Plate-type nuclear fuel has been adopted in most research reactors. Such fuel consists of a fuel core in aluminum alloy. The production quality of nuclear fuel is an important part of an efficient and stable generation of thermal energy in research reactors. Thus, a nondestructive quality inspection of the internal defects of plate-type nuclear fuel is a key process during the production of nuclear fuel for safety insurance. Nondestructive quality inspections based on X-rays and ultrasounds have been widely used for defect detection of plate-type nuclear fuel [1-6].

X-ray testing is a simple and fast inspection method, and provides an image in real-time [1-2]. Thus, a non-expert field worker can carry out the testing. However, it is difficult to detect closed-type defects because an X-ray image is made from the density differences of the materials. In particular, it is hard to detect delamination defects that should be detected during the production of plate-type nuclear fuel. Thus, an X-ray inspection is usually used for an inspection of the filling status of nuclear fuel in aluminum alloy.

Ultrasonic testing is a powerful tool to detect internal defects including open- and closed-type defects, in platetype nuclear fuel [3-6]. Such testing can also provide thickness information for each internal fuel plate. A high frequency ultrasonic signal of over 20 MHz is usually adopted because the thickness of each plate in plate-type nuclear fuel is within the millimeter to sub-millimeter range. Thus, the inspection process is complicated, as an immersion test should be carried out in a water tank. It is also a time-consuming inspection method because area testing to acquire an image is based on the scanning of point-by-point inspections. In addition, this testing method needs to be carried out by an expert since the detection capability is greatly dependent on the inspector’s knowledge.

Among nondestructive imaging methods, techniques based on laser interferometry and infrared thermography has been widely used in the detection of internal defects of plate-type materials, such as those used in aircrafts and automobiles. To detect internal defects, an infrared thermography technique (IRT) analyzes the thermal behavior of the specimen surface and a laser interferometry technique (LIT) analyzes the deformation field.

IRT presents some disadvantages such as emissivity variations, and external reflections affecting the temperature pattern. In addition, there is a lack of information in the application of an IRT in a non-controlled environment. An LIT is a non-contact inspection method and is faster than thermography. This technique requires less energy than thermography and the signal processing is almost instantaneous. As a disadvantage, an LIT is more sensitive to mechanical vibrations. Thus, to properly detect internal defects, several inspection parameters, such as the acquisition time, external stimulation, and vibration environment, must be optimized when the assessment procedure is developed. Both techniques are useful tools for the detection and evaluation of internal defects in plate-type materials [7].

A laser speckle interferometer is a noncontact optical device used to measure micro changes in the surface topography [8]. An interferometer is sensitive to very small displacement at a resolution of nanometers by superposing the speckle patterns of two different object states. Amplitude differences in the deformation among intact and defective areas have been widely used for the detection of internal defects in plate specimens [9-13]. Phase differences have also been used when the phases among intact and defective areas are different during the periodic deformation of a plate specimen [14-18]. Here, the phase difference is the time delay between an intact surface and a defective surface during the periodic movement of the sinusoidal pattern. Amplitude information on the modulation frequencies, and nonlinear information on the harmonic frequencies, has also been used to detect internal defects including small closed-type defects [14-19].

In this paper, an imaging technique to visualize the internal defects in plate-type nuclear fuel was developed for a nondestructive quality inspection using an active laser speckle interferometer with a periodic thermal wave. The developed visualization technique measures an interference image, which is a micro deformation image of a specimen surface including the internal defect information, at the effective detection time, and accumulates the defective components after extracting them from each measured image. By virtue of a real-time display for an accumulated defect image, a non-expert field worker can carry out the inspection. The experimental results using the developed optical visualization technique with an active laser interferometer to detect the internal defects of a plate-type nuclear fuel specimen are described in this paper.

A block diagram of the configured system used to detect internal defects using an active laser interferometer with a thermal wave is shown in Fig. 1. The noncontact visualization system was configured using two halogen lamps with a driver (DDS 5600, LEVITON MFG. co) controlled by a signal generator, a test specimen held by a holder, a digital speckle pattern laser interferometer (CW laser: 532 nm, 100 mW, Cobolt Samba™ 150) with a PZT driver (E-665.XR, LVPZT-AMP, PI), and a computer.

As shown in Fig. 1, the deformation shape of the specimen surface will vary according to the applied thermal wave on the backside of the specimen. Two halogen lamps driven by an amplifier generate the thermal wave. The intensity of the thermal wave is a periodic sinusoidal pattern controlled by a signal generator. The digital speckle pattern interferometer acquires a deformation interference image for the front surface of the specimen. The deformation image is an interference image between a reference laser beam and signal laser beam. As shown in Fig. 1, a reference laser beam reflected at the beam splitter (BS) goes into a CCD camera after passing through a mirror (M), spatial filter (S. F), mirror (M), and focusing lens. The signal laser beam is projected onto the specimen surface after passing through the beam splitter (BS), mirrors (M), and spatial filter (SF), and the reflected beam then goes into the CCD camera. Here, the phase of the signal beam can be shifted using a piezoelectric transducer (PZT) controlled by a computer. Internal defect information is contained in the acquired interference image as the deformation difference is also caused by a reflection of the thermal wave on the internal defects. An extracted deformation image including internal defect information will be produced after the signal processing of the laser fringe images by the computer.

