• Proceedings of the Korea Water Resources Association Conference
• /
• 2017.05a
• /
• pp.216-216
• /
• 2017

제주지역의 하천은 일정강우 이상 발생 시 하천 표층이 포화된 후 갑작스러운 유출이 발생되며, 강우에 의해 점진적으로 하천유량이 증가하는 내륙지역 하천과는 다른 유출특성을 보인다. 그러나 제주지역 하천특성에 관한 연구는 최근부터 진행되어 현재까지도 내륙지역의 하천계획의 방법을 적용되고 있다. 이 연구에서는 제주도 한천유역을 대상으로 국내에서 적용되는 설계홍수량 산정방법을 적용하여 문제점을 도출하고, 지형 및 지역특성을 분석하여 설계홍수량 산정 시 매개변수의 적용방안을 개선하였다. 홍수량 검증을 위해 보유하고 있는 현장관측 자료 중 가장 큰 첨두유량이 발생한 3개의 사상을 선정하여 비교하였다. 기존 홍수량 산정방법을 이용하여 한천유역의 유출량을 산정한 결과 78.7 ~ 317.8 cms, 14.3 ~ 37.5%의 오차율로 관측자료 대비 큰 오차를 범하는 것으로 검토되었다. 한천유역의 상류지역은 45%로 매우 급한 경사형태를 보이고, CN산정 시 경사보정을 통해 홍수량을 산정한 결과 기존의 유출특성 대비 1.47 ~ 6.45% 개선되었다. 관측자료 기반 강우-유출의 최적화 기법을 통해 산정된 도달시간을 적용하여 홍수량을 산정한 결과 4.39 ~ 16.67% 개선되었으며, 한천유역을 소유역으로 구분하여 홍수량을 산정한 결과 9.92 ~32.96% 개선된 결과가 도출되었다. 이러한 결과는 지역별 함양특성과 급한 경사 특성, 토지이용에 따른 유역특성이 적용되고, 지역별 초기손실이 반영되어 홍수유출곡선이 실제 관측자료와 유사하게 나타나는 것으로 기존방법의 경우 전체 유역에 평균값이 일괄적으로 적용되기 때문에 실제관측자료와 큰 오차를 나타내는 것으로 분석되었다.

• Journal of Digital Convergence
• /
• v.17 no.10
• /
• pp.251-258
• /
• 2019

In the future, various products are created in various fields using artificial intelligence. In this age, it is a very important problem to know the operation principle of artificial intelligence learning method and to use it correctly. This paper introduces artificial intelligence learning methods that have been known so far. Learning of artificial intelligence is based on the fixed point iteration method of mathematics. The GD(Gradient Descent) method, which adjusts the convergence speed based on the fixed point iteration method, the Momentum method to summate the amount of gradient, and finally, the Adam method that mixed these methods. This paper describes the advantages and disadvantages of each method. In particularly, the Adam method having adaptivity controls learning ability of machine learning. And we analyze how these methods affect digital signals. The changes in the learning process of digital signals are the basis of accurate application and accurate judgment in the future work and research using artificial intelligence.

• Korean Journal of Optics and Photonics
• /
• v.14 no.3
• /
• pp.306-311
• /
• 2003

In this paper, we modeled the incident beam in order to analyze and evaluate the optical thin film device for wavelength division multiplexing in optical telecommunication network. As applied ray tracing method to the optical path, we were compared the accuracy of coupling efficiency simulated by two modeling methods. In the results of sinulation, ceil modeling method was preferred to annual modeling method in micro-optic device because of accuracy for coupling efficiency and Gaussian intensity distribution. In the results of optimal simulation for optical device using thin film filter, the distance (d1) between optical fiber and GRIN lens, the distance (d2) between GRIN lens and thin film filter and the coupling efficiency were 0.24 mm, 0.25 mm and -0.11 ㏈ respectively. As d2 was displaced at 0.25 mm and d1 was varied in order to evaluate the optimal value, d1 and maximum coupling efficiency were 0.24 mm and -0.35㏈, respectively. Then the results of experiment were corresponded to that of optimal simulation by cell modeling and it was possible to analyze the performance for optical device using thin film filter by the simulation.

• Journal of Broadcast Engineering
• /
• v.26 no.2
• /
• pp.167-174
• /
• 2021

Recently, deep learning technology is widely used in various computer vision applications, such as object recognition, classification, and image generation. In particular, the deep learning-based super-resolution has been gaining significant performance improvement. Fast super-resolution convolutional neural network (FSRCNN) is a well-known model as a deep learning-based super-resolution algorithm that output image is generated by a deconvolutional layer. In this paper, we propose an FPGA-based convolutional neural networks accelerator that considers parallel computing efficiency. In addition, the proposed method proposes Optimal-FSRCNN, which is modified the structure of FSRCNN. The number of multipliers is compressed by 3.47 times compared to FSRCNN. Moreover, PSNR has similar performance to FSRCNN. We developed a real-time image processing technology that implements on FPGA.

