• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2023년도 학술발표회
• /
• pp.246-246
• /
• 2023

Multi-purpose dams are operated accounting for both physical and socioeconomic factors. This study aims to evaluate the utility of a deep learning algorithm-based model for three multi-purpose dam operation (Seomjin River dam, Juam dam, and Juam Control dam) in Seomjin River. In this study, the Gated Recurrent Unit (GRU) algorithm is applied to predict hourly water level of the dam reservoirs over 2002-2021. The hyper-parameters are optimized by the Bayesian optimization algorithm to enhance the prediction skill of the GRU model. The GRU models are set by the following cases: single dam input - single dam output (S-S), multi-dam input - single dam output (M-S), and multi-dam input - multi-dam output (M-M). Results show that the S-S cases with the local dam information have the highest accuracy above 0.8 of NSE. Results from the M-S and M-M model cases confirm that upstream dam information can bring important information for downstream dam operation prediction. The S-S models are simulated with altered outflows (-40% to +40%) to generate the simulated water level of the dam reservoir as alternative dam operational scenarios. The alternative S-S model simulations show physically inconsistent results, indicating that our deep learning algorithm-based model is not explainable for multi-purpose dam operation patterns. To better understand this limitation, we further analyze the relationship between observed water level and outflow of each dam. Results show that complexity in outflow-water level relationship causes the limited predictability of the GRU algorithm-based model. This study highlights the importance of socioeconomic factors from hidden multi-purpose dam operation processes on not only physical processes-based modeling but also aritificial intelligence modeling.

• 한국재료학회지
• /
• 제34권2호
• /
• pp.111-115
• /
• 2024

Silicon carbide (SiC) has emerged as a promising material for next-generation power semiconductor materials, due to its high thermal conductivity and high critical electric field (~3 MV/cm) with a wide bandgap of 3.3 eV. This permits SiC devices to operate at lower on-resistance and higher breakdown voltage. However, to improve device performance, advanced research is still needed to reduce point defects in the SiC epitaxial layer. This work investigated the electrical characteristics and defect properties using DLTS analysis. Four deep level defects generated by the implantation process and during epitaxial layer growth were detected. Trap parameters such as energy level, capture-cross section, trap density were obtained from an Arrhenius plot. To investigate the impact of defects on the device, a 2D TCAD simulation was conducted using the same device structure, and the extracted defect parameters were added to confirm electrical characteristics. The degradation of device performance such as an increase in on-resistance by adding trap parameters was confirmed.

• 대기
• /
• 제16권3호
• /
• pp.235-246
• /
• 2006

Remote sensing techniques using satellites or the scanning weather radars depend mostly on the presence of clouds or precipitation, and leave the extensive regions of clear air unobserved. But wind profilers provide the most direct measurements of mesoscale vertical air motion in the troposphere, even in the context of heavy precipitation. In this paper, the precipitation events during the Changma period was classified into 4 precipitation types - stratiform, mixed stratiform/ convective, deep convective, and shallow convective. The parameters for the classification of analysis are the vertical structure of reflectivity, Doppler velocity, and spectral width measured with the wind profiler at Haenam for a three-year period (2003-2005). In addition, the synoptic fields and total amount of precipitation were analyzed using the Global Final Analyses (FNL) data and the Global Precipitation Climatology Project (GPCP) data. During the Changma period, the results show that the stratiform type was dominant under the moist-neutral atmosphere in 2003, whereas the deep convective type was under the moist unstable condition in 2004. The stratiform type was no less popular than the deep convective type among four seasons because the moist neutral layer was formed by the convergence between the upper-level jet and the low-level jet, and by the moisture transport along the western rim of the North Pacific subtropical anticyclone.

• Archives of Plastic Surgery
• /
• 제42권5호
• /
• pp.580-587
• /
• 2015

Background Descent of the lateral aspect of the brow is one of the earliest signs of aging. The purpose of this study was to describe an open surgical technique for lateral brow lifts, with the goal of achieving reliable, predictable, and long-lasting results. Methods An incision was made behind and parallel to the temporal hairline, and then extended deeper through the temporoparietal fascia to the level of the deep temporal fascia. Dissection was continued anteriorly on the surface of the deep temporal fascia and subperiosteally beyond the temporal crest, to the level of the superolateral orbital rim. Fixation of the lateral brow and tightening of the orbicularis oculi muscle was achieved with the placement of sutures that secured the tissue directly to the galea aponeurotica on the lateral aspect of the incision. An additional fixation was made between the temporoparietal fascia and the deep temporal fascia, as well as between the temporoparietal fascia and the galea aponeurotica. The excess skin in the temporal area was excised and the incision was closed. Results A total of 519 patients were included in the study. Satisfactory lateral brow elevation was obtained in most of the patients (94.41%). The following complications were observed: total relapse (n=8), partial relapse (n=21), neurapraxia of the frontal branch of the facial nerve (n=5), and limited alopecia in the temporal incision (n=9). Conclusions We consider this approach to be a safe and effective procedure, with long-lasting results.

• Nuclear Engineering and Technology
• /
• 제32권3호
• /
• pp.244-253
• /
• 2000

A Simple Large Model (SLM), which can be used to make thermal calculation for a deep geological repository with finite number of HLW canisters, was developed. In order to develop the SLM, a Simple Basic Model (SBM), which will be a unit of the SLM, was optimized first. The SBM was optimized to achieve the same maximum buffer temperature as that of the Detailed Basic Model (DBM) representing the real geometric aspects of the repository. In contrast to the models with the assumption of infinite number of canisters which cannot consider boundary effect, the SLM can model the real repository with finite number of canisters and thus consider the boundary effect. Thermal results from the SLM can be used to evaluate the reliability of the models, which do not consider boundary effect. This model can also be used to simulate the thermal layout design and to analyze the thermal safety of a deep geological repository as well as an underground laboratory.

