• 비파괴검사학회지
• /
• 제27권5호
• /
• pp.383-392
• /
• 2007

본 논문에서는 신경회로망을 이용하여 원자력 발전소 증기발생기 세관의 결함 깊이와 폭을 예측하는 연구를 수행한다. 결함 크기 추정을 위하여 우선, I-In 형태, I-Out 형태, V-In 형태, V-Out 형태의 4가지 결함형상에 대한 와전류탐상시험(ECT) 신호를 생성한다. 특히, 유한요소법에 기반한 수치해석 기법을 이용하여 여러 가지 폭과 깊이를 갖는 결함 400개의 ECT 신호를 생성한다. 이와 같이 생성된 ECT 신호로부터, 결함 크기와 폭을 예측하기 위한 새로운 특징벡터를 추출하는데, 이 특징벡터에는 최대 임피던스 값을 갖는 점과 최대 임피던스값의 1/2의 값을 갖는 점 사이의 위상각이 포함된다. 추출된 특징벡터를 이용하여 결함의 크기를 예측하기 위해서 하나의 은닉층을 갖는 다층퍼셉트론을 이용하였다. 컴퓨터 모의실험 연구를 통하여 제안된 방법이 우수한 예측성능을 갖는다는 것을 보였다.

• Structural Engineering and Mechanics
• /
• 제63권4호
• /
• pp.429-438
• /
• 2017

Structural design has an imperative role in deciding the failure possibility of a Reinforced Concrete (RC) structure. Recent research works achieved the goal of predicting the structural failure of the RC structure with the assistance of machine learning techniques. Previously, the Artificial Neural Network (ANN) has been trained supported by Particle Swarm Optimization (PSO) to classify RC structures with reasonable accuracy. Though, keeping in mind the sensitivity in predicting the structural failure, more accurate models are still absent in the context of Machine Learning. Since the efficiency of multi-objective optimization over single objective optimization techniques is well established. Thus, the motivation of the current work is to employ a Multi-objective Genetic Algorithm (MOGA) to train the Neural Network (NN) based model. In the present work, the NN has been trained with MOGA to minimize the Root Mean Squared Error (RMSE) and Maximum Error (ME) toward optimizing the weight vector of the NN. The model has been tested by using a dataset consisting of 150 RC structure buildings. The proposed NN-MOGA based model has been compared with Multi-layer perceptron-feed-forward network (MLP-FFN) and NN-PSO based models in terms of several performance metrics. Experimental results suggested that the NN-MOGA has outperformed other existing well known classifiers with a reasonable improvement over them. Meanwhile, the proposed NN-MOGA achieved the superior accuracy of 93.33% and F-measure of 94.44%, which is superior to the other classifiers in the present study.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제27권4호
• /
• pp.422-431
• /
• 2000

본 논문에서는 주성분 해석 기법에 기반한 새로운 벡터 양자화 코드북 설계 방법을 제안한다. 주성분 해석 알고리즘은 입력 영상벡터를 더 작은 차원의 특징 벡터로 변환시키는데 사용되며, 변환된 영역에서 특징 벡터의 군집을 최적으로 결정된 분할 초평면을 이용하여 두 군집으로 분할하는 과정을 반복 함으로써 코드북을 생성한다. 본 논문에서는 연산 시간이 오래 걸리는 최적 분할 초평면 탐색을 (1) 분할 초평면은 특징 벡터의 주축에 수직이며, (2) 좌우측 부군집의 오차의 균형점과 일치하며, (3) 좌우측 부군집의 오차를 점진적으로 조정함으로서 연산 수행 시간을 크게 단축시켰다. 제안한 주축 연속 분할은 분할전후의 오차의 감축이 가장 큰 군집에 대해, 전체 군집의 오차가 설정한 수준보다 작을 때까지 연속적으로 수행된다. 실험 결과 제안한 주성분 해석 기반 벡터 양자화 방법은 SOFM을 이용한 방법보다 수행시간이 빠르며 K-mean 알고리즘을 이용한 방법보다 복원 성능이 뛰어남을 볼 수 있다.

