Nondestructive Evaluation of Thick Concrete Structures

Transcription

1 Nondestructive Evaluation of Thick Concrete Structures Dwight A. CLAYTON Electronics and Electronic Systems Research Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee, USA, Phone: ; Abstract Materials issues are a key concern for the existing nuclear reactor fleet in the United States as material degradation can lead to increased maintenance, increased downtime, and increased risk. Extending reactor life to 60 years and beyond will likely increase susceptibility and severity of both known and new forms of degradation. A multitude of concrete-based structures are typically part of a light water reactor plant to provide foundation, support, shielding, and containment functions. The size and complexity of nuclear power plant containment structures and the heterogeneity of Portland cement concrete make characterization of the degradation extent a difficult task. This paper examines the benefits of using time-frequency analysis with Synthetic Aperture Focusing Technique (SAFT). By using wavelet packet decomposition, the original ultrasound signals are decomposed into various frequency bands that facilitates highly selective analysis of the signal s frequency content and can be visualized using the familiar SAFT image reconstruction algorithm. Keywords: Concrete, nondestructive evaluation, nuclear power plant, sustainability, frequency band analysis Introduction Unlike most metallic materials, reinforced concrete is a nonhomogeneous material; a composite with a low-density matrix, a mixture of cement, sand, aggregate and water, and a high-density reinforcement (typically 5% in nuclear power plant containment structures) made up of steel rebar or tendons. Figure 1illustrates the vast amount of rebar that appears in a typical nuclear power plant (NPP). NPPs have typically been built with local cement and aggregate fulfilling the design specification regarding strength, workability, and durability, but as a consequence, each plant s concrete composition is unique and complex. Figure 1. Rebar during the construction of a NPP.

2 NPP concrete structures are often inaccessible, containing large volumes of massively thick concrete sections that are exposed to different environments (moisture, temperature) and a variety of degradation mechanisms (high temperatures, radiation exposure, chemical reactions) at different plant sites, all of which add to the complexity of determining the integrity/quality of the concrete. Often only one side of the concrete structure is readily available or one side has a steel liner, as illustrated in Figure 2. Since the aspect ratio is so different from typical transportation concrete structures due to the thickness of concrete sections, additional reflections from lateral boundaries add to the complexity of received ultrasonic signals. Figure 3 illustrates the multiple ultrasonic reflections that are possible when the source is near a lateral boundary. Figure 2. A typical NPP containment wall. Figure 3. Ultrasonic waves in a thick concrete specimen. In ORNL/TM-2013/430 [1], seven different nondestructive evaluation (NDE) technologies were compared using two concrete test specimens from the Florida Department of Transportation s NDE Validation Facility in Gainesville, Florida. As documented in that report, ultrasonic linear array devices with 40 or more transducers performed best at volumetric imaging. These devices are based on the pitch-catch method of sending and receiving shear wave impulses at the surface, requiring only one-sided access and receiving the echoes at the original surface. 1. Large Concrete Specimen Representative of NPP Containment Comparative testing on the various NDE concrete measurement techniques will require concrete specimens with known material properties, voids, internal microstructure flaws, and reinforcement locations. To minimize artifacts caused by boundary effects, the dimensions of

