• Nuclear Engineering and Technology
• /
• 제54권9호
• /
• pp.3205-3214
• /
• 2022

The aim of this study is to provide a clear and accessible method to obtain accurate true-stress strain data, and to extend the limited material data beyond the ultimate tensile strength (UTS) for AISI 304L. AISI 304L is used for the outer construction for some types of nuclear transport packages, due to its post-yield ductility and high failure strain. Material data for AISI 304L beyond UTS is limited throughout literature. 3D digital image correlation (DIC) was used during a series of uniaxial tensile experiments. Direct method extracted data such as true strain and instantaneous cross-sectional area throughout testing such that the true stress-strain response of the material up to failure could be created. Post processing of the DIC data has a considerable effect on the accuracy of the true stress-strain data produced. Influence of subset size and smoothing of data was investigated by using finite element analysis to inverse model the force displacement response in order to determine the true stress strain curve. The FE force displacement response was iteratively adapted, using subset size and smoothing of the DIC data. Results were validated by matching the force displacement response for the FE model and the experimental force displacement curve.

• Nuclear Engineering and Technology
• /
• 제55권2호
• /
• pp.506-515
• /
• 2023

This study is to find a method to learn engine sound after the start-up of a diesel generator installed in nuclear power plant with an unsupervised deep learning algorithm (CNN autoencoder) and a new method to predict the failure of a diesel generator using it. In order to learn the sound of a diesel generator with a deep learning algorithm, sound data recorded before and after the start-up of two diesel generators was used. The sound data of 20 min and 2 h were cut into 7 s, and the split sound was converted into a spectrogram image. 1200 and 7200 spectrogram images were created from sound data of 20 min and 2 h, respectively. Using two different deep learning algorithms (CNN autoencoder and binary classification), it was investigated whether the diesel generator post-start sounds were learned as normal. It was possible to accurately determine the post-start sounds as normal and the pre-start sounds as abnormal. It was also confirmed that the deep learning algorithm could detect the virtual abnormal sounds created by mixing the unusual sounds with the post-start sounds. This study showed that the unsupervised anomaly detection algorithm has a good accuracy increased about 3% with comparing to the binary classification algorithm.

• Nuclear Engineering and Technology
• /
• 제55권2호
• /
• pp.530-545
• /
• 2023

As an alternative to conventional management options for a lot of concrete waste from decommissioning of nuclear power plants, a set of scenarios for controlled recycling of decommissioning concrete waste as engineered barriers of a radioactive waste repository was proposed, and a comprehensive safety assessment model and framework covering both pre-and post-closure phases was newly developed. The new methodology was applied to a reference vault-type repository, and the ratios of derived concentration limits to unconditional clearance levels of eighteen radionuclides for controlled recycling were provided for three sets of dose criteria (0.01, 1, and 20 mSv/y for the pre-closure and 0.01 mSv/y for the post-closure phases). It turns out that decommissioning concrete waste whose concentration is much higher than the unconditional clearance level can be recycled even when the dose criterion 0.01 mSv/y is applied. Moreover, a case study on ABWR bio-shield shows that the fraction of recyclable concrete waste increases significantly by increasing the dose criterion for the radiation worker in the pre-closure phase or the duration of storage prior to recycling. The results of this study are expected to contribute to demonstrating the feasibility of controlled recycling of a lot of decommissioning concrete waste within nuclear sectors.

• Nuclear Engineering and Technology
• /
• 제55권1호
• /
• pp.119-130
• /
• 2023

Among the various elements of probabilistic safety assessment (PSA), human failure events (HFEs) and their dependencies are major contributors to the quantification of risk of a nuclear power plant. Currently, the dependency among HFEs is reflected using a post-processing method in PSA, wherein several drawbacks, such as limited propagation of minimal cutsets through the fault tree and improper truncation of minimal cutsets exist. In this paper, we propose a method to model the HFE dependency directly in a fault tree using the if-then-else logic. The proposed method proved to be equivalent to the conventional post-processing method while addressing the drawbacks of the latter. We also developed a software tool to facilitate the implementation of the proposed method considering the need for modeling the dependency between multiple HFEs. We applied the proposed method to a specific case to demonstrate the drawbacks of the conventional post-processing method and the advantages of the proposed method. When applied appropriately under specific conditions, the direct fault-tree modeling of HFE dependency enhances the accuracy of the risk quantification and facilitates the analysis of minimal cutsets.

