• 한국근적외분광분석학회:학술대회논문집
• /
• 한국근적외분광분석학회 2001년도 NIR-2001
• /
• pp.1132-1132
• /
• 2001

This work is a further development of the method created by G. Krivoshiev in 1996 for elimination of peel interference and prediction of fruit flesh optical density. In this investigation, as it was earlier, the objects are observed as being structured by three successive layer “AlongrightarrowOlongrightarrowB” denoting “peel-flesh-peel”. In the first version of the method the transmittances of the surface layers A and B were measured according to Kubelka-Munk theory by means of their diffuse reflectance. At that the overall transmittance T was approximated in the form of a multiplication approximation being valid for plane-parallel layers of a non-scattering material. In this work this approximation was done away with applying the theory of discontinuum, respectively Benfor's equations. As a result two mathematical models were created for non-destructive prediction of fruit flesh optical density. These models are different from the ones based solely on Kubelka-Munk theory, the destruction being marked by the terms 1n (1 - $R_{A}R_{0}$) and 1n (1 - $R_{A}R_{B}$), where: $R_{A}$ and $R_{B}$ are reflectance values for the surface layers A and B; $R_{0}$ is the average reflectance of the internal layer that could be obtained empirically by means of a preliminary measurement of sufficiently large number of physically peeled fruits of a given species and variety.

• 대한조선학회논문집
• /
• 제34권1호
• /
• pp.122-129
• /
• 1997

외판곡가공 과정의 역학적해석에 사용되는 열탄소성 해석은 계산시간이 많이 소요된다. 선상가열작업의 최적공정을 구하기 위하여는 시뮬레이션과 유한요소해석 사이에 상당수의 반복계산이 필요하다. 그러므로 요구되는 천문학적인 계산시간으로 인하여 선상가열의 수치시뮬레이션은 많은 제약을 받고 있다. 따라서 선상가열의 시뮬레이션에 소요되는 역학해석 계산시간을 크게 줄이는 방안이 필요하다. 외판곡가공에서의 역학적 정보를 효율적으로 구하기 위하여, 본 논문에서는 먼저 선상 가열 작업에 영향을 끼치는 요소들을 살펴보았으며, 역전파 방식의 인공 신경망을 이용하여 일종의 추론기구를 구현하였다. 신경망은 은닉층의 갯수와 은닉층에 있는 뉴런의 갯수를 바꿔 주며 수치해석 결과들을 학습시켰다. 그 결과 두개의 은닉층을 가진 인공 신경망의 경우 각 은닉층에 뉴런의 갯수가 충분하다면 학습 예제들을 쉽게 학습하였고, 또 학습된 결과로부터 새로운 해를 도출해 낼 때 그 값은 실제 값과 비교해 볼 때 비교적 작은 오차를 보였다. 결과적으로 구하고자 하는 문제영역 근처에 충분한 학습 예제가 마련된다면 비교적 실제 값에 근사하는 결과를 보인다는 것을 알 수 있었다. 본 연구를 통하여, 수치해석결과의 학습방법을 기초로한 인공신경망은 곡면가공의 시뮬레이션단계에서 필요한 역학적 해석정보의 경제적인 산출에 적용가능함이 확인되었다.

• Bulletin of the Korean Chemical Society
• /
• 제35권2호
• /
• pp.353-356
• /
• 2014

We fabricated organic photovoltaic (OPV) based on ZnO ripple structure on indium tin oxide as electron-collecting layers and PTB7-F20 as donor polymer. In addition, atomic layer deposition (ALD) was used for preparing additional ZnO layers on rippled ZnO. Addition of 2 nm-thick ALD-ZnO resulted in enhanced initial OPV performance and stability. Based on photoluminescence results, we suggest that ALD-ZnO layers reduced number of surface defect sites on ZnO, which can act as electron-hole recombination center of OPV, and increased resistance of ZnO towards surface defect formation.

• 한국표면공학회지
• /
• 제41권1호
• /
• pp.16-22
• /
• 2008

Anodizing to form oxide layers on the pure titanium was performed in the electrolyte containing 1.5M $H_2SO_4$, 0.2M $H_3PO_4$, and 2.5wt.% $CuSO_4$ using the ac-biased arc anodizing technique. Titanium oxide layers anodized with different applied voltages, voltage-elevating rates, and anodizing times were investigated. In addition, thermal oxidation test under an atmospheric environment for the arc-anodized specimens was carried out. The thickness of oxide layers were not affected by the voltage-elevating rates, but increased slightly with the increase of anodizing times. The thickness of oxide layers were increased with the increase of voltages, and increased remarkably in the condition of 200V. The size and number of the pore observed in the center of the porous cell were decreased with increase of applied voltage. From the result of thermal oxidation test, it revealed that oxide layer formed by arc anodizing more effective to prevent oxidation of pure titanium.