3 the specimens should not be too compact. The exact size of the specimen depends on the NDE method used. The minimum dimensions of the test specimen are directly related to the thickness of the specimen. For many NDE ultrasonic techniques, the first reflected wave received is normally assumed to be from the rear surface. If the ultrasonic wave is modeled as a spherical propagation from the point source, the distance from the source location to the rear surface must be the minimum dimension Size Considerations Since it is not feasible to build specimens that are to the scale of in-service structures, compact sample specimens must be built while still replicating the NDE needs of real structures. This includes minimizing artifacts caused by boundary effects. Although significant NDE research has been conducted on relatively thin concrete structures to assess pavement, bridge decks, and other infrastructures applications, construction of large reinforced concrete specimens specifically for NDE comparisons is less common [2]. Comparative studies have shown various NDE techniques to be successful in identifying the types of internal characteristics of interest. However, these applications are typically conducted to evaluate thin specimens (~1 ft. in thickness), while NPP containment walls are often much thicker (over 3 ft. in thickness). Even though previous results for thinner specimens show promise in the ability to nondestructively evaluate internal characteristics, the results need to be validated for thicker and more heavily reinforced structures. Two major NDE challenges associated with thicker structures remain: 1) low signal-to-noise ratio with greater depths due to heterogeneous materials with a dense and complex arrangement of reinforcements, and 2) effects from boundaries at similar distances to the region of interest. While the design of a large concrete specimen mitigates some of the boundary effect concerns, it creates some additional complexities involved with forming such a large reinforced concrete specimen. Beyond the efforts required to cast and properly consolidate a large concrete specimen, the ability to maneuver and transport the specimen can be restrictive. Often specimens need to be tilted to allow for actuator loading at the correct orientation or, in this case, NDE data collection access. Additionally, there are times where the specimen needs to be cast in a different location than the testing location to mitigate concrete truck or other access issues. The weight of the specimen is a major factor in this regard, where reinforced concrete is typically 150 lbs/ft 3 and 162 lbs/ft 3 assuming a reported NPP 5% steel by volume ratio. This can be restrictive for the use of a typical structural laboratory crane s 20-ton (40,000 lbs) load capacity. There can also be restrictions due to the large specimen dimensions even if the specimen is cast in the location and orientation necessary to conduct the testing. While infrastructure specimens can easily be cut to desired dimensions along the thickness cross section for disposal, this technique is not straightforward for disposal of the specimens meeting NPP containment structure thickness requirements. This can also create restrictions beyond the weight limits of crane operation. For example, the smallest dimension of the specimen and crane fixture mechanism is required to be less than the smallest dimension of clearance between the crane and floor along the path to the disposal site, assuming the specimen is instrumented to allow for rotation to the desired specimen orientation Dimensions and reinforcement Concrete structures in NPPs are typically 3 to 4 feet thick along the wall and dome of the containment structure. Therefore, a specimen thickness of 3 feet 4 inches (1.016 m) was

4 chosen. This thickness is consistent with NPP containment wall specifications. The height and width of the specimen was constricted to 7 feet to accommodate use of a maximum 20- ton crane while still mitigating boundary effects. NPP concrete is normally embedded with heavily reinforced cross sections using mild steel #18 bars (2.257 inch diameter, 4.00 in 2 cross sectional area) and #8 bars (1.000 inch diameter, 0.79 in 2 cross sectional area) at 6 to 12 in spacing. The type and spacing of the reinforcement has a significant effect on shielding evaluation of defects below the level of reinforcements. While elastic wave-based methods are less sensitive to reinforcement than ground penetrating radar (GPR), characterization of defects within more heavily reinforced structures are more difficult than for less heavily reinforced structures. The constructed specimen contained #18 rebar at 12 in. spacing in both horizontal and vertical orientation. This provides a realistic reinforcement size that also allows for space between reinforcement for semi-controlled evaluation of the effects of concrete depth on defect characterization. This will also allow for differentiation between complexities caused by dense levels of reinforcement versus complexities caused by depth of penetration within concrete, while still observing the effects of the uniquely large diameter reinforcement used in NPP structures. The #18 reinforcement size and arrangement at an intermediate stage of the formwork process is shown in Figure 4. Figure 4. Reinforcement typical of a NPP containment wall using #18 rebar Specimen material and simulated defects Concrete material properties of the various NPP concrete structures depend on the materials used during construction of the specimen. The material used at each NPP site is variable depending on the distance from aggregate and cement sources in the area. Additionally, the early age material properties resulting from each mix design are different than the material properties 50 years after the beginning of the curing process. Theoretically, the mix design of example NPP locations could be replicated to determine the effect of material properties on NDE methods. However, since the mix designs are variable from site to site, materials from 50 years ago are not easily obtained, and early age properties are not representative of older concrete, the proposed mix was designed to make the Portland cement concrete (PCC) matrix surrounding the simulated defects and reinforcement pattern as controlled and consistent as possible. By taking into account the large size of the specimen, complex nature of the reinforcement and simulated defects, a self-consolidating (SCC) performance-based mix was chosen. The use of SCC mix provided consistent consolidation and low stress on embedded defects without the need for vibration. This also mitigated concerns of the simulated defects damaging or moving from their desired location during the pour.