• 한국항행학회논문지
• /
• 제12권4호
• /
• pp.311-316
• /
• 2008

단일 모드 광섬유(SMF sigle mode fiber)로 구성된 광전송 링크에서 축적되는 색 분산과 비선형 효과를 효과적으로 보상하는 광전송 링크 구성 방법을 제안하였다. 제안된 광전송 링크는 전체 전송로 중간에 광 위상 공액기 (optical phase conjugator)를 두고 SMF의 각 광 중계 간격에서 축적된 분산을 분산 보상 광섬유 (DCF dispersion compensating fiber)를 통해 보상하는 inline 분산 제어 (DM; dispersion management)가 적용된 구조이다. 광전송 링크의 전체 잉여 분산 (NRD; net residual dispersion)은 송신기 다음에 있는 DCF에 의해 결정되는 선치 보상 (precompensation)과 수신기 전에 있는 DCF에 의해 결정되는 후치 보상 (postcompensation)을 통해 조절하도록 설계하였다. 선치 보상이나 후치 보상 중 하나를 이용해 NRD을 결정하는 구조보다 선치 보상과 후치 보상을 동시에 변화시키면서 분산 맵의 구조가 OPC를 중심으로 대칭이 되는 링크 구조가 더욱 효과적이고 융통적이라는 것을 확인하였다.

• 소성∙가공
• /
• 제17권2호
• /
• pp.124-129
• /
• 2008

Net shape forging technologies give many effects into the costs and qualities for the finished products. So, the studies to reduce the additional machining amount are very important in forging industry. Specially, there are two main topics in cold forging industry, such as, tool life and precision forging. In this study, new forging technique was proposed to eliminate the machining process for fixing up the length and improve the lead accuracy of gear. The luck-up hub is manufactured through many processes, such as upsetting, piercing and direct extrusion. The gear is formed in direct extrusion process; however, lead accuracy of the gear is over allowance limit. Therefore, the additional sizing process must be added. In this study, process design for closed-die forging of a lock-up hub used for a component of automobile transmission was made using three-dimensional finite element simulations, and the strain distributions and velocity distributions are investigated through the post processor. The rigid-plastic finite-element method for back pressure forging has been used in order to reduce development time and die cost. Using the FEM simulation, we found the optimum value of back pressure. The prototypes of lock-up hub parts were forged into the net-shape. In the experiment, lead precision of tooth are measured by the CCMM(Contact Coordinate Measuring Machine). The dimensional accuracy of forged part was improved up to the 40% when back press was applied.

• 한국통신학회논문지
• /
• 제36권12B호
• /
• pp.1433-1441
• /
• 2011

단일 모드 광섬유 (SMF; single mode fiber)의 색 분산과 자기 위상 변조 (SPM; self phase modulation)에 의한 신호 왜곡을 보상하기 위하여 분산 제어 (DM; dispersion management)와 광 위상 공액 (OPC; optical phase conjugation)이 적용된 파장 분할 다중 (WDM; wavelength division multiplexed) 시스템에서 전송 링크의 설계 기준을 살펴보았다. 본 연구에서 고려한 전송 링크의 설계 기준은 precompensation과 postcompensation에 의해 결정되는 최적 전체 잉여 분산 (NRD; net residual dispersion)과 채널의 최적 입사 전력 범위이다. 최적 NRD는 WDM 시스템의 전체 전송 거리와 중계 간격에 상관없이 precompensation과 postcompensation에 따라 각각 +10 ps/mn와 -10 ps/mn로 얻어졌다. 그러나 전송 거리와 중계 간격이 짧을수록 시스템 성능이 전반적으로 더욱 양호하게 나타나고, 비교적 넓은 유효 입사 전력 범위에 대해 유효 NRD 범위가 넓게 분포한다는 것을 알 수 있었다.