• 한국표면공학회지
• /
• 제36권4호
• /
• pp.307-316
• /
• 2003

Salt spray, CASS(copper accelerated acetic salt spray) and EC(electrolytic corrosion) tests were performed in order to evaluate the corrosion resistance of plated Ni and Ni-Cr layers on Fe substrate. Compared with the conventional methods such as salt spray and CASS, the electrochemical method such as EC test may be beneficial in terms of test time span and quantitative accuracy. Furthermore, EC test can also become the alternative method to evaluate the resistance to corrosion of coatings by measuring the corrosion potentials of the coated layers in the electrolyte during the off-time of EC cycles. Compared with the corrosion potentials of pure iron, nickel, chromium, those potentials of coated layers can be used to anticipate the extent of corrosion. Results showed that in terms of the test time span, EC test gave 14 times and 21 times faster results than the salt spray test in cases of $5\mu\textrm{m}$ Ni and $20\mu\textrm{m}$ Ni plated layers, respectively. In addition, EC test also offered the shorter test time span than CASS test in cases of $5 \mu\textrm{m}$ Ni + $0.5\mu\textrm{m}$ Cr, and $20\mu\textrm{m}$ Ni + $0.5\mu\textrm{m}$ Cr on Fe substrate by 78 times and 182 times, respectively. Therefore, EC test can be regarded as the better method to evaluate the resistance to corrosion of coated layers than the conventional methods such as salt spray and CASS.

• 보존과학회지
• /
• 제26권3호
• /
• pp.247-258
• /
• 2010

무령왕릉 출토 목관재 6점의 칠도막 편에 대하여 광학 현미경 관찰, 적외선 분광분석 및 SEM-EDS 분석을 실시하였다. 목관재에 도장된 옻칠 기법은 크게 네 가지 방법으로 분류되었다. 목재 표면에 그을음(미립자의 흑색 안료)을 혼합한 흑색 안료층이 밑층으로 존재하면서 그 위로 칠이 1회(그룹 II) 도장된 것과 3회(그룹 I) 도장된 것, 흑색 안료층이 존재하지 않으면서 칠 도장이 1회(그룹 IV)와 2회(그룹 III) 되어있는 것으로 분류되었다. 이는 목관의 제작기법과 부재의 결구 방법을 밝히는데 좋은 정보가 될 것으로 생각된다. 적외선 분광분석을 통해 무령왕릉 칠도막은 자외선에 의한 열화 요인이 적은 것으로 나타났다. 이는 이동이 어려운 목관의 크기와 무덤이라는 매장 환경 등에 기인한 결과로 보인다. 또한 SEM-EDS 분석결과 현재까지 고대 칠분석에서 보고된 바 없는 Ca, Fe, Cu의 성분이 확인되어 의미가 있다.

• JSTS:Journal of Semiconductor Technology and Science
• /
• 제14권3호
• /
• pp.322-330
• /
• 2014

New effective techniques to repair "small" design errors in integrated circuits are presented. As semiconductor chip complexity increases and the design period becomes tight, errors frequently remain in a fabricated chip making revisions required. Full mask revision significantly increases the cost and time-to-market. However, since many "small" errors can be repaired by modifying several connections among the circuit blocks and spare cells, errors can frequently be repaired by revising metal layers. Metal only revision takes significantly less time and involves less cost when compared to full mask revision, since mask revision costs multi-million dollars while metal revision costs tens of thousand dollars. In our research, new techniques are developed to further reduce the number of metal layers to be revised. Specifically, we partition the circuit blocks with higher error probabilities and extend the terminals of the signals crossing the partition boundaries to the preselected metal repair layers. Our partitioning and pin extension to repair layers can significantly improve the repairability by revising only the metal repair layers. Since pin extension may increase delay slightly, this method can be used for non-timing-critical parts of circuits. Experimental results by using academia and industrial circuits show that the revision of the two metal layers can repair many "small" errors at low-cost and with short revision time. On the average, when 11.64% of the spare cell area and 24.72% of the extended pins are added to the original circuits, 83.74% of the single errors (and 72.22% of the double errors) can be corrected by using two metal revision. We also suggest methods to use our repair techniques with normal commercial vender tools.