5 The simulated defects were embedded to determine how the current state-of-the-practice NDE techniques will be able to determine various forms of degradation in NPP concrete structures. This is a difficult task since, to date, limited comparisons of state-of-the-art methods have been conducted at the size and reinforcement density of LWR containment structures on controlled specimens, or verified through forensic activities. The constructed specimen is designed to allow for assessment of controlled benchmark defects from previous studies in a more heavily reinforced concrete structure as well as evaluation of realistic defects to ensure that the correct type of features for effective NPP NDE are included. The controlled defects embedded in the proposed wall sample include sufficiently challenging inclusions to ensure that limitations of even the most advanced methods can be quantified. At the same time, some of the proposed defects are designed to be identifiable by a majority of the methods. This will ensure that the methods that are not close to the desired achievement can be eliminated from consideration, while the baseline level of achievement of the methods performing well can be identified. The realistic defects are designed to represent activities that could have occurred during the construction process and/or cumulative deterioration and degradation of the concrete with time. Some of the aging related degradation mechanisms cannot be reproduced due to time constraints, while more realistic construction defect simulation can cause less repeatable results and can be difficult to quantify. However, designing the defects solely to be repeatable and not realistic can lead to the wrong conclusions when evaluating the various NDE techniques. For example, the NDE attributes determined to be desirable based on good performance on the test block may not be useful for evaluation of commercial concrete NPP structures if the defects are not realistic enough. With these factors in mind, the defects were designed to give a mix of realistic and controlled defects for assessment of both the necessary measures needed to overcome the challenges with more heavily reinforced concrete structures, while also ensuring the correct type of features for effective NPP evaluation have been included. Information on each of the twenty defects is shown in Table 1. Table 1. Defects in the constructed thick specimen. Defect Defect Description Defect Defect Description 1 Porous half cylinder (no cover) 11 Plexiglass 2 Porous half cylinder (no cover) 12 Dissolved Styrofoam (medium) 3 Porous half cylinder (no cover) 13 Styrofoam (medium) 4 Porous half cylinder (cover) 14 Plexiglass 5 Porous half cylinder (cover w/crack) 15 Dissolved Styrofoam (thin) 6 PVC 16 Lumber (2x4) 7 PVC 17 Gloves 8 Dissolved Styrofoam (thick) 18 Debond duct tape (one layer) 9 Foam (thick) 19 Debond duct tape (multi-layer) 10 Styrofoam (thin) 20 Moving rebar

6 2. Analysis The additional thickness of concrete in NPP applications drastically decreases the signal-tonoise ratio on returned ultrasound signals since the signals must travel through the concrete twice so the echo can be received and analyzed. This reduction in signal to noise necessitates the development of advance signal processing techniques so NPP concrete structures can be examined. This section examines the benefits of using time-frequency analysis with Synthetic Aperture Focusing Technique (SAFT). SAFT is an image reconstruction technique commonly used in conjunction with ultrasonic arrays. By using wavelet packet decomposition, the original ultrasound signals are decomposed into various frequency bands. Selected frequency bands are then reconstructed back into a time-series dataset [3]. This facilitates highly selective analysis of the signal s frequency content and can be visualized using the familiar and reliable SAFT image reconstruction algorithm. [4] 2.1. Detecting embedded lumber To simulate construction debris inappropriately being part of the concrete structure, a short piece of lumber (2x4) was placed into the thick specimen as shown in Figure 5 (Defect 16 in Table 1). When imaged from the shallow side, both standard SAFT (Figure 6) and the frequency banded SAFT (Figure 7) illuminates the lumber defect. However when imaged from deep cover, the standard SAFT image (Figure 8) does not clearly indicate the defect whereas the frequency banded SAFT (Figure 9) proves clear evidence. Figure 5. A segment of lumber was embedded.