• Asian Pacific Journal of Cancer Prevention
• /
• 제15권22호
• /
• pp.9885-9889
• /
• 2014

Background: Tumor necrosis factor (TNF) related apoptosis-inducing ligand (TRAIL) is an antitumor candidate in cancer therapy. This study focused on effects of TRAIL, as a proapototic ligand that causes apoptosis, in B-CELL chronic lymphocytic leukemia cells (B-CLL). Materials and Methods: A population of HEK 293 cells was transducted by lentivirus that these achieved ability for producing the TRAIL protein and then HEK 293 cells transducted were placed in the vicinity of CLL cells. After 24 hours of co-culture, apoptosis of CLL cells was assessed by annexin V staining. Results: The amount of Apoptosis was examined separately in four groups: 293 HEK TRAIL ($16.17{\pm}1.04%$); 293 HEK GFP ($2.7{\pm}0.57%$); WT 293 HEK ($2{\pm}2.6%$); and CLL cells ($0.01{\pm}0.01%$). Among the groups studied, the maximum amount of apoptosis was in the group that the vector encoding TRAIL was transducted. In this group, the mean level of soluble TRAIL in the culture medium was 253pg/ml; also flow cytometry analyzes showed that proapotosis in this group was $32.8{\pm}1.6%$, which was higher than the other groups. Conclusions: In this study, we have demonstrated that TNF secreted from HEK 293 cells are effective in death of CLL cells.

• Nuclear Engineering and Technology
• /
• 제52권7호
• /
• pp.1429-1435
• /
• 2020

Applying an accurate parametric prediction model to identify abnormal or false pressurizer water levels (PWLs) is critical to the safe operation of marine pressurized water reactors (PWRs). Recently, deep-learning-based models have proved to be a powerful feature extractor to perform high-accuracy prediction. However, the effectiveness of models still suffers from two issues in PWL prediction: the correlations shifting over time between PWL and other feature parameters, and the example imbalance between fluctuation examples (minority) and stable examples (majority). To address these problems, we propose a cost-sensitive mechanism to facilitate the model to learn the feature representation of later examples and fluctuation examples. By weighting the standard mean square error loss with a cost-sensitive factor, we develop a Cost-Sensitive Long Short-Term Memory (CSLSTM) model to predict the PWL of PWRs. The overall performance of the CSLSTM is assessed by a variety of evaluation metrics with the experimental data collected from a marine PWR simulator. The comparisons with the Long Short-Term Memory (LSTM) model and the Support Vector Regression (SVR) model demonstrate the effectiveness of the CSLSTM.

• 설비공학논문집
• /
• 제27권6호
• /
• pp.283-292
• /
• 2015

A model of apartment heating load in a district heating system could be useful in the management and utilization of energy resources, since it could predict energy usage and so could assist in the efficient use of energy resources. The heating load in a district heating system varies in a highly nonlinear manner and is subject to many different factors, such as heating area, number of people living in that complex, and ambient temperature. Thus there are few published papers with accurate models of heating load, especially in domestic literature. This work is concerned with the modeling of apartment heating load in a district heating system in winter, using the reduced least square support vector machine (LS-SVM), and with the purpose of using the model to predict heating energy usage in domestic city area. We collected 23,856 pieces of data on heating energy usage over a 12-week period in winter, from 12 heat exchangers in five apartments. Half of the collected data were used to construct the heating load model, and the other half were used to test the model's accuracy. The model was able to predict the heating energy usage pattern rather accurately. It could also estimate the usage of heating energy within of mean absolute percentage error. This implies that the model prediction accuracy needs to be improved further, but it still could be considered as an acceptable model if we consider the nonlinearity and uncertainty of apartment heating energy usage in a district heating system.

• 대한기계학회논문집B
• /
• 제20권5호
• /
• pp.1649-1660
• /
• 1996

The flow around a circular cylinder was controlled by an acoustic excitation issued from a thin slit along the cylinder axis. The static pressure distributions around the cylinder wall and flow characteristics in the near wake have been measured. Experiments were performed under three cases of Reynolds number, 7.8 * 10$\^$4/, 2.3 * 10$\^$5/ and 3.8 * 10$\^$5/. The effects of excitation frequency, sound pressure level and the location of the slit were examined. Data indicate that the excitation frequency and the slit location are the key parameters for controlling the separated flow. At Re$\_$d/, = 7.8 * 10$\^$4/, the drag is reduced and the lift is generated to upward direction, however, at Re$\_$d/, =2.3 * 10$\^$5/ and 3.8 * 10$\_$5/, the drag is increased and lift is generated to downward direction inversely. It is thought that the lift switching phenomenon is due to the different separation point of upper surface and lower surface on circular cylinder with respect to the flow regime which depends on the Reynolds number. Vortex shedding frequencies are different at upper side and lower side. Time-averaged velocity field shows that mean velocity vector and the points of maximum intensities are inclined to downward direction at Re$\_$d/ = 7.8 * 10$\^$4/, but are inclined to upward direction at Re$\_$d/ = 2.3 * 10$\^$5/.

