• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.21 no.1
• /
• pp.73-82
• /
• 1997

High temperature densification behaviors of alumina powder compacts were investigated under hot pressing and hot isostatic pressing. An alumina part of valve-head shape was fabricated under hot pressing and its forming process was simulated by finite element calculation. an alumina powder compact encapsulated by a stainless steel container was also densified under hot isostatic pressing. Inhomogeneous deformations during hot isostatic pressing due to the canning effect were observed experimentally and predicted by finite element analysis.

• Journal of the Korean Society of Fisheries and Ocean Technology
• /
• v.26 no.4
• /
• pp.380-386
• /
• 1990

The high stress, low cycle fatigue test were carried out on HT80 plates with artificial part-through notch. The LBB conditions by the net section stress approach was examined in detail. Finite element methods for the state of deformation were also applied and the results were compared with the results from experiments. From the test results, it was noted that the LBB condition in fatigue test was dependent upon the maximun load change before and after the crack penetration. The LBB criterion By the net section stress approach were slightly dependent on initial crack size. It was shown that the state of deformation can be used to predict whether the component is likely to show LBB condition or they will be broken rapidly.

• Annual Conference on Human and Language Technology
• /
• 2000.10d
• /
• pp.262-268
• /
• 2000

본 논문에서는 어휘 의미 애매성 해소와 영어 대역어 사전 그리고 외국언어에 존재하는 개념체계를 이용하여 한국어 개념체계를 자동으로 구축하는 방법을 기술한다. 본 논문에서 사용하는 방법은 기존의 개념체계 구축 방법들에 비해 적은 노력과 시간을 필요로 한다. 또한 상기한 자동 구축 방법에서 사용하는 어휘 의미 애매성 해소를 위한 6가지 feature도 함께 설명한다.

• The Korean Journal of Malacology
• /
• v.24 no.2
• /
• pp.131-136
• /
• 2008

Growth and survival of the spat of arkshell, Scapharca broughtonii were investigated for the improvement of survival in intermediate culture with different shape of protective net and type of preventive net of spat loss. The growth performance of the spat of arkshell was observed for 60 days by using different forms of protective net, as the exposure form with an average shell length of $12.8{\pm}3.2$ mm(P<0.05), the fish pot form with $12.2{\pm}3.5$ mm, the cylinder form with $11.9{\pm}3.8$ mm and the last one is the rectangular form with $10.9{\pm}3.7$ mm. Their numerical value did not show any marked difference with each other. But in case of survival rate, the fish pot form showed highest survival rate which is 43.7% and the significantly lowest was showed by the exposure form, which is 5.4%(P<0.05). The cylinder form and rectangular form showed 41.2 and 31.6% respectively. And then the growth of the spat of arkshell was observed for 163 days by rearing in a sort of preventive net, the first group was in a blackout curtain with average shell length of $13.9{\pm}3.1$ mm, the next group was in a balsam pear net with $12.9{\pm}3.0$ mm, in a polyethylene net with $11.8{\pm}3.1$ mm and the control with $12.6{\pm}3.3$ mm which was not installed by preventive net of spat loss. The survival rate was 91.5% in a blackout curtain, 90.1% in a polyethylene net, 88.5% in a balsam pear net and 61.5% in a controlled group. It is seen that the highest growth and survival rate were observed in a fish pot form and cylinder form. These results were a little bit difference from those of the spat of arkshell cultured in the form of different preventive nets. There was no significant difference in survival rate in relation to the sorts of preventive net against carrying-away, but there was a difference in survival rate by more than 30% as against the non-installed controlled group. We expect that survival rate would be highly improved in intermediate culture carried out with fish pot form and cylinder form of protective net and preventive net of spat loss.

• Proceedings of the KIEE Conference
• /
• 1990.07a
• /
• pp.535-538
• /
• 1990

The main problem in speech recognition is the enormous variability in acoustic signals due to complex but predictable contextual effects. Especially in plosive consonants it is very difficult to find invariant cue due to various contextual effects, but humans use these contextual effects as helpful information in plosive consonant recognition. In this paper we experimented on three artificial neural net models for the recognition of plosive consonants. Neural Net Model I used "Multi-layer Perceptron ". Model II used a variation of the "Self-organizing Feature Map Model". And Model III used "Interactive and Competitive Model" to experiment contextual effects. The recognition experiment was performed on 9 Korean plosive consonants. We used VCV speech chains for the experiment on contextual effects. The speech chain consists of Korean plosive consonants /g, d, b, K, T, P, k, t, p/ (/ㄱ, ㄷ, ㅂ, ㄲ, ㄸ, ㅃ, ㅋ, ㅌ, ㅍ/) and eight Korean monothongs. The inputs to Neural Net Models were several temporal cues - duration of the silence, transition and vot -, and the extent of the VC formant transitions to the presence of voicing energy during closure, burst intensity, presence of asperation, amount of low frequency energy present at voicing onset, and CV formant transition extent from the acoustic signals. Model I showed about 55 - 67 %, Model II showed about 60%, and Model III showed about 67% recognition rate.

