). The signal to noise ratio is determined by comparing the signal from the contrast between the crack opening and the base metal to the noise from machining marks, cleaning marks and other extraneous indications on or near the discontinuity.

 

REMOTE VISUAL TESTING SYSTEMS AND IMPLEMENTATION GUIDELINES

System Basics
As components in a nuclear power station are generally maintained underwater and reside in high radiation fields, remote testing with radiation hardened video systems is necessary. Remote visual testing has been successfully used to find cracks in pressure vessel cladding in pressurized water reactors and core shrouds in boiling water reactors, as well as to investigate leaks in piping and reactor components. These visual tests are performed using a wide variety of procedures and equipment. Techniques generally include the use of submersible closed circuit high resolution video cameras to examine reactor components. There are numerous industrial camera systems available for this purpose. These systems have video resolutions ranging from 470 to 600 vertical lines on the screen, roughly equivalent to a 640 by 480 pixel count, and most systems typically possess zoom capability to achieve fields of view of 24 by 18 mm (0.9 by 0.7 in.) or smaller. The camera systems record the tests either digitally or using Super VHS tape for review offline and to document the results. The cameras typically have a pair of spotlights mounted near the lens to provide illumination. The authors are aware of at least one system that has a light emitting diode ringlight mounted around the lens. There is no standard technique for visual test lighting.

 

Wire Resolution Tests
Nuclear utilities now follow guidelines for remote visual testing found in BWRVIP-03: Reactor Pressure Vessel and Internals Examinations Guidelines (EPRI, 1995). These guidelines specify that test surfaces must be clean and, for underwater testing, that the water be clean and clear. The guidelines also describe training requirements for personnel and specify which areas around a weld should be tested, how to measure the sizes of indications found and how to test the resolving power of the visual equipment used for the test. There are no guidelines dealing with scanning speed or field of view used during the test. To test the visual acuity of the camera system and lighting, the guidelines call for the camera system to image a sensitivity, resolution and contrast standard before and after the test. This sensitivity, resolution and contrast standard typically contains two perpendicular 12 µm (5 x 10^-4 in.) diameter wires as a resolution calibration standard. If the camera and lighting are sufficient to detect the wires, then the camera system is deemed to have a sufficiently high resolution for the test. The very important issue of lighting is also presumed covered by this line detection test.

There are two main problems with the sensitivity, resolution and contrast standards described in the BWRVIP-03 report. The first problem with using two wires is that line detection is a poor measure of visual acuity. Although using line detection to test a camera system has a long history, it is not well respected by many experts in VT (ASNT, 1993) and may allow one to detect a blurred and low contrast image associated with poor focus and insufficient magnification and pixel size. A simple line detection test is inferior to a resolution test using a calibrated resolution target. An example of a camera system using a 12 µm (5 x 10^-4 in.) wire detection test is shown in Figure 3. It should be noted that the resolution target also contained a series of letters for use as a reading chart. For ease of viewing and printing in this paper, contrast in Figure 3 has been digitally enhanced. The wires are detectable as low contrast fuzzy linear indications against the 18% neutral gray card. The video camera used in this example has a pixel size of 7.5 pixels/mm (190.5 pixels/in.) or 125 µm (5 x 10^-3 in.). The 12 µm (5 x 10^-4 in.) wire is one tenth of the width of each of the pixels and is well below the resolution of the system at this magnification. Had the contrast between the wire and the background been weaker, or if any visual noise were near the wires, the wires would probably not be detectable.

The second problem is that the commonly used 12 µm (5 x 10^-4 in.) diameter wire test can be very misleading, as wires have a different response to lighting angle than an embedded, surface breaking crack. This is because wires may cast shadows and generally provide an excellent source of reflected light, while cracks do not cast shadows and only reflect light from a small range of angles. These differences between wires and cracks are shown in Figure 4. As the current technique for determining the optimum lighting for VT is to arrange the lighting to allow the system to see the wires, this method optimizes the lighting to detect features with the wrong geometry.

 

Lighting
Lighting is an extremely important factor in visual testing. Issues facing inspectors include glare off shiny (specular) metal surfaces, shadowing caused by geometry and other issues. There are currently no standards for lighting to image cracks in the nuclear industry. Most systems possess two spotlights mounted near the camera lens. The measure of success for the lighting system is measured by the spotlights' ability to detect the two wires on the sensitivity, resolution and contrast standard before and after testing. As was described previously, this technique does not ensure that the lighting is optimized for detecting cracks. Typically, bright field imaging of specular reflecting surfaces is done using diffuse lighting that minimizes glare and the effect of machining marks and other shallow scratches. In machine vision applications where one is testing a shiny surface for discontinuities, diffuse lighting or domes emitting light at many angles are used to provide directionless illumination and very little glare. Examples of the use of diffuse lighting to overcome glare caused by direct lighting are shown in Figure 5. The same thermal fatigue crack is shown in each image. Note the noise due to surface scratches when direct lighting is used as opposed to when the lighting is diffuse. Glare can also be reduced through the use of polarized light. Light from a source is polarized in one direction and the camera lens has a polarizer which can be oriented 90 degrees to the light source. Specular reflected light from the surface would be reduced and reflected light from other features would pass through.