A test specimen was prepared for the experiments using a noncontact optical visualization system for the test of plate-type nuclear fuel. The test specimen was composed of three plates (A, B, C), as shown in Fig. 2. Instead of uranium nuclear fuel, a tungsten plate (B) with a similar density was used. In a common structure of plate-type nuclear fuel, a plate-type tungsten core (B) is placed at the center of an aluminum alloy plate (A, C). The thickness of the tungsten plate (B) was 1.0 mm, and the thicknesses of the aluminum alloy (A, C) at the front and backsides were both 0.5 mm. Here, the typical thickness of each component widely used in the field was selected. Two rectangle polyethylene films, D and E, are inserted between plates A and B to make delamination-type defects. Three

plates were then bonded by metal paste, except the two areas of D and E, as shown in Fig. 2. The size of D was 3.0 mm x 3.0 mm, and that of E was 5.0 mm x 5.0 mm. The D and E areas are non-bonded defects, and the thickness of the inserted film was about 100 μm. Here, the dotted rectangular is the viewing area of the camera, and specimen holders (B1, B2) are placed on both the left and right sides of the specimen.

An ultrasound immersion inspection based on pointby- point scanning is currently used for the detection of internal defects in plate-type nuclear fuel, although it is a time consuming and complicated inspection method. The inspection should be carried out by an expert since the detection capability is greatly dependent on the inspector’s knowledge; it is a powerful nondestructive testing tool.

In this paper, high-frequency ultrasonic test was carried out to detect the internal defects of a test specimen. A C-scan image using a high-frequency ultrasonic inspection system (SPICA™, Acoulab Co. Ltd.) is shown in Fig. 3. An immersion test using a high frequency of 50 MHz was carried out in a water tank. The image was acquired from measuring 350 x 1100 points with a scanning resolution of 0.1 mm. The ultrasonic inspection provided precise information for the delamination defects, as well as difficult information to analyze, as shown in Fig. 3.

A periodic thermal wave with sine pattern intensities was applied to the specimen. The shape of the specimen deformation will be periodically varied in proportion to the thermal power. The applied pattern of the thermal wave is shown in Fig. 4.

To obtain a relative deformation shape, a reference phase image was extracted using a digital laser speckle interferometer at the reference voltage time of (ref.) set by the user. In this paper, the lowest thermal power time was selected as a reference acquisition time because the

remaining strain distribution at this reference voltage time is distinct over a long period of time, as shown in Fig. 5. Here, the reference phase image was made from 4 speckle images with phases of 0, 90, 180, and 270 degrees at the time of (ref.) using a 4-bucket algorithm [20]. After making a reference phase image, the control computer extracted the current deformation image using the subtraction of the reference phase image from the current phase image, which was also calculated from the 4 phase-shifted speckle images at the current time. In this experiment, the control computer acquired about 33 deformation images in one period of 10 seconds.

Measured images at times of t₉, t₁₇, t₂₅, and t₃₃ in the

4^th period are shown in Fig. 5. High-pass filtered images corresponding to the images in Fig. 5 are shown in Fig. 6. Here, the starting pass frequency was 5 pixels in a 512x512 pixel image. The removed low frequency noises were usually caused by the body movement of the specimen according to the thermal power. As shown in Figs. 5 and 6, the internal defect information can be discriminated in all acquired images, and the visibilities of the internal defects are varied according to the thermal power. Here, the best visibility for the internal defects is provided at the acquisition voltage level of the reference image because the noises caused by the whole body variation are the lowest at this level.

To develop a valuable nondestructive inspection system, an efficient noise rejection technique and a display technique for internal defect images are needed. For this purpose, a signal processing technique was developed as shown in Fig. 7. Here, the defect information was extracted by subtracting a low frequency image (L_i) from the unwrapped phase image (P_i). To robustly extract the noise components of low frequencies, a smoothing filter of N x N pixels was applied. The extracted defect components were accumulated and displayed on a monitor at each measured image.

An accumulated image of the extracted defect components for 496 measured images during the first 15 periods is shown in Fig. 8. As shown in Fig. 8, the internal defects can be discriminated from the accumulation image, which displays the defect components at each acquisition.

An averaged image can also be used to discriminate the internal defects. Averaged images after removing the DC value for 15 and 30 selected images are shown in Figs. 9 and 10, respectively. Here, the selected images were measured images at the reference voltage level. Thus, one image was used per period. As shown in Figs. 9 and 10, though defect images can be discriminated from the averaged image, the discrimination capability was sensitive to low frequency noises.

The computer can acquire interference images when the input voltage of the heat is within the user set level band (V₁~V₂). A more distinct visual defect image can then be acquired by accumulating the defect components after extracting them. Accumulated images after the extraction of the defect components for the same 15 and 30 images are shown in Figs. 11 and 12, respectively. As shown in Fig. 8 to Fig. 12, defect images can be clearly discriminated, and the extracted defect information in the accumulated images is robust from occasionally acquired noisy images, which are characteristics of laser speckle interferometry.

As the experimental results show, the visualization technique based on the active laser interferometer quickly provided an inspection image including information of the internal defects. Though this technique is not powerful when compared with the high-frequency ultrasonic test, it can be a valuable complementary tool because it provides an easily recognizable image, and the image information comes from other areas based on the thermal expansions.

In this paper, a visualization technique to detect the internal defects in a plate-type nuclear fuel was developed for a nondestructive quality inspection. The developed technique visualized a vivid image for internal defects by sequentially accumulating defect components after extracting them from each acquired image. The accumulated image for the extracted defect components from the measured image at the acquisition voltage level of the reference provided improved defect information owing to the use of images with a high signal-to-noise ratio. The visualization technique, which is based on the accumulation of defect components by subtracting smoothing filtered components, is robust against occasionally induced noisy images. As the experimental results show, the developed visualization technique can be valuable for a nondestructive quality inspection of plate-type nuclear fuel.