• 대한자기공명의과학회:학술대회논문집
• /
• 대한자기공명의과학회 2001년도 제6차 학술대회 초록집
• /
• pp.111-111
• /
• 2001

목적： To evaluate the feasibility of the event-related functional MR study using power grip studying the hand motor system 대상 및 방법： Event-related functional MRI was performed on a 1.5T MR unit in seven norm volunteers (man=7, right-handedness=2, left-handedness=5, mean age： 25 years). A single-shot GRE-EPI sequence (TR/TE/flip angle： 1000ms/40ms/90, FOV = 240 mm matrix= 64$\times$64, slice thickness/gap = 5mm/0mm, 7 true axial slices) was used for functiona MR images. A flow-sensitive conventional gradient echo sequence (TR/TE/flip angl 50ms/4ms/60) was used for high-resolution anatomical images. To minimize the gross hea motion, neck-holders (MJ-200, USA) were used. A series of MR images were obtained in axial planes covering motor areas. To exclude motion-corrupted images, all MR images wer surveyed in a movie procedure and evaluated using the estimation of center of mass of ima signal intensities. Power grip task consisted of the powerful grip of all right fingers and hand movement ta used very fast right finger tapping at a speed of 3 per 1 second. All tasks were visual-guid by LCD projector (SHARP, Japan). Two tasks consisted of 134 phases including 7 activatio and 8 rest periods. Active stimulations were performed during 2 seconds and rest period were 15 seconds and total scan time per one task was 2 min 14 sec. Statistical maps we obtained using cross-correlation method. Reference vector was time-shifted by 4 seconds an Gaussian convolution with a FWHM of 4 seconds was applied to it. The threshold in p val for the activation sites was set to be 0.001. All mapping procedures were peformed usin homemade program an IDL (Research Systems Inc., USA) platform. We evaluated the activation patterns of the motor system of power grip compared to hand movement in t event-related functional MRI.

• 한국음향학회지
• /
• 제23권5호
• /
• pp.400-409
• /
• 2004

본 논문에서는 임의의 화자가 발성한 음성을 다른 화자가 발성한 음성처럼 들리도록 변환하는 음성 변환 알고리즘을 제안하였다. 개인이 지니고 있는 음성의 특성을 변환하기 위해 성도 전달 함수의 특성을 변환 변수로 사용하였으며, 기존의 기법과 비교하여 목표 화자의 음성과 주관적, 객관적으로 더욱 유사한 변환음을 얻기 위한 새로운 방법을 제안하였다. 성도 전달 함수의 변환은 전체 특징 벡터 공간을 분류 한 뒤, 각 구획에 대한 선형 변환식을 통해 구현된다. 특징 변수로서 LPC 켑스트럼을 사용하였으며, 벡터 공간의 분류와 선형 변환식의 추정을 동시에 최적화시키는 분류-변환 알고리즘이 새로이 제안되었다. 제안된 음성 변환 기법의 성능을 평가하기 위해 3명의 남성 화자와 1명의 여성 화자로부터 수집된 약 150개의 문장을 사용하여 변환 규칙을 생성하였으며, 이를 동일한 화자가 발성한 다른 150개의 문장에 대해 적용하여 객관적인 성능 평가와 주관적 청취 테스트를 수행하였다.

• 전자공학회논문지
• /
• 제50권8호
• /
• pp.196-202
• /
• 2013

기저 세포암은 가장 일반적인 피부암이고 그 발병이 급속도로 증가하고 있다. 본 연구에서는 피부 조직에서 측정한 라만 스펙트럼에서 기저 세포암 진단을 위해 NMF(non-negative matrix factorization) 알고리즘을 사용하는 방법을 제안하였다. 측정된 라만 스펙트럼은 영역 선택과 정규화 등의 몇 가지 전처리 과정을 거쳐 분류 실험에 사용한다. 전처리 과정을 수행한 라만 스펙트럼은 NMF 알고리즘을 이용하여 분해된 행렬의 열벡터를 기저로 사용한다. 이 기저들을 선형 결합하여 각 클래스의 평균 스펙트럼에 근사하기 위한 가중치는 행렬 연산으로 결정한다. 분류 실험은 스펙트럼과 NMF에 의한 기저와 가중치의 선형 결합 스펙트럼의 차에 대한 제곱평균제곱근을 최소로 하는 클래스를 선택하는 것으로 수행한다. 기저 세포암의 진단을 위한 분류 실험에서 제안한 방법을 사용하는 경우가 약 99.1%의 평균 분류율로 이전의 BCC 진단에 사용한 방법보다 약 2-3% 정도의 향상된 성능을 보였다.