Another technique often used to find cracks in other industries is through the use of dark field lighting, in which the surface is illuminated at a sharp oblique angle. The surface itself does not reflect the light and appears dark, while surface discontinuities show up as bright indications. Using dark field lighting, a system can detect very tight cracks quite easily. Dark field lighting would not be very effective on a surface with scratches or machining marks, however, as they could mask any cracks. A combination of polarization, dark field and diffuse bright field lighting would offer a strong improvement over the current spot lighting used today. Any of these alternative light sources would result in much less glare than is produced by spot lighting and enhance crack detection.

 

Camera Resolution
Finally, the current systems used for VT in the nuclear industry have video resolutions on the order of 480 lines. This resolution has been sufficient to pass the current acuity tests, but is not likely to detect tight cracks or pass more rigorous visual acuity tests. Better cameras are currently available for machine vision applications, with some systems capable of images of more than six million pixels. In addition, high definition camcorders with video resolutions of 1080 lines are commercially available to the public. Improving the camera video resolution would allow the inspector to view the tested area with much greater clarity and a smaller pixel size will provide a higher contrast between the cracks and the tested surface.

 

SERVICE INDUCED CRACK SIZES
Based on the previous discussion, one can begin to understand the myriad of parameters that play a role in the reliability of remote visual testing. But in order to determine if a remote visual testing system is capable of detecting actual cracks, a discussion of typical service induced crack dimensions is needed. The primary feature of a crack to be visually detected is its width, or crack opening dimension. The crack opening dimension is a function of several factors, some of which are material hardness, applied loads, crack length, residual stresses around the crack opening and the degree of corrosive attack at the crack opening. The specific variables of most importance to crack opening dimension depend on the type of crack involved. For instance, literature reports that the width of intergranular stress corrosion cracking is fairly random and is primarily a factor of how many grain boundaries at the crack opening are affected.

Several hundred cracks of various types and origins in many materials have been characterized and documented in the literature of the United States and Europe. The results show that service induced crack opening dimensions are highly variable over most crack types and materials and several outlier sizes were found that increase the range of the data sets. However, it was found that most reported crack opening dimensions tend to be populated around a median crack width. Table 1 provides a compilation of crack opening dimension ranges for various types of service induced degradation. It is important to note that the median values for these cracks are very tight, with crack opening dimensions on the order of 10 to 40 µm (4 x 10^-4 to 1.6 10^-3 in.). The results show that crack opening dimension is highly variable over all crack types and materials. This presents a significant challenge when attempting to detect them using remote video camera systems. More importantly, the findings also indicate that crack opening dimension is largely independent of the crack throughwall depth, that is, judging the overall crack depth by crack opening dimension is not reliable (Ekström and WŒle, 1995). Therefore, even if the crack opening dimension is large enough to be detected visually, other volumetric NDT methods must be used to fully characterize the crack boundaries.

 

CRACK DETECTION WITH REMOTE VISUAL TESTING
Based on a brief literature search, few comprehensive studies of the probability of various video systems used for remote visual testing to detect cracks relative to crack opening dimension have been published to date. Work has been conducted in Germany on visually testing reactor components (D'Annucci, 2001) with results reporting the ability to resolve indications as small as 7 µm (3 x 10^-4 in.) in size, but only with a system capable of applying a dye in order to enhance visual performance (a liquid penetrant test is typically not performed underwater). A more standard visual system was used in Sweden to test reactor components and the reported detectable limit was 20 µm (8 x 10^-4 in.) for discontinuities (Efsing et al., 2001). Useful information on the evaluation of remote visual testing was found in a recent human factors study performed in Sweden (Enkvist, 2003). A series of cracked ceramic specimens, molded to reproduce the surface appearance of a welded region, was examined underwater by 10 operators using a high resolution 752 by 582 pixel video camera, an 18x optical zoom and with lighting provided by two 15 W halogen lamps. Only one viewing angle and a single distance of 200 mm (7.9 in.) from the test samples were used, so the tests were somewhat more restrictive than actual field visual testing. The area tested by the system at maximum magnification was 47 by 35 mm (1.85 by 1.46 in.) with a resulting pixel size of 60 µm (2.4 x 10^-3 in.), therefore this study likely represents the capabilities of optimized remote visual testing such as one might observe under laboratory conditions. The operators were instructed to note all linear indications that could be detected and label them as definite cracks, probable cracks, probable no cracks and definite no cracks. Detection and false call rates were then evaluated under "strict" guidelines, which counted only definite cracks as hits; "normal" guidelines, which counted both definite cracks and probable cracks as hits; and "lenient" guidelines, which counted all but definite no cracks as hits. The results are shown in Figure 6. Cracks above 40 µm (1.6 x 10^-3 in.) crack opening dimension were easily detected, while cracks below 20 µm (8 x 10^-4 in.) crack opening dimension had, at best, a 20% probability of detection using the "lenient" grading scale.