• 한국과학교육학회지
• /
• 제41권3호
• /
• pp.221-235
• /
• 2021

본 연구는 초등학교 5~6 학년 학생을 대상으로 과학 학습 상황에서 과학상태호기심(Science State Curiosity; SSC) 및 과학상태불안(Science State Anxeity; SSA) 수준에 따른 학습효과를 구명하는데 목적이 있다. 이를 위해 과학 학습을 과학 문제 대면(I), 결과 확인(II), 과학 개념 학습(III)의 세 가지 상황으로 구분하여 각 학습 상황에서 SSC 및 SSA를 측정하여 분석하였다. SSC 및 SSA가 학습효과에 미치는 순 영향을 파악하기 위해 학습효과에 영향을 줄 것으로 예상되는 과학호기심, 인지욕구, 과학자아개념, 과학불안, 흥미를 통제하였다. 과학 문제 대면 상황에서의 SSC 및 SSA를 'SSCI' 및 'SSAI'으로, 결과 확인 상황에서의 SSC 및 SSA를 'SSCII' 및 'SSAII'로, 과학 개념 학습 상황에서의 SSC 및 SSA를 'SSCIII' 및 'SSAIII'로 정의하였다. 그리고 학습효과를 직후학습효과와 지연학습효과로 구분하여 사전검사 점수에 비하여 직후검사 또는 지연검사 점수가 향상된 정도를 산출하여 분석하였다. 분석결과, 직후학습효과는 SSCI·SSCII가 높을수록, SSAI·SSAII·SSAIII가 낮을수록 높았으며, 지연학습효과는 SSCI·SSCII가 높을수록, SSAIII가 낮을수록 높았다. SSC가 SSA보다 학습효과에 미치는 영향이 컸으며, 직후학습효과에는 SSCII가, 지연학습효과에는 SSCI이 가장 많은 영향을 미쳤다. 또한 SSCIII가 높을수록 추가적인 자발적 학습을 하는 경향이 나타났다. 본 연구의 결과는 과학 학습에서 학생의 정서적 상태에 대한 이해의 폭을 넓히고, 상태호기심 및 상태불안 연구에 대한 이론적 토대를 마련할 것으로 기대된다.

• Korean Journal of Radiology
• /
• 제23권10호
• /
• pp.949-958
• /
• 2022

Objective: To investigate the feasibility of using a deep learning-based analysis of auscultation data to predict significant stenosis of arteriovenous fistulas (AVF) in patients undergoing hemodialysis requiring percutaneous transluminal angioplasty (PTA). Materials and Methods: Forty patients (24 male and 16 female; median age, 62.5 years) with dysfunctional native AVF were prospectively recruited. Digital sounds from the AVF shunt were recorded using a wireless electronic stethoscope before (pre-PTA) and after PTA (post-PTA), and the audio files were subsequently converted to mel spectrograms, which were used to construct various deep convolutional neural network (DCNN) models (DenseNet201, EfficientNetB5, and ResNet50). The performance of these models for diagnosing ≥ 50% AVF stenosis was assessed and compared. The ground truth for the presence of ≥ 50% AVF stenosis was obtained using digital subtraction angiography. Gradient-weighted class activation mapping (Grad-CAM) was used to produce visual explanations for DCNN model decisions. Results: Eighty audio files were obtained from the 40 recruited patients and pooled for the study. Mel spectrograms of "pre-PTA" shunt sounds showed patterns corresponding to abnormal high-pitched bruits with systolic accentuation observed in patients with stenotic AVF. The ResNet50 and EfficientNetB5 models yielded an area under the receiver operating characteristic curve of 0.99 and 0.98, respectively, at optimized epochs for predicting ≥ 50% AVF stenosis. However, Grad-CAM heatmaps revealed that only ResNet50 highlighted areas relevant to AVF stenosis in the mel spectrogram. Conclusion: Mel spectrogram-based DCNN models, particularly ResNet50, successfully predicted the presence of significant AVF stenosis requiring PTA in this feasibility study and may potentially be used in AVF surveillance.

• 한국전자통신학회논문지
• /
• 제11권1호
• /
• pp.113-118
• /
• 2016

본 연구는 우편 접수창구에서 고객이 우편물을 보낼 때 마다 사용자의 불편한 문제점을 도출하고, 많은 사용자가 소유하고 있는 휴대전화 단말기에 저장되어 있는 주소 정보를 기반으로 별도의 보안 체계를 별도로 구축하지 않고 우편물의 주소 정보를 우체국 접수창구 단말기(포스 넷)로 주소 정보를 전송하는 방법을 제안하고자 한다. 본 제안 방법은 사용자의 정보를 포스 넷의 2차원 바코드(QR 코드)로 변환한 후 포스 넷에서 QR 코드를 판독하는 방법과 블루투스 등과 같은 근거리 통신망을 이용하여 다량의 주소 정보를 텍스트 화하여 포스 넷에 전송하는 방식이다. 본 제안 방식에 의하여 우편접수시스템과 연동할 경우 모바일용 우편 주소를 자동으로 생성하고 출력해 줄 뿐만 아니라 우편물 접수담당자의 업무를 개선하고, 편리에 따른 경제적 효과를 거둘 수 있었다.