• Structural Engineering and Mechanics
• /
• 제77권1호
• /
• pp.47-56
• /
• 2021

This paper deals with damage detection using a Gapped Smoothing Method (GSM) combined with deep learning. Convolutional Neural Network (CNN) is a model of deep learning. CNN has an input layer, an output layer, and a number of hidden layers that consist of convolutional layers. The input layer is a tensor with shape (number of images) × (image width) × (image height) × (image depth). An activation function is applied each time to this tensor passing through a hidden layer and the last layer is the fully connected layer. After the fully connected layer, the output layer, which is the final layer, is predicted by CNN. In this paper, a complete machine learning system is introduced. The training data was taken from a Finite Element (FE) model. The input images are the contour plots of curvature gapped smooth damage index. A free-free beam is used as a case study. In the first step, the FE model of the beam was used to generate data. The collected data were then divided into two parts, i.e. 70% for training and 30% for validation. In the second step, the proposed CNN was trained using training data and then validated using available data. Furthermore, a vibration experiment on steel damaged beam in free-free support condition was carried out in the laboratory to test the method. A total number of 15 accelerometers were set up to measure the mode shapes and calculate the curvature gapped smooth of the damaged beam. Two scenarios were introduced with different severities of the damage. The results showed that the trained CNN was successful in detecting the location as well as the severity of the damage in the experimental damaged beam.

• 한국수산과학회지
• /
• 제54권2호
• /
• pp.194-207
• /
• 2021

It is necessary to assess and manage the different elements of the marine ecosystem, such as climate change, habitat, primary and secondary production, energy flow, food web, potential yield, and fishing, to maintain the health of the ecosystem as well as support sustainable development of fishery. We set up an ecosystem model around the Korean peninsula to produce scientific predictions necessary for the assessment and management of marine ecosystems and presented the usability of the model with scenario experiments. We used the Atlantis ecosystem model based on the marine food web; Atlantis is a three-dimensional end-to-end model that includes the information and processes within an entire system, from an abiotic environment to human activity. We input the ecological and biological parameters, such as growth, mortality, spawning, recruitment, and migration, to the Atlantis model via functional groups using existing research and local measurements. During the simulation period (2018-2019), we confirmed that the model reproduced the observed data reasonably and reflected the actual ecosystem characteristics appropriately. We thus identified the usability of a marine ecosystem model with experiments on different environmental change scenarios.

• KEPCO Journal on Electric Power and Energy
• /
• 제7권2호
• /
• pp.277-283
• /
• 2021

Power plant simulators have been used for operator training, control verification and engineering verification. In general, simulators can be used in the place where they are installed by only single user group at a time. Considering high cost of simulator development, a lot of available scenarios, the diversity of user level and accessibility based on users' work location, development of simulator system that can be used by multiple user groups regardless of location is required in order to enhance utilization of simulators. In this paper, the simulator system that can be used by multiple user group simultaneously without location limitation is proposed. The simulator system is composed of simulator servers, database servers, HMI servers, a web server, web clients. Simulator server consists of control model, process model that are developed for Circulating Fluidized Bed power plant located overseas. A web server manages user accounts, operation procedures, multiple server access between web client group and simulator server group. In other words, a web server makes a user group select a simulator server at a time. The developed simulator system is integrated after implementing process model, control model, HMI, and web server. Web client systems are installed on local site where power plant is located, while simulator servers, HMI servers, database servers, and a web server are located in KEPCO RI. The developed simulator system is verified by steady-state test, malfunction test and so on via remote access.

• Nuclear Engineering and Technology
• /
• 제54권10호
• /
• pp.3604-3613
• /
• 2022

The Physical Protection System (PPS) plays an important role and must effectively deal with various adversary attacks in nuclear security. In specific single adversary path scenarios, we can calculate the PPS effectiveness by EASI (Estimated Adversary Sequence Interruption) through Probability of Interruption (P_I) calculation. EASI uses a single value of the probability of detection (P_D) and the probability of alarm communications (P_C) in the PPS. In this study, we develop a multi-path analysis code based on EASI to evaluate the effectiveness of PPS. Our quantification method for P_I considers the variability and uncertainty of P_D and P_C value by Monte Carlo simulation. We converted the 2-D scheme of the nuclear facility into an Adversary Sequence Diagram (ASD). We used ASD to find the adversary path with the lowest probability of interruption as the most vulnerable paths (MVP). We examined a hypothetical facility (Hypothetical National Nuclear Research Facility - HNNRF) to confirm our code compared with EASI. The results show that implementing the variability extension can estimate the P_I value and its associated uncertainty. The multi-path analysis code allows the analyst to make it easier to assess PPS with more extensive facilities with more complex adversary paths. However, the variability of the P_D value in each protection element allows a significant decrease in the PI value. The possibility of this decrease needs to be an important concern for PPS designers to determine the P_D value correctly or set a higher standard for PPS performance that remains reliable.