A recent study on the detectability of tight thermal fatigue cracks under normal test conditions was performed by Virkkunen et al. (2004) using a commercially available remote camera system. In this study, the camera was focused on an area 24 by 18 mm (0.94 by 0.71 in.), resulting in a pixel size of 37.5 µm (1.5 x 10^-3 in.). The cracks ranged in opening dimension from less than 20 to 200 µm (less than 8 x 10^-4 to 8 x 10^-3 in.). The cracked area was scanned at 20 mm (0.8 in.) per second, the data recorded and the images later reviewed frame by frame. Identified areas of interest were then reexamined statically with a focal area of 12 by 9 mm (0.47 by 0.35 in.). This careful scanning, testing and retesting showed that the smallest cracks that could be reliably detected were of 100 µm (3.9 x 10^-3 in.) opening dimension or larger; the smallest discontinuities possible to detect were of 40 µm (1.6 x 10^-3 in.) opening dimension. The detection rate for cracks smaller than 100 µm (3.9 x 10^-3 in.) opening dimension was approximately 20%.

The two studies discussed above show that video equipment is able to detect moderately tight cracks only with stationary imaging. However, when comparing the results from the stationary camera system that performed well for cracks with opening dimensions above 40 µm (1.6 x 10^-3 in.) to the compiled crack opening dimensions in Table 1, one observes that even a stationary camera system may miss a significant number of cracks. The median fatigue crack size of 17.5 µm (7 x 10^-4 in.) for mechanical fatigue cracks is well below the threshold for reliable detection and the median stress corrosion crack opening dimension of 30 to 40 µm (1.2 x 10^-3 to 1.6 ´ 10^-3 in.) is within a zone of mediocre to poor detection. Furthermore, when the camera is scanned at 20 mm/s (0.8 in./s), the reliable threshold for detection increases to a crack opening dimension of approximately 100 µm (3.9 x 10^-3 in.). Therefore, it appears that a continuously moving camera system, even when the scan is reviewed frame by frame, will miss all but the widest of cracks present in the material.

 

CONCLUSION AND RECOMMENDATIONS
Based on the analysis documented in this paper, the following conclusions can be drawn.

• Although current remote visual testing systems have been demonstrated to be able to detect wide cracks, many service induced cracks may have crack opening dimensions much smaller than those reliably detected in the referenced studies. More than half of the fatigue cracks and 25% of the stress corrosion cracks in one study (Ekström and Wåle, 1995) have crack opening dimensions of less than 20 µm (8 x 10^-4 in.). Virtually all of the cracks had crack opening dimensions less than 100 µm (4 x 10^-3 in.), which may be the reliable limit for crack detection with a continuously scanning system (Virkkunen et al., 2004). The small crack sizes in reactor components relative to the reliably detected crack sizes suggest that current remote visual testing systems and techniques would probably miss a significant fraction of cracks in nuclear components.

• Since no correlation exists between crack opening dimension and throughwall dimension of a crack, volumetric NDT methods such as ultrasonic or radiographic testing may be necessary to fully characterize the extent of cracks detected with remote visual testing systems in order to determine the structural integrity of components.

• The current standard resolution target, two crossed 12 mm (4.7 x 10^-4 in.) diameter wires, is not sufficient to determine the actual resolution and visual acuity of most remote visual testing systems used in the field. A rigorous resolution test should include a reading chart and a resolution chart and possibly a tight electrodischarge machined notch or an actual crack to determine optimum illumination angles. Lighting angles and style (bright field, dark field or diffuse) need to be optimized for crack detection based on geometry and surface conditions, not wire detection.

• A better understanding of how visual acuity performance relates to crack detection as a function of crack opening dimension, lighting angles, lighting methods and other important variables, is necessary to ascertain the overall reliability of remote visual testing.

• It is possible to accurately detect very tight cracks with a remote visual testing system possessing a sufficiently high resolution and using high magnification and by taking a series of stationary images (or scanning and incrementally stopping over each region). However, values for the resolution, magnification and maximum scanning speeds for such a test system are currently unknown.

• Because remote visual testing systems lose visual acuity as the camera scans over a surface, the camera almost certainly needs to be held stationary when taking critical images.

