• 대한조선학회논문집
• /
• 제59권5호
• /
• pp.271-279
• /
• 2022

In order to study the self-propulsion test and analysis techniques for the submerged body with pumpjet propulsors in the Large Cavitation Tunnel (LCT), at the Korea Research Institute of Ships and Ocean Engineering, a set of test equipment was designed and manufactured. The pumpjet propulsor is composed of rotor, stator and duct which results in the strong interaction between the components. To measure the thrust and torque for duct and stator, a ring-shaped sensor was applied. The test equipment including pumpjet is installed on the stern of the submerged body. As the whole pumpjet including duct and stator was considered as the propulsor from pumpjet open-water test, the self-propulsion test was conducted in the same way. The total thrust, combined thrust of rotor, duct and stator was used for the pumpjet self-propulsion test analysis. Accordingly, the self-propulsion test and analysis were conducted in the same way as those of the conventional propeller. The full-scale performances of the pumpjet propulsor were compared with those of the reference propeller. On the basis of the present study, it is thought that the pumpjet propulsor would be designed optimally.

• International Journal of Advanced Culture Technology
• /
• 제10권4호
• /
• pp.137-147
• /
• 2022

In this study, for students who gave up their studies midway through college, we investigate the reasons for dropping out of college, analyze the factors that caused students to drop out, and suggest improvement measures to effectively increase the recruitment rate of enrolled students. This study explores measures to increase student enrollment in community colleges. For this purpose, it conducted a survey of students who dropped out of G College between 2018 and 2020 from June 28 to July 9, 2021. Its analysis is based on the results of 53 students who participated in the survey. First, our results suggest that programs to develop interpersonal relationships between students, faculty, and staff are necessary. Such programs will cultivate a culture of solidarity and collective identity among students, which in turn will reinforce positive experiences in college. Moreover, by developing systems to form relationships between faculty and students, colleges can have a feedback mechanism, such as an open-study program, through which they learn what the students want and need. Through this effort, colleges can help their students adjust to campuses and enhance student satisfaction in college. Second, it is necessary to develop various extracurricular programs not only for freshmen but existing students and to run hobby and leisure activity programs. To have continuous, standing extracurricular programs instead of one-time events, colleges should set up specific goals, delivery methods and strategies. Third, colleges should renovate old and outdated buildings and facilities on campus to enhance the quality of campus life. Moreover, more comprehensive improvement of facilities and a campus environment by having various convenient and leisure facilities that meet the needs and demands of students. Fourth, it is suggested to develop programs or systems that help students to more fully engage in campus lives and activities, which in turn increases confidence and self-efficacy among students. Through such programs, students can better adjust to their majors and, therefore, will be less likely to drop out of college.

• 산업진흥연구
• /
• 제7권1호
• /
• pp.1-8
• /
• 2022

본 논문은 중도중단 데이터가 포함된 생존데이터의 경우 적용할 수 있는 기계학습 방법에 대해 살펴보았다. 우선 탐색적인 자료분석으로 각 특성에 대한 분포, 여러 특성들 간의 관계 및 중요도 순위를 파악할 수 있었다. 다음으로 독립변수에 해당하는 여러 특성들과 종속변수에 해당하는 특성(사망여부) 간의 관계를 분류문제로 보고 logistic regression, K nearest neighbor 등의 기계학습 방법들을 적용해본 결과 적은 수의 데이터이지만 통상적인 기계학습 결과에서와 같이 logistic regression보다는 random forest가 성능이 더 좋게 나왔다. 하지만 근래에 성능이 좋다고 하는 artificial neural network나 gradient boost와 같은 기계학습 방법은 성능이 월등히 좋게 나오지 않았는데, 그 이유는 주어진 데이터가 빅데이터가 아니기 때문인 것으로 판명된다. 마지막으로 Kaplan-Meier나 Cox의 비례위험모델과 같은 통상적인 생존분석 방법을 적용하여 어떤 독립변수가 종속변수 (t_i, δ_i)에 결정적인 영향을 미치는지 살펴볼 수 있었으며, 기계학습 방법에 속하는 random forest를 중도중단 데이터가 포함된 생존데이터에도 적용하여 성능을 평가할 수 있었다.