• Proceedings of the SAREK Conference
• /
• 2007.06a
• /
• pp.161-161
• /
• 2007

• Korean Journal of Soil Science and Fertilizer
• /
• v.40 no.6
• /
• pp.471-475
• /
• 2007

This study was carried out to elucidate the relationship between yields and inorganic nutrient contents, and then establish of critical range of inorganic nutrient contents in leaves of net melon in protected cultivation in Naju from 2004 to 2006. In considering the yields and nutrient contents of net melon, the critical ranges of macro-nutrient contents in leaves of net melon at 25 days after transplanting were in N 3.4~4.8%, P 0.4~0.7%, K 3.4~5.1%, Ca 2.7~4.1% and Mg 0.6~0.8%, respectably. The critical ranges of micronutrient content, such as Fe, Mn, Zn, Cu and B, were not found the regular trend with different growth stages. However, the critical range of micronutrient contents in leaves of net melon at 25 days after transplanting were in Fe 102~164, Mn 52~116, Zn 28~98, Cu 7~15 and B $36{\sim}72mg\;kg^{-1}$, respectably. Finally, these results might be used at the indicator for critical nutrient contents for diagnosis of nutritional disorder in net melon in protected cultivation.

• Proceedings of the KIEE Conference
• /
• 2005.07d
• /
• pp.2802-2805
• /
• 2005

본 논문에서는 당사에서 개발한 부스터펌프 원격관리 시스템을 소개한다. 부스터펌프 원격관리 시스템은 인터넷을 통하여 급수 장비인 부스터펌프 시스템에 대한 원격관리와 정보 공유가 가능하며 이를 통해 장비 운영자, 납품업체, 시공업체 등에서 누구나 부스터펌프 시스템을 감시 및 관리를 할 수 있게 한 웹기반의 통합 시스템이다. 부스터펌프 원격관리 시스템은 각각의 부스터펌프 시스템이 설치되어 있는 현장에 당사 Gateway 제품인 ProNET-iMC를 설치하여 부스터펌프 시스템의 정보를 수집하고 이를 상위 시스템에서 취합하여 데이터베이스에 저장하며, 웹을 통해 사용자 권한에 따른 실시간 및 이력 정보를 제공하도록 설계되었다. 부스터펌프 원격관리 시스템을 통하여 부스터펌프 시스템의 고장을 방지하거나 조기에 발견하여 신속한 A/S를 통해 제품의 안정성을 높이고 사용자의 신뢰를 구축 할 수 있을 것이다.

• International Journal of Internet, Broadcasting and Communication
• /
• v.14 no.1
• /
• pp.136-141
• /
• 2022

A very complex task in deep learning such as image classification must be solved with the help of neural networks and activation functions. The backpropagation algorithm advances backward from the output layer towards the input layer, the gradients often get smaller and smaller and approach zero which eventually leaves the weights of the initial or lower layers nearly unchanged, as a result, the gradient descent never converges to the optimum. We propose a two-factor non-saturating activation functions known as Bea-Mish for machine learning applications in deep neural networks. Our method uses two factors, beta (𝛽) and alpha (𝛼), to normalize the area below the boundary in the Mish activation function and we regard these elements as Bea. Bea-Mish provide a clear understanding of the behaviors and conditions governing this regularization term can lead to a more principled approach for constructing better performing activation functions. We evaluate Bea-Mish results against Mish and Swish activation functions in various models and data sets. Empirical results show that our approach (Bea-Mish) outperforms native Mish using SqueezeNet backbone with an average precision (AP50val) of 2.51% in CIFAR-10 and top-1accuracy in ResNet-50 on ImageNet-1k. shows an improvement of 1.20%.

• Geophysics and Geophysical Exploration
• /
• v.25 no.4
• /
• pp.177-188
• /
• 2022

Recently, the application of distributed acoustic sensors (DAS), which can replace geophones and seismometers, has significantly increased along with interest in micro-seismic monitoring technique, which is one of the CO₂ storage monitoring techniques. A significant amount of temporally and spatially continuous data is recorded in a DAS monitoring system, thereby necessitating fast and accurate data processing techniques. Because event detection and seismic phase picking are the most basic data processing techniques, they should be performed on all data. In this study, a machine learning-based P, S wave phase picking algorithm was developed to compensate for the limitations of conventional phase picking algorithms, and it was modified using a transfer learning technique for the application of DAS data consisting of a single component with a low signal-to-noise ratio. Our model was constructed by modifying the convolution-based EQTransformer, which performs well in phase picking, to the ResUNet structure. Not only the global earthquake dataset, STEAD but also the augmented dataset was used as training datasets to enhance the prediction performance on the unseen characteristics of the target dataset. The performance of the developed algorithm was verified using K-net and KiK-net data with characteristics different from the training data. Additionally, after modifying the trained model to suit DAS data using the transfer learning technique, the performance was verified by applying it to the DAS field data measured in the Pohang Janggi basin.