• Parametric studies are needed to assess many of the variables influencing remote visual testing and it may be discovered, due to inherent variability in systems and applications, that round robin testing may be necessary to determine the true reliability of remote visual testing.

• Anecdotal information suggests that, in commercial nuclear power environments, crack detection is enhanced because of the "patina" or inside surface oxide that is formed during operation. The Pacific Northeast National Laboratory plans to evaluate this information in future research activities.

 

ACKNOWLEDGMENTS
This work is being sponsored by the US Nuclear Regulatory Commission under contract DE-AC06-76RLO 1830; NRC JCN Y6604; Wallace Norris, NRC program monitor.

 

REFERENCES
American Society for Nondestructive Testing, Nondestructive Testing Handbook, second edition: Volume 8, Visual and Optical Testing, Columbus, Ohio, ASNT, 1993.

American Society of Mechanical Engineers, Boiler and Pressure Vessel Code, New York, ASME, 2001.

D'Annucci, F., "Latest Technologies and Results in Reactor Vessel Head Penetration Inspection," Proceedings of the 3rd International Conference on NDE in Relation to Structural Integrity for Nuclear and Pressurized Components, Seville, Spain, Tecnatom SA, 2001, pp. C 25-32.

De Petris, C. and C. Macro, "Verification of the Resolution Capability for Equipment Used for Visual Testing," Proceedings of the 15th World Congress on Non-destructive Testing, Brescia, Italy, Italian Society for Non-destructive Testing and Monitoring Diagnostics, 2000, NDT.net, <www.www.ndt.net/article/wendt00 /papers/idn300/idn300.htm>.

Dorman, W.J., "Brunswick Steam Electric Plant, Units 1 and 2, Docket Nos. 50-325 and 50-324, License Nos. DPR-71 and DPR-62, Request for Approval of Revised Relief Request for the Third 10-Year Inservice Inspection Program," correspondence, Carolina Power and Light to US NRC, 7 August 2000.

Efsing P., J.-Å. Berglund, C. Sandelin and A. Werner, "Visual Inspection of Brackets for Emergency Core Cooling System in Barsebäck Unit 2," Proceedings of the 3rd International Conference on NDE in Relation to Structural Integrity for Nuclear and Pressurized Components, Seville, Spain, Tacnatom SA, 2001, pp. C 43-51.

Ekström P. and J. Wåle, Crack Characterization for In-service Inspection Planning, SKI Report 95:70, Stockholm, Sweden, Swedish Nuclear Power Inspectorate, 1995.

Electric Power Research Institute, BWRVIP-03: Reactor Pressure Vessel and Internals Examinations Guidelines, Palo Alto, California, Electric Power Research Institute, 1995.

Enkvist, J., "A Study of Operator Performance in a Visual NDT Inspection Task by Remote Video Camera," Insight, Vol. 45, No. 4, 2003, pp. 252-257.

Institute of Electrical and Electronics Engineers, STD 208-1995: Measurement of Resolution of Camera Systems, New York, IEEE, 1995.

Jensen, N., Optical and Photographic Reconnaissance Systems, New York, John Wiley & Sons, 1968.

Lapides, M.E., "Evaluation of Miniac Radiography for In-service Inspection of Cast Austenitic Piping and Components in PWR Nuclear Units," EPRI Draft Document, September 1984.

MacDonald, D.E., IGSCC Detection in BWR Piping Using the Miniac, EPRI Report NP-3828, February 1985.

Moore, M., B. Phares, B. Graybeal, D. Rolander and G. Washer, Reliability of Visual Inspection for Highway Bridges - Volume I: Final Report. FHWA-RD-01-020, McLean, Virginia, NDE Validation Center, Office of Infrastructure Research and Development, Federal Highway Administration, 2001.

Sine Patterns, Standard Charts, Pittsford, New York, Sine Patterns LLC, 2004, <www.sinepatterns.com/Standards.htm>.

Stenefjall, T., private correspondence, September 1994.

US Nuclear Regulatory Commission, Title 10, Part 50: "Domestic Licensing of Production and Utilization Facilities," Code of Federal Regulations, Washington, DC, US Nuclear Regulatory Commission, 1956.

Virkkunen, I., M. Kemppainen, R. Paussu, P. Seppälä, D.S. Dybal and A.A. Nikitin, Cracked Samples for Visual Testing, Espoo, Finland, Trueflaw, 2004.

*  Nondestructive Measurement and Characterization Sciences Group, Pacific Northwest National laboratory, PO Box 999, MSIN: K5-26, Richland, WA 99352; (509) 375-2523; e-mail <michael.anderson@pnl.gov>.

⁺ Nondestructive Measurement and Characterization Sciences Group, Pacific Northwest National laboratory, PO Box 999, MSIN: K5-26, Richland, WA 99352.

 

Copyright © 2005 by the American Society for Nondestructive Testing, Inc. All rights reserved.