• 한국전기전자재료학회:학술대회논문집
• /
• 한국전기전자재료학회 2004년도 하계학술대회 논문집 Vol.5 No.1
• /
• pp.150-153
• /
• 2004

본 논문에서는 780 nm 고출력 레이저 다이오드의 신뢰성을 향상시키기 위하여 DLTS(deep level transient spectroscopy)을 이용하여 MOCVD(metalorganic chemical vapor deposition) 성장 조건 변화에 따른 $Al_{0.48}Ga_{0.52}As$와 $Al_{0.1}Ga_{0.9}As$ 물질에서의 깊은준위(deep level)의 거동을 조사하였다. DLTS 측정결과, MOCVD로 성장된 막에서만 나타나는 결함(defect)으로 추정되는 trap A(0.3 eV), DX center로 알려진 trap B, 갈륨(Ga) vacancy와 산소(O2) 원자의 복합체(complex)에 의한 결함인 trap D(0.6 eV) 및 EL2 라고 불리우는 trap E(0.9 eV)의 네 가지 깊은준위들이 관측되었고, 성장 조건의 변화에 따라 깊은 준위들의 농도가 감소하는 것을 관측함으로써 최적 성장 조건을 찾을 수 있었다.

• 방사성폐기물학회지
• /
• 제21권1호
• /
• pp.23-41
• /
• 2023

Coupled thermo-hydraulic-mechanical (THM) processes are essential for the long-term performance of deep geological disposal of high-level radioactive waste. In this study, a numerical sensitivity analysis was performed to analyze the effect of rock properties on THM responses after the execution of the heater test at the Kamaishi mine in Japan. The TOUGH-FLAC simulator was applied for the numerical simulation assuming a continuum model for coupled THM analysis. The rock properties included in the sensitivity study were the Young's modulus, permeability, thermal conductivity, and thermal expansion coefficients of crystalline rock, rock salt, and clay. The responses, i.e., temperature, water content, displacement, and stress, were measured at monitoring points in the buffer and near-field rock mass during the simulations. The thermal conductivity had an overarching impact on THM responses. The influence of Young's modulus was evident in the mechanical behavior, whereas that of permeability was noticed through the change in the temperature and water content. The difference in the THM responses of the three rock type models implies the importance of the appropriate characterization of rock mass properties with regard to the performance assessment of the deep geological disposal of high-level radioactive waste.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권4호
• /
• pp.1258-1275
• /
• 2023

Cloud computing is a new technology that has adapted to the traditional way of service providing. Service providers are responsible for managing the allocation of resources. Selecting suitable containers and bandwidth for job scheduling has been a challenging task for the service providers. There are several existing systems that have introduced many algorithms for resource allocation. To overcome these challenges, the proposed system introduces an Optimized Task Scheduling Algorithm with Deep Learning (OTS-DL). When a job is assigned to a Cloud Service Provider (CSP), the containers are allocated automatically. The article segregates the containers as' Long-Term Container (LTC)' and 'Short-Term Container (STC)' for resource allocation. The system leverages an 'Optimized Task Scheduling Algorithm' to maximize the resource utilisation that initially inquires for micro-task and macro-task dependencies. The bottleneck task is chosen and acted upon accordingly. Further, the system initializes a 'Deep Learning' (DL) for implementing all the progressive steps of job scheduling in the cloud. Further, to overcome container attacks and errors, the system formulates a Container Convergence (Fault Tolerance) theory with high-level security. The results demonstrate that the used optimization algorithm is more effective for implementing a complete resource allocation and solving the large-scale optimization problem of resource allocation and security issues.

• Journal of Information Processing Systems
• /
• 제19권3호
• /
• pp.355-369
• /
• 2023

Since the widespread adoption of deep-learning and related distributed representation, there have been substantial advancements in part-of-speech (POS) tagging for many languages. When training word representations, morphology and shape are typically ignored, as these representations rely primarily on collecting syntactic and semantic aspects of words. However, for tasks like POS tagging, notably in morphologically rich and resource-limited language environments, the intra-word information is essential. In this study, we introduce a deep neural network (DNN) for POS tagging that learns character-level word representations and combines them with general word representations. Using the proposed approach and omitting hand-crafted features, we achieve 90.47%, 80.16%, and 79.32% accuracy on our own dataset for three morphologically rich languages: Uyghur, Uzbek, and Kyrgyz. The experimental results reveal that the presented character-based strategy greatly improves POS tagging performance for several morphologically rich languages (MRL) where character information is significant. Furthermore, when compared to the previously reported state-of-the-art POS tagging results for Turkish on the METU Turkish Treebank dataset, the proposed approach improved on the prior work slightly. As a result, the experimental results indicate that character-based representations outperform word-level representations for MRL performance. Our technique is also robust towards the-out-of-vocabulary issues and performs better on manually edited text.

• 로봇학회논문지
• /
• 제18권4호
• /
• pp.463-471
• /
• 2023

This paper proposes a method to estimate the level of interest shown by visitors towards a specific target, a guide robot, in spaces where a large number of visitors, such as exhibition halls and museums, can show interest in a specific subject. To accomplish this, we apply deep learning-based behavior recognition and object tracking techniques for multiple visitors, and based on this, we derive the behavior analysis and interest level of visitors. To implement this research, a personalized dataset tailored to the characteristics of exhibition hall and museum environments was created, and a deep learning model was constructed based on this. Four scenarios that visitors can exhibit were classified, and through this, prediction and experimental values were obtained, thus completing the validation for the interest estimation method proposed in this paper.