• 한국소프트웨어감정평가학회 논문지
• /
• 제16권2호
• /
• pp.45-52
• /
• 2020

소스 코드의 보안 취약점을 탐지하는 전통적인 방법은 많은 시간과 노력을 필요로 한다. 만약 보안 취약점 유형들에 대한 좋은 품질의 데이터가 있다면, 이와 머신러닝 기술을 활용해 효과적으로 문제를 해결할 수 있을 것이다. 이에 본 논문은 정적 프로그램 분석에 머신러닝 기술을 활용하여 소스 코드에서 보안 취약점을 탐지하는 방법을 제시하고, 실험을 통하여 가능성을 보인다. 메소드 단위의 코드 조각의 의미를 해석하여 메소드의 이름을 예측하는 code2vec 모델을 사용하고, 모델을 생성하고 검증 및 평가를 하기 위한 데이터로 흔히 발생할 수 있는 보안 취약점을 모아놓은 Juliet Test Suite를 사용하였다. 모델 평가 결과 약 97.3%의 정밀도와 약 98.6%의 재현율로 매우 희망적인 결과를 확인하였고 오픈 소스 프로젝트의 취약점을 탐지함으로써 가능성을 보였다. 향후 연구를 통해 다른 취약점 유형과 다양한 언어로 작성된 소스 코드에 대해서 대응함으로써 기존의 분석 도구들을 대체할 수 있을 것이다.

• KEPCO Journal on Electric Power and Energy
• /
• 제8권1호
• /
• pp.5-7
• /
• 2022

KEPCO's plan is undergoing a trial operation to replace the open-loop section with ring main units configuration where underground distribution lines are installed, by linking the multi-way circuit breakers auto (MCA) on the power side of each pad-mounted transformer. However, ring main units application mentioned above may cause the ripple effects, when implementing the configuration without a study of protection coordination. Because ring main units with classical pre-set protection devices contribution in fault condition didn't consider yet. For the reliable ring main units operation, it is necessary to resolve several protection issues such as the protection coordination with substation side, prevention of the transformer inrush current. These issues can radically deteriorate the distribution system reliability Hence, it is essential to design proper protection coordination to reduce these types of problems. This paper presents a scheme of ring main units' configuration and MCA's settings of time-current curves to preserve the performance of protection coordination among the switchgears considering constraints, e.g. prevention of the ripple effects (on the branch section when a transformer failure occurs and the mainline when a branch line failure occurs). It was confirmed that the propagation of the failure for each interrupter segment could be minimized by applying the proposed TCC and the interrupter settings for the MCAs (branch, transformer). Further, it was verified that the undetected area of the distribution automation system (DAS) could be supplemented by having the MCA configurated ring main units operate first, instead of the internal protection equipment in the transformer such as the fuse, STP when a transformer failure occurs.

• 국제학술발표논문집
• /
• The 3th International Conference on Construction Engineering and Project Management
• /
• pp.1346-1351
• /
• 2009

In this paper we propose a framework for an Object Relational IFC Server (OR-IFC Server). Enormous amounts of information are generated in each project. Today, many BIM systems are developed by various CAD software vendors. Industry Foundation Classes (IFC) developed by International Alliance for Interoperability (IAI) is an open standard data model for exchanging data between the various BIM tools. The IFC provides a foundation for exchanging and sharing of information directly between software applications and define a shared building project model. The IFC model server is a database management system that can keep track of transactions, modifications, and deletions. It plays a role as an information hub for storing and sharing information between various parties involved in construction projects. Users can communicate with each other via the internet and utilize functions implemented in the model server such as partial data import/export, file merge, version control, etc. IFC model servers using relational database systems have been developed. However, they suffered from slow performance and long transaction time due to a complex mapping process between the IFC structure and a relational-database structure because the IFC model schema is defined in the EXPRESS language which is object-favored language. In order to simplify the mapping process, we developed a set of rules to map the IFC model to an object-relational database (ORDB). Once the database has been configured, only those pieces of information that are required for a specific information-exchange scenario are extracted using the pre-defined information delivery manual (IDM). Therefore, file sizes will be reduced when exchanging data, meaning that files can now be effectively exchanged and shared. In this study, the framework of the IFC server using ORDB and IDM and the method to develop it will be examined.

• 한국전자통신학회논문지
• /
• 제18권4호
• /
• pp.677-686
• /
• 2023

본 논문에서는 사용자의 외출 목적을 인지하고 외출 시 도움을 줄 수 있는 유용한 정보를 전달하는 스마트 홈 외출 케어 시스템의 설계 및 구현에 대해 연구한다. ESP8266을 이용하여 기상청의 RSS 서비스 데이터를 실시간으로 전송할 수 있고, Arduino MEGA를 이용하여 데이터를 분석 후 사용자에게 기상정보를 제공할 수 있는 시스템을 구현한다. 앱인벤터를 활용하여 필요한 물품을 잊지 않고 챙길 수 있으며 원하는 날씨와 목적에 맞게 설정을 변경 가능하다. 마이크 위치는 외부에 배치하여 인지도를 12% 높였으며, 압력센서의 감도는 최대 210 kΩ으로 설정했다. 문 사이에 장애물이 있을 경우 자동으로 문이 열린다. 서랍 천장에 초음파 센서를 배치해 0.5cm~10cm 범위 내 물체를 인식해 물체 유무를 확인하고 카메라를 설치해 보안 강화 시스템을 연구헸다.

