• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.11a
• /
• pp.625-627
• /
• 2022

In the construction and development of modern industrial production technology, the traditional technology management mode is faced with many problems such as low qualification rates and high application costs. In the research, an improved workpiece defect detection method based on deep learning is proposed, which can control the application cost and improve the detection efficiency of irregular defects. Based on the research of the current situation of deep learning applications, this paper uses the improved Faster R-CNN network structure model as the core detection algorithm to automatically locate and classify the defect areas of the workpiece. Firstly, the robustness of the model was improved by appropriately changing the depth and the number of channels of the backbone network, and the hyperparameters of the improved model were adjusted. Then the deformable convolution is added to improve the detection ability of irregular defects. The final experimental results show that this method's average detection accuracy (mAP) is 4.5% higher than that of other methods. The model with anchor size and aspect ratio (65,129,257,519) and (0.2,0.5,1,1) has the highest defect recognition rate, and the detection accuracy reaches 93.88%.

• Nuclear Engineering and Technology
• /
• v.52 no.6
• /
• pp.1252-1258
• /
• 2020

This study explores the ability of non-destructive assay techniques to detect a partial material defect in which 100 g of plutonium are diverted from the center of a 1000 g can of PuO₂ powder. Four safeguards measurements techniques: neutron multiplicity counting, calorimetry, gravimetry, and gamma ray spectroscopy are used in an attempt to detect the defect. Several materials are added to the partial defect PuO₂ can to replicate signatures of the diverted material. ²⁵²Cf is used to compensate for the doubles neutron counts, ²⁴¹Am is used to compensate for the decay heat, and aluminum is used to compensate for the weight. Although, the doubles and triples difference before and after diversion are statistically indistinguishable with the AWCC in fast and thermal mode, the difference in the singles counts are statistically detectable in both modes. The relatively short half-life of ²⁵²Cf leads to a decrease (three sigma uncertainty) in the doubles neutron counts after 161 days. Combining this with the precise quantity of ²⁴¹Am needed (10.7 g) to mimic the heat signature and the extreme precision in ²⁵²Cf mass needed to defeat neutron multiplicity measurements gives reassurance in the International Atomic Energy Agency's ability to detect partial material defects.

• Proceedings of the Korean Society of Machine Tool Engineers Conference
• /
• 2003.10a
• /
• pp.301-306
• /
• 2003

Spot welding is frequently used for industrial purpose, such as automobile and aerospace industries and household appliances due to its high performance. In these day, robotization and systemization of welding process made it possible to produce more precise or smaller electric parts. And when it comes to welding of steel sheet, the size of nugget must be getting smaller. Therefore, welding conditions are limited to avoid defects, such as deformation, damage, weakening of joining area. In this research, the measurement of the nugget size by the nondestructive inspection has been conducted. As a result, the right estimation of the nugget size and void defects, the detection of corona bond near joining area, the selection of the optimum ultrasonic mode, and set up for ultrasonic inspection are studied. From the trustworthy solutions of nugget size estimated by results of measurement, the optimum inspection conditions depending upon the width of welding parts are determined as well.

• Computers and Concrete
• /
• v.16 no.4
• /
• pp.531-544
• /
• 2015

The objective of this study was to examine the effect of repair mortar on the dynamic properties such as natural frequencies, mode shape and damping ratios of two story single span scale reinforced concrete building. To this end, two story single span scale reinforced concrete building having dimensions of 150 cm (width), 150 cm (length) and 135 cm (height) was constructed. Workmanship defects such as separation of material, faulty vibration application and bad gradation of the structure were properly evaluated. Dynamic properties of damaged structure were experimentally determined using Operational Modal Analysis (OMA). Detected defects in the structure were fixed by plastering with repair mortar. Dynamic properties of repaired structure were reevaluated by using the OMA method. Finite element software called Abaqus was used to numerically determine dynamic properties of the structure. Structure modeled as solid was subjected to Linear Perturbation Frequency Method. The changes in dynamic properties of structure after the repair process were comparatively studied by evaluating experimental and numerical results.

• Proceedings of the KIEE Conference
• /
• 1987.11a
• /
• pp.383-386
• /
• 1987

In this paper, a low frequency acoustic microscope system has been built for the purpose of detecting sub-surface defects in materials. 100 won coin, aluminium, ceramics, I.C. are used for experimental and the acoustic microscope is operated in the focused and defocused mode at a frequency of 3 MHz. In experiment, it has shown that acoustic microscope has the resolution of 0.5mm and the defects presented on the surface and near surface of samples are detected by a monitor.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.31 no.3
• /
• pp.271-277
• /
• 2011

This study presents a nondestructive test for underwater PVC pipes. To use guided ultrasonic waves, specially denoted by cylindrical Lamb waves, a test setup was made in a water tank using the pitch and catch mode and specimens were made to give artificial cutouts located in the circumferential direction of the pipes. Total three states of damaged levels were considered to see how the guided waves interact with the defects. For the experimental adjustments, three different pipe diameters (60, 90, 114 mm) were tested, and two factors - incident angle (10 and $40^{\circ}$) and distance (50 and 200 mm) - were tried. From the results, regardless of the diameters and two experimental factors, it is shown that the degrees of defects were recognized through amplitude and arrived time of the very first part of the received cylindrical Lamb waves. Between amplitude and arrived time, it is found that the amplitude gives more sensitive results.

