• 한국융합학회논문지
• /
• 제11권12호
• /
• pp.23-30
• /
• 2020

본 연구에서는, 천연가스(NG) 데이터와 가스 관련 환경 요소 간의 관계를 기계학습 알고리즘을 사용하여 가스 누출 데이터를 직접 측정하지 않고 가스 누출 위험 수준을 예측하였다. 이번 연구는 서버가 제공하는 오픈 데이터인 IoT 기반 원격 제어 피카로(Picarro) 가스 센서 사양을 기반으로 사용했다. 천연 가스는 공기 중으로 누출이 되며, 대기 오염, 환경, 그리고 건강에 큰 문제가 된다. 본 연구에서 제안하는 방법은 천연 가스의 누출 위험 예측을 위한 랜덤 포레스트(Random Forest) 분류 기반 다변량 특이치 제거 방법이다. 비지도 k-평균 클러스터링 후에 실험 데이터 집합은 불균형 데이터이다. 따라서 우리는 제안된 모델이 중간과 높은 위험 수준을 가장 잘 예측할 수 있다는 점에 초점을 맞춘다. 이 경우 각 분류 모델에 대한 수신자 조작 특성(ROC) 곡선, 정확도, 평균 표준 오차(MSE)를 비교했다. 실험 결과로 정확도, 수신자 조작 특성의 곡선 아래 영역(AUC, Area Under the ROC Curve), MSE가 각각 MOL_RF의 경우 99.71%, 99.57%, 및 0.0016의 결과 값을 얻었다.

• 한국수자원학회논문집
• /
• 제54권8호
• /
• pp.599-606
• /
• 2021

지하에 매설되어 있는 사회기반시설물 중 하나인 상수도시스템은 정수처리된 물을 수용가에게 수송 및 공급하는 기능을 가지고 있다. 최근들어, 계측능력이 향상됨에 따라 유량데이터에 의한 딥러닝기법을 적용한 누수 인지 및 탐지와 관련한 연구가 다수 수행되고 있다. 본 연구에서는 현재까지 상수도 분야에 적용되지 않은 LSTM 기반의 딥러닝 알고리즘을 활용하여 누수발생에 대한 인지 모형을 개발하였다. 가정한 데이터를 기반으로 모형에 대한 검증을 수행하였으며 2% 이상의 누수가 발생한 경우에 대하여 모두 인식이 가능한 것으로 나타났다. 향후, 제안된 모형을 토대로 유량 데이터 예측부분에 있어서 보다 정밀한 결과가 도출 될 수 있을것으로 판단된다.

• 한국위험물학회지
• /
• 제2권1호
• /
• pp.31-37
• /
• 2014

Since the unprecedented hydrogen fluoride leak accident in 2012, there has been growing demand for customized technical information for rapid response and chemical accident management agencies including the Ministry of Environment, the National Emergency Management Agency, and the National Police Agency need more information on chemicals and accident management. In this regard, this study aims to provide reliable technical data and guidelines to initial response agencies, similar to accident management technical reports of the US and Canada. In this study, we conducted a questionnaire survey and interviews on initial response agencies like fire stations, police stations, and local governments to identify new information items for appropriate initial response and improvements of current guidelines. We also collected and reviewed the Canada's TIPS, US EPA's hydrogen fluoride documents, domestic and foreign literature on applicability tests of control chemicals, and interview data, and then produced items to be listed in the technical guidelines. In addition, to establish database of on-site technical information, we carried out applicability tests for accident control data including ① emergency shut down devide, safety guard, shut down valve, ground connection, dyke, transfer pipe, scrubber, and sensor; ② literature and field survey on distribution type and transportation/storage characteristics (container identification, valve, ground connection, etc.); ③ classification and identification of storage/transportation facilities and emergency management methodslike leak prevention, chemicals control, and cutoff or bypass of rain drainage; ④ domestic/foreign analysis methods and environmental standards including portable detection methods, test standards, and exposure limits; and ⑤ comparison/evaluation of neutralization efficiency of control chemicals on toxic substances.

• 한국안전학회지
• /
• 제28권2호
• /
• pp.21-25
• /
• 2013

The purpose of this study is to provide basic data for the safety assessment of a water purifier when water leaks due to inappropriate maintenance and the examination of the cause of accidents related to the leak. Due to its inspection and management by non-specialists, if a leak occurs in a water purifier with the water level controller being inclined, it may result in the failure of the compressor, power supply line, PCB, etc. The analysis of the thermal diffusion pattern of water purifier compressors using a thermal image camera shows that its maximum temperature was approximately $80^{\circ}C$. In addition, its operating current was a maximum of 13 A and the system's operating current was approximately 1.7 A after the compressor was charged. It was found that the housing type power cable cover of the compressor had the effect of preventing electric shock but has poor flame resistance. Furthermore, the performance of the overload protector, PTC relays, etc., was excellent but they have potential for problems as metallic terminals were exposed, resulting in the potential of a safety related accident. The terminals and their surface damaged by the tracking showed a trace of carbonization and the resistance between terminals was measured to be approximately $8{\Omega}$. In addition, while the tracking was proceeding, the fuse and circuit breaker installed for system protection did not operate.