7 Figure 6. Standard SAFT image of the embedded lumber with shallow cover. Figure 7. Frequency banded SAFT image of the embedded lumber with shallow cover.

8 Figure 8. Standard SAFT image of the embedded lumber with deep cover. Figure 9. Frequency banded SAFT image of embedded lumber with deep cover Detecting Dissolved Styrofoam (thin) Similar to some of the defects from Yehia et al. [5], a section of Styrofoam was embedded into the concrete with a tube inserted. The tube was placed to allow pouring acetone into the Styrofoam thereby dissolving the Styrofoam and creating a controlled void that is characteristic of a delamination. Once again the defect is detected using both standard SAFT and frequency banded SAFT when the defect is near the surface. However when imaged from deep cover, the standard SAFT image (Figure 10) does not clearly indicate the defect whereas the frequency banded SAFT (Figure 11) proves clear evidence.

9 Figure 10. Standard SAFT image of dissolved Styrofoam under deep cover. Figure 11. Frequency banded SAFT image of dissolved Styrofoam under deep cover. 3. Conclusions The uniqueness of concrete structures associated with NPPs creates distinctive challenges in NDE. Comparative NDE of various defects in reinforced concrete specimens is a key component in identifying the most promising techniques and directing the research and development efforts needed to characterize concrete degradation in commercial NPPs. This requires access to the specimens for data collection using state-of-the-art technology. Validation data is needed to properly evaluate the effectiveness of the techniques. In this case, the various defects should be created, well defined, and/or feasible to be evaluated forensically. It is also critical that the evaluation specimen and embedded defects are representative of in-service NPP structure concrete.

10 The construction of the specimen detailed above allows for an evaluation of how different NDE techniques may interact with the size and complexities of NPP concrete structures. These factors were taken into account when determining specimen size and features to ensure a realistic design. The lateral dimensions of the specimen were also chosen to mitigate unrealistic boundary effects that would not affect the results of field NPP concrete testing. While standard SAFT techniques are adequate for many defects with shallow concrete cover, some defects that are located under deep concrete cover are not easily identified using the standard SAFT techniques. For many defects, particularly defects under deep cover, the use of frequency banded SAFT improves the detectability over standard SAFT. In addition to the improved detectability, the frequency banded SAFT also provides improved scan depth resolution that can be important in determining the suitability of a particular structure to perform its designed safety function. Acknowledgements The author would like to thank the United States Department of Energy for supporting this research under the Light Water Sustainability Reactor Program. The author would also like to thank his colleagues at the Oak Ridge National Laboratory and the University of Minnesota for their contributions to this research. References 1. Clayton, D. A. et. al. Evaluation of Ultrasonic Techniques on Concrete Structures, ORNL/TM-2013/430, Oak Ridge National Laboratory, (2013). 2. Clayton, D. A. and Smith, C. M., Summary of Large Concrete Samples, ORNL/TM- 2013/223, Oak Ridge National Laboratory, (2013). 3. Clayton, Dwight A., Austin P. Albright and Hector J. Santos-Villalobos. Initial Investigation of Improved Volumetric Imaging of Concrete Using Advanced Processing Techniques. ORNL/TM-2014/362. Oak Ridge, TN: Oak Ridge National Laboratory Albright, Austin P. and Dwight A. Clayton. "The Benefits of Using Time-Frequency Analysis with Synthetic Aperture Focusing Technique." Review of Progress in Quantitative Nondestructive Evaluation, Boise, Idaho, USA, July 20 25, Sherif Yehia et al., "Detection of common defects in concrete bridge decks using nondestructive evaluation techniques," Journal of Bridge Engineering 12(2), (2007).