• Korean Journal of Fisheries and Aquatic Sciences
• /
• v.52 no.4
• /
• pp.311-320
• /
• 2019

This study aimed to optimize mixing conditions (adding amount of squid skin and sea tangle Saccharina japonica) for concentrates of octopus Octopus vulgaris cooking effluent (COCE) and roasting conditions (temperature and time) of seasoned Laver Pyropia yezoensis with concentrates of octopus cooking effluent (SL-COCE) using response surface methodology (RSM). The results of RSM program for COCE showed that the optimum independent variables ($X_1$, squid skin amount; $X_2$, sea tangle amount) based on the dependent variables ($Y_1$, odor intensity; $Y_2$, amino-N content; $Y_3$, sensory overall acceptance) for high-quality COCE were 0.53% (w/w) for $X_1$ and 0.48% (w/w) for $X_2$ for uncoded values. The results of the RSM program for SL-COCE showed that the optimum independent variables ($X_1$, roasted temp.; $X_2$, roasted time) based on the dependent variables ($Y_1$, burnt odor intensity; $Y_2$, water activity; $Y_3$, sensory overall acceptance) for high-quality SL-COCE were $344^{\circ}C$ for $X_1$ and 8 sec for $X_2$ for uncoded values. The SL-COCE prepared under optimum procedure was superior in sensory overall acceptance to commercial seasoned laver.

• Journal of the Korean GEO-environmental Society
• /
• v.20 no.8
• /
• pp.13-20
• /
• 2019

In order to ground reinforcement, the chemical grouting, the anchor, the soil nailing system, the micropile, etc. can be mentioned by the methods widely used in domestic. The above ground reinforcement methods are developed by various methods depending on the type of reinforcement, installation method, presence of prestress, grouting method, etc. However, in common, the strength of reinforcement, the friction force of grout and reinforcement and the friction force of grout and ground are the main design variables. Therefore, the optimized ground reinforcement is a material with a high tensile strength of the reinforcement itself, the friction force between the reinforcement and the grout is high, and the application of an optimal grouting method is necessary to improve the friction force between the grout and the ground. In this study, a total of 20 model tests were conducted to analyze the reinforcement effects according to the shape of the reinforcement and the grouting method. As a result of the test, As a result of the experiment, it is judged that the reinforcing effect is superior to the perforated + wing type reinforcement and post grouting method.

• Journal of the Korean Society of Propulsion Engineers
• /
• v.25 no.3
• /
• pp.56-61
• /
• 2021

In this study, Al₂O₃ and H-ZSM-5 were coated on the inner wall of the stainless steel tube for the stable use of liquid hydrocarbon fuel and an endothermic catalyst used as coolant for hypersonic flying vehicles. Coke production is inevitable by the endothermic decomposition reaction of the liquid hydrocarbon fuel, and Fe, Ni metals induce the production of the filamentous coke by using a stainless steel tube reactor as a cooling channel. By coating the stainless steel with H-ZSM-5, Fe and Ni metals are prevented from being directly exposed to the liquid hydrocarbon fuel, and the formation of the filamentous coke is inhibited. In addition, Al₂O₃ is coated between the stainless steel and H-ZSM-5 to enhance adhesion bond strength.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.22 no.2
• /
• pp.20-25
• /
• 2021

The target-object classification method was implemented using a deep-learning-based detection model in real-time images. The object detection model was a deep-learning-based detection model that allowed extensive data collection and machine learning processes to classify similar target-objects. The recognition model was implemented by changing the processing structure of the detection model and combining developed the vision-processing module. To classify the target-objects, the identity and similarity were defined and applied to the detection model. The use of the recognition model in industry was also considered by verifying the effectiveness of the recognition model using the real-time images of an actual soccer game. The detection model and the newly constructed recognition model were compared and verified using real-time images. Furthermore, research was conducted to optimize the recognition model in a real-time environment.

• Composites Research
• /
• v.34 no.1
• /
• pp.47-50
• /
• 2021

In this paper, the Curing process for Epoxy Molding Compound (EMC) Package was developed by comparing the performance of the EMC/Cu Bi-layer package manufactured by the conventional Hot Press process system and Carbon Nanotubes (CNT) Heater process system of the surface heating system. The viscosity of EMC was measured by using a rheometer for the curing cycle of the CNT Heater. In the EMC/Cu Bi-layer Package manufactured through the two process methods by mentioned above, the voids inside the EMC was analyzed using an optical microscope. In addition, the interfacial void and warpage of the EMC/Cu Bi-layer Package were analyzed through C-Scanning Acoustic Microscope and 3D-Digital Image Correlation. According to these experimental results, it was confirmed that there was neither void in the EMC interior nor difference in the warpage at room temperature, the zero-warpage temperature and the change in warpage.

• Journal of the Korean Crystal Growth and Crystal Technology
• /
• v.30 no.6
• /
• pp.258-263
• /
• 2020

Silicon carbide (SiC) fibers mainly fabricated from polycarbosilane, a ceramic precursor, are applied as reinforcing materials for ceramic matrix composites (CMCs) because of their high temperature oxidation resistance, tensile strength, and light weight. In this study, continuous SiC fibers used to fabricate rope-type flexible heating elements capable of generating high-temperature heat (> 650℃). For high-efficiency heating elements, the resistance of SiC fiber rope was measured by 2-point probe method according to the cross-sectional area and length. In addition, the fabrication conditions of rope-type SiC fiber heating elements were optimized by controlling the oxygen impurities and the size of crystal grains present in the amorphous SiC fiber. As a result, the SiC fiber heating element having a resistance range of about 100~200 Ω exhibited an excellent power consumption efficiency of 1.5 times compared to that of the carbon fiber heating element.