• Corrosion Science and Technology
• /
• 제12권5호
• /
• pp.227-232
• /
• 2013

A number of components installed in the secondary system of nuclear power plants are exposed to aging mechanisms such as FAC (Flow-Accelerated Corrosion), Cavitation, Flashing, and LDIE (Liquid Droplet Impingement Erosion). Those aging mechanisms can lead to thinning of the components. In April 2013, one (1) inch small bore piping branched from the main steam line experienced leakage resulting from wall thinning in a 1,000 MWe Korean PWR nuclear power plant. During the normal operation, extracted steam from the main steam line goes to condenser through the small bore piping. The leak occurred in the downstream of an orifice. A control valve with vertical flow path was placed on in front of the orifice. This paper deals with UT (Ultrasonic Test) thickness data, SEM images, and numerical simulation results in order to analyze the extent of damage and the cause of leakage in the small bore piping. As a result, it is concluded that the main cause of the small bore pipe wall thinning is liquid droplet impingement erosion. Moreover, it is observed that the leak occurred at the reattachment point of the vortex flow in the downstream side of the orifice.

• 한국정밀공학회지
• /
• 제18권6호
• /
• pp.67-73
• /
• 2001

Three dimensional, elastic-plastic finite element(FE) analyses for circumferential through-wall cracked pipes under combined tension and bending are performed using actual tensile data of stainless steel, for two purposes. The first one is to validate the recently-proposed enhanced reference stress (ERS) method to estimate the J-integral and COD for circumferential through-wall cracked pipes under combined tension and bending. The second one is to compare those results with the GE/EPRI estimations. The FE results of the J-integral and the COD, resulting from six cases of proportional and non-proportional combined tension and bending, compare very well with those estimated from the proposed method. Excellent agreements of the proposed method with the detailed FE results provide sufficient confidence in the use of the proposed method to the Leak-Before-Break(LBB) analysis of through-wall cracked pipes under combined tension and bending.

• International Journal of Ocean Engineering and Technology Speciallssue:Selected Papers
• /
• 제4권1호
• /
• pp.44-50
• /
• 2001

Existing surface defects in structural members often act as sites of fatigue crack initiation, and if undetected, these cracks may grow through the thickness of the member, leading to catastrophic failure of the structure. Thus, in-service monitoring of fatigue cracks through reliable and effective nondestructive techniques is an important ingredient in the leak-before-break (LBB) design and safe operation of defects critical structures. An advanced, waveform-based, acoustic emission (AE) technique has been used in this paper to study the characteristics of the signals emanating from the initiation, growth and through-the -thickness penetration of surface fatigue crack in a 6061 aluminum plate. The goal of this experimental study is to determine whether the evolution of the fatigue crocks could be identified from the properties of the waveforms produced during the tests. The AE waveform signals detected at different stages of crack growth was found to have different temporal and spectral characteristics. The data analysis technique presented here can be applied to real-time monitoring of the initiation and propagation of fatigue cracks in structural components.

• 상하수도학회지
• /
• 제26권5호
• /
• pp.705-712
• /
• 2012

Selecting Daechung Dam as a sample study site, this study simulates virtual water quality incident which can be occurred using EFDC maintained by USEPA. In order to predict the behavior and diffusion of leaked conservative pollutant within dam under the worst condition, the hydrological data and information from 2008 were used. EFDC was successfully calibrated for observed water level obtained from the above sources. From the results of simulations, even though the concentrations (500 ppm, 1,000 ppm and 10,000 ppm) of leaked pollutant were different with each other at the same sources, the travel time of each peak concentration appeared similar. Also, changing the leak source point from dam gate(0 km) to 7 km, it was found that as leak source point was nearer to the dam gate, the travel time of each peak concentration showed up sooner. It was simulated to take 1 day to 15 days for initial appearance of the leaked pollutant according to the leaked points, and 3 days to 25 days for the reach of the peak concentration, respectively.

• 한국건축시공학회:학술대회논문집
• /
• 한국건축시공학회 2017년도 춘계 학술논문 발표대회
• /
• pp.142-143
• /
• 2017

An analysis of the current water leak status of understand structure (underground parking lots, staircases, plumbing systems, water reservoirs, etc.) of multi-family housing in South Korea shows that water leaks are found from cracks in all areas of the underground structure caused by the degradation environment (water pressure by underground water, humidity, temperature, earth pressure, soil behaviour and vibration, etc.), which result in various problems, including mold, malodour, debonding of finishing materials, exfoliation, breakout, water leaks in electrical boxes, efflorescence, sedimentation of calcium hydroxide, decoloration, rusting, damages and pollution among others. Therefore, this study aims to analyse the current status of water leaks in underground structure and use the results as the basic data for developing a standard guideline for water leaks and maintenance by parts of the underground structure of multi-family housing.

• 한국전산유체공학회지
• /
• 제14권2호
• /
• pp.9-18
• /
• 2009

The cargo containment system (CCS) for ships carrying cryogenic fluid consists of at least two levels of barriers and insulation layers. It is because, even though there is a small amount of leak through the primary barrier, the liquid tight secondary barrier blocks further leakage of the cryogenic fluid. However, once the secondary barrier is damaged, it is highly possible that the leaked cryogenic fluid flows through the flat joint made of glass wool and reaches the inner hull of the ship. The primary objective of the present study is to investigate the influence of the damage extent in the secondary barrier on the amount of leaked cryogenic fluid reaching the inner hull and the temperature distribution there. Simulation results using a computational fluid dynamics tool were compared with the experimental data for the leaked cryogenic fluid flow and evaporation in the secondary insulation layer. The experimental and computational results suggest that, unless there is a massive leak, the cryogenic fluid mostly evaporates in the insulation layer and does not reach the inner hull in the state of liquid.