• Journal of Advanced Marine Engineering and Technology
• /
• 제38권8호
• /
• pp.981-994
• /
• 2014

Density currents formed by buoyancy discharges from rivers are numerically studied using non-dimensional two layer model including Coriolis acceleration, bottom stress, interfacial friction. Some typical numbers such as Froude number, densimetric Froude number and Kelvin number are obtained and some characteristic scales are defined as a result of non-dimensionalization of the governing equations. Besides the Coriolis effect, the configurations of bottom topography, bottom friction coefficient and interfacial friction are found to significantly affect the propagation of the warm water plume. Frontal position can fastly propagate in the case of large density difference between the two layers and small interfacial friction. Left side boundary current is easily formed under the small interfacial friction. With large Kelvin number, both right and left side boundary currents are formed. Wave-like disturbances and eddies are easily formed under the high Froude number.

• 태양에너지
• /
• 제5권2호
• /
• pp.11-19
• /
• 1985

In this paper, the interface stability not to occur mixing and entrainment between the adjacent layers has been studied in the case of the selective withdrawal of a stratum and the injection in stratified fluid formed by the density difference in a small solar pond. There are stability parameter, Richardson number, Rayleigh number and Froude number as the parameters governing stability in order to measure the interface stability on the stratified fluid. The model which could measure the interface stability on the stratified fluid was the small solar pond composed by 1 meters wide, 2 meters high, and 5 meters long. In order to measure the interface stability on the stratified fluid at the inlet port, the middle section and the outlet port, Richardson number, Rayleigh number, and Froude number involved in the parameters governing the stability were calculated by means of the data resulted from the test of the study on hydrodynamic stability between the convective and nonconvective layers in that solar pond. Richardson number written by the ratio of inertia force to buoyancy force can be used in order to measure the stability on the stratified fluid related to the buoyancy force generated from the injection of fluid. Rayleigh number written by the product of Grashof number by Prandtl number can be used in order to measure the stability of the fluid related to the heat flux and diffusivity of viscosity. Froude number written by the ratio of gravity force to inertia force can be used in order to measure the stability of the nonhomogeneous fluid related to the density difference. As the result of calculating the parameters governing stability, the interface stability on the stratified fluid couldn't be identified below the 70cm height from the bottom of the solar pond, but it could be identified above the 70cm height from it at the inlet port, the middle section and the outlet port. When compared with such the three parameters as Richardson number, Rayleigh number, Froude number, the calculated result was in accord with them at inlet port, the middle section and the outlet port. Henceforth, it is learned that even though any of the three parameters is used for the purpose of measuring the interface stability on the stratified fluid, the result will be the same with them. It is concluded that all the use of Richardson number, Rayleigh number, and Froude number, is desirable and infallible to measure the interface stability on the stratified fluid in the case of considering the exist of the fluid flow and the heat flux like the model of the solar pond.

• 한국컴퓨터정보학회논문지
• /
• 제12권5호
• /
• pp.313-319
• /
• 2007

본 논문에서는 획득된 차량 영상에서 수직 에지의 특징을 이용하여 번호판 영역과 개별 코드를 추출하고, 추출된 개별 코드는 퍼지 신경망 알고리즘을 이용하여 인식한다. 차량 번호판 영역을 검출하기 위해 프리윗 마스크에 의해 수직 에지를 찾고, 차량 번호판에 관한 특성 정보를 이용하여 잡음을 제거한 추에 차량 번호판 영역과 개별 코드를 추출한다 추출된 개별 코드를 인식하기 위해 퍼지 신경망 알고리즘을 제안하고 인식에 적용한다. 제안된 퍼지 신경망은 입력층과 중간층간의 학습 구조로는 FCM 알고리즘을 적용하고, 중간층과 출력층간의 학습 구조에는 Max_Min 신경망을 적용한다. 제안된 방법의 추출 및 인식 성능을 평가하기 위하여 실제 차량 영상 150장을 대상으로 실험한 결과, 기존의 차량 번호판 인식 방법보다 효율적이고 인식 성능이 개선된 것을 확인하였다.