• 국제학술발표논문집
• /
• The 1th International Conference on Construction Engineering and Project Management
• /
• pp.655-661
• /
• 2005

For a hydro power plant project, the headrace tunnel having a finished diameter of 3.3m was constructed in volcanic rocks with well-developed vertical joint and high groundwater table. The intake facility was located 20.3 km upstream of the powerhouse and headrace tunnel of 20 km in length and penstock of 440 m in height connected the intake and the powerhouse. The typical caldera lake, Lake Toba set the geology at the site; the caving of the ground caused tension cracks in the vertical direction to be developed and initial stresses at the ground to be released. High groundwater table(the maximum head of 20 bar) in the area of well-connected vertical joints delayed the progress of tunnel excavation severely due to the excessive inflow of groundwater. The excavation of tunnel was made using open-shield type TBM and mucking cars on the rail. High volume of water inflow raised the water level inside tunnel to 70 cm, 17% of tunnel diameter (3.9 m) and hindered the mucking of spoil under water. To improve the productivity, several adjustments such as modification of TBM and mucking cars and increase in the number of submersible pumps were made for the excavation of severe water inflow zone. Since the ground condition encountered during excavation turned out to be much worse, it was decided to adopt PC segment lining instead of RC lining. Besides, depending on the conditions of the water inflow, rock mass condition and internal water pressure, one of the invert PC segment lining with in-situ RC lining, RC lining and steel lining was applied to meet the site specific condition. With the adoption of PC segment lining, modification of TBM and other improvement, the excavation of the tunnel under severe groundwater condition was successfully completed.

• Structural Engineering and Mechanics
• /
• 제84권3호
• /
• pp.423-435
• /
• 2022

In this article, we present torsion-bending analysis of a composite FGM beam with an open section, according to the advanced and refined theory of 1D / 3D beams based on the 3D Saint-Venant's solution and taking into account the edge effects. The (initially one-dimensional) model contains a set of three-dimensional (3D) displacement modes of the cross section, reflecting its 3D mechanical behaviour. The modes are taken into account depending on the mechanical characteristics and the geometrical form of the cross-section of the composite FGM beam. The model considered is implemented on the CSB (Cross-Section and Beam Analysis) software package. It is based on the RBT/SV theory (Refined Beam Theory on Saint-Venant principle) of FGM beams. The mechanical and physical characteristics of the FGM beam continuously vary, depending on a power-law distribution, across the thickness of the beam. We compare the numerical results obtained by the three-beam theories, namely: The Classical Beam Theory of Saint-Venant (Classical Beam Theory CBT), the theory of refined beams (Refined Beam Theory RBT), and the theory of refined beams, using the higher (high) modes of distortion of the cross-section (Refined Beam Theory using distorted modes RBTd). The results obtained confirm a clear difference between those obtained by the three models at the level of the supports. Further from the support, the results of RBT and RBTd are of the same order, whereas those of CBT remains far from those of higher-order theories. The 3D stresses, strains and displacements, obtained by the present study, reflect the 3D behaviour of FGM beams well, despite the initially 1D nature of the problem. A validation example also shows a very good agreement of the proposed models with other models (classical or higher-order beam theory) and Carrera Unified Formulation 1D-beam model with Lagrange Expansion functions (CUF-LE).

• International Journal of Advanced Culture Technology
• /
• 제10권1호
• /
• pp.274-283
• /
• 2022

Korea is facing a number difficulties arising from rising housing prices. As 'housing' takes the lion's share in personal assets, many difficulties are expected to arise from fluctuating housing prices. The purpose of this study is creating housing price prediction model to prevent such risks and induce reasonable real estate purchases. This study made many attempts for understanding real estate instability and creating appropriate housing price prediction model. This study predicted and validated housing prices by using the LSTM technique - a type of Artificial Intelligence deep learning technology. LSTM is a network in which cell state and hidden state are recursively calculated in a structure which added cell state, which is conveyor belt role, to the existing RNN's hidden state. The real sale prices of apartments in autonomous districts ranging from January 2006 to December 2019 were collected through the Ministry of Land, Infrastructure, and Transport's real sale price open system and basic apartment and commercial district information were collected through the Public Data Portal and the Seoul Metropolitan City Data. The collected real sale price data were scaled based on monthly average sale price and a total of 168 data were organized by preprocessing respective data based on address. In order to predict prices, the LSTM implementation process was conducted by setting training period as 29 months (April 2015 to August 2017), validation period as 13 months (September 2017 to September 2018), and test period as 13 months (December 2018 to December 2019) according to time series data set. As a result of this study for predicting 'prices', there have been the following results. Firstly, this study obtained 76 percent of prediction similarity. We tried to design a prediction model of real estate transaction price with the LSTM Model based on AI and Bigdata. The final prediction model was created by collecting time series data, which identified the fact that 76 percent model can be made. This validated that predicting rate of return through the LSTM method can gain reliability.