• Proceedings of the KIEE Conference
• /
• 2003.11b
• /
• pp.271-274
• /
• 2003

This paper presents a Built-in Current Sensor that detect defects in CMOS integrated circuits using the current testing technique. This scheme employs a cross-coupled connected PMOS transistors, it is used as a current comparator. Our proposed scheme is a negligible impart on the performance of the circuit undo. test (CUT). In addition, in the normal mode of the CUT not dissipation extra power, high speed detection time and applicable deep submicron process. The validity and effectiveness are verified through the HSPICE simulation on circuits with defects. The entire area of the test chip is $116{\times}65{\mu}m^2$. The BICS occupies only $41{\times}17{\mu}m^2$ of area in the test chip. The area overhead of a BICS versus the entire chip is about 9.2%. The chip was fabricated with Hynix $0.35{\mu}m$ 2-poly 4-metal N-well CMOS technology.

• Journal of the Korean Magnetic Resonance Society
• /
• v.12 no.2
• /
• pp.74-80
• /
• 2008

The local structure of defects in undoped, Si-doped, and neutron irradiated free standing GaN bulk crystals, grown by hydride vapor phase epitaxy, has been investigated by employing electron magnetic resonance(EMR), Raman scattering and cathodoluminescence. The GaN samples were irradiated to a dose of $2{\times}10^{17}$ neutrons in an atomic reactor at Korea Atomic Energy Research Institute. There was no appreciable change in the Raman spectra for undoped GaN samples before and after neutron irradiation. However, a forbidden transition, $A_1$(TO) mode, appeared for a neutron irradiated Si-doped GaN crystal. Cathodoluminescence spectrum for the neutron irradiated Si-doped GaN crystal became much broader or was much more broadened than that for the unirradiated one. The observed EMR center with the g value of 1.952 in a neutron irradiated Si-doped GaN may be assigned to a Si-related complex donor.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.27 no.8
• /
• pp.1295-1302
• /
• 2003

In order to evaluate a failure behavior of pipe with notch-type wall thinning, the present study performed full-scale pipe tests using the 102mm, Schedule 80 pipe specimen simulated notch- and circular-type thinning defects. The pipe tests were conducted under the conditions of both monotonic and cyclic bending moment at a constant internal pressure of 10 MPa. From the results. of experiment the failure mode, load carrying capacity, deformation ability, and fatigue life of a notch-type wall thinned pipe were investigated, and they were compared with those of a circular-type wall thinned pipe. The failure mode of notched pipe was similar to that of circular-type thinned pipe under the monotonic bending load. Under the cyclic bending load, however, the mode was clearly distinguished with variation in the shape of wall thinning. The load carrying capacity of a pipe containing notch-type wall thinning was about the same or slightly lower than that of a pipe containing circular-type wall thinning when the thinning area was subjected to tensile stress, whereas it was higher than that of a pipe containing circular-type thinning defect when the thinning area was subjected to compressive stress. On the other hand, the deformation ability and fatigue life of a notch-type wall thinned pipe was lower than those of a circular-type wall thinned pipe.

• Journal of KIISE:Software and Applications
• /
• v.33 no.7
• /
• pp.605-614
• /
• 2006

The success of a software development project can be determined by the use of QCD. And as a software's size and complexity increase, the importance of early quality assurance rises. Therefore, more effort should be given to prevention, as opposed to correction. In order to provide a framework for the prevention of defects, defect detection activities such as peer review and testing, along with analysis of previous defects, is required. This entails a systematization and use of quality data from previous development efforts. FMEA, which is utilized for system safety assurance, can be applied as a means of software defect prevention. SW-FMEA (Software Failure Mode Effect Analysis) attempts to prevent defects by predicting likely defects. Presently, it has been applied to requirement analysis and design. SW-FMEA utilizes measured data from development activities, and can be used for defect prevention on both the development and management sides, for example, in planning, analysis, design, peer reviews, testing, risk management, and so forth. This research discusses about related methodology and proposes defect prevention model based on SW-FMEA. Proposed model is extended SW-FMEA that focuses on system analysis and design. The model not only supports verification and validation effectively, but is useful for reducing defect detection.