• Earthquakes and Structures
• /
• 제22권2호
• /
• pp.185-201
• /
• 2022

Structural Health Monitoring (SHM) is used to provide reliable information about the structure's integrity in near realtime following extreme incidents such as earthquakes, considering the inevitable aging and degradation that occurs in operating environments. This paper experimentally investigates an integrated wireless sensor network (Wi-SN) based monitoring technique for damage detection in concrete structures. An effective SHM technique can be used to detect potential structural damage based on post-earthquake data. Two novel methods are proposed for damage detection in reinforced concrete (RC) building structures including: (i) Jerk Energy Method (JEM), which is based on time-domain analysis, and (ii) Modal Contributing Parameter (MCP), which is based on frequency-domain analysis. Wireless accelerometer sensors are installed at each story level to monitor the dynamic responses from the building structure. Prior knowledge of the initial state (immediately after construction) of the structure is not required in these methods. Proposed methods only use responses recorded during ambient vibration state (i.e., operational state) to estimate the damage index. Herein, the experimental studies serve as an illustration of the procedures. In particular, (i) a 3-story shear-type steel frame model is analyzed for several damage scenarios and (ii) 2-story RC scaled down (at 1/6th) building models, simulated and verified under experimental tests on a shaking table. As a result, in addition to the usual benefits like system adaptability, and cost-effectiveness, the proposed sensing system does not require a cluster of sensors. The spatial information in the real-time recorded data is used in global damage identification stage of SHM. Whereas in next stage of SHM, the damage is detected at the story level. Experimental results also show the efficiency and superior performance of the proposed measuring techniques.

• 시스템엔지니어링학술지
• /
• 제18권2호
• /
• pp.94-107
• /
• 2022

Accidents prevention and mitigation is the highest priority of nuclear power plant (NPP) operation, particularly in the aftermath of the Fukushima Daiichi accident, which has reignited public anxieties and skepticism regarding nuclear energy usage. To deal with accident scenarios more effectively, operators must have ample and precise information about key safety parameters as well as their future trajectories. This work investigates the potential of machine learning in forecasting NPP response in real-time to provide an additional validation method and help reduce human error, especially in accident situations where operators are under a lot of stress. First, a base-case SGTR simulation is carried out by the best-estimate code RELAP5/MOD3.4 to confirm the validity of the model against results reported in the APR1400 Design Control Document (DCD). Then, uncertainty quantification is performed by coupling RELAP5/MOD3.4 and the statistical tool DAKOTA to generate a large enough dataset for the construction and training of neural-based machine learning (ML) models, namely LSTM, GRU, and hybrid CNN-LSTM. Finally, the accuracy and reliability of these models in forecasting system response are tested by their performance on fresh data. To facilitate and oversee the process of developing the ML models, a Systems Engineering (SE) methodology is used to ensure that the work is consistently in line with the originating mission statement and that the findings obtained at each subsequent phase are valid.

• 한국안전학회지
• /
• 제37권3호
• /
• pp.52-59
• /
• 2022

Conventionally, all the unsafe acts by human beings in relation to industrial accidents have been regarded as unintentional human errors. Exceptionally, however, in the cases with fatalities, seriously injured workers, and/or losses that evoked social issues, attention was paid to violating related laws and regulations for finding out some people to be prosecuted and given judicial punishments. As Heinrich stated, injury or loss in an accident is quite a random variable, so it can be unfair to utilize it as a criterion for prosecution or punishment. The present study was conducted to comprehend how categorizing intentional violations in unsafe acts might disrupt conventional conclusions about the industrial accident process. It was also intended to seek out the right direction for countermeasures by examining unsafe acts comprehensively rather than limiting the analysis to human errors only. In an analysis of 150 industrial accident cases that caused fatalities and featured relatively clear accident scenarios, the results showed that only 36.0% (54 cases) of the workers recognized the situation they confronted as risky, out of which 29.6% (16 cases) thought of the risk as trivial. In addition, even when the risks were recognized, most workers attempted to solve the hazardous situations in ways that violated rules or regulations. If analyzed with a focus on human errors, accidents can be attributed to personal deviations. However, if considered with an emphasis on safety rules or regulations, the focus will naturally move to the question of whether the workers intentionally violated them or not. As a consequence, failure of managerial efforts may be highlighted. Therefore, it was concluded that management should consider unsafe acts comprehensively, with violations included in principle, during accident investigations and the development of countermeasures to prevent future accidents.

• 한국방재안전학회논문집
• /
• 제16권1호
• /
• pp.71-80
• /
• 2023

연료전지는 저탄소 발전원으로 주유소 내 연료전지를 설치 시 분산에너지와 전기차 충전인프라를 확충할 수 있다. 주유소 내 연료전지 설치 시 안전성 확보를 위하여 국‧내외 주유소 및 연료전지의 사고데이터를 기반으로 사고시나리오를 도출하고 사고피해예측을 위한 정량적 위험성평가를 실시하였다. 최악의 사고시나리오가 아닌 현실적으로 발생 가능한 화재 및 폭발사고의 피해범위를 산출하고, 피해영향을 분석한 결과 주유기로부터 9.0 m, 주유 중 차량으로부터 15.5 m, 통기관으로부터 4.1 m, 연료전지의 정압기 등 가스조정장치로부터 1.1 m 이상 이격 시 사고로 인한 심각한 피해 위험을 낮출 수 있는 것으로 나타났다. 이러한 연구 결과는 주유소 내 연료전지 배치 및 사고피해를 경감할 수 있는 안전대책 수립에 활용할 수 있을 것으로 기대된다.

• 산업과 과학
• /
• 제2권3호
• /
• pp.8-14
• /
• 2023

대다수의 IoT 기기들은 이미 AIoT를 사용하고 있지만, AI 애플리케이션을 구축하기 위해서는 아직 해결해야 할 문제가 많이 남아 있다. 본 연구에서는 IoT 에지 자원을 보다 효과적으로 분산하기 위해 머신러닝 기반의 IoT 에지 자원 관리 기법을 제안한다, 제안 기법은 머신러닝을 이용하여 IoT 에지 자원 동향을 파악함으로써 IoT 자원의 할당을 지속적으로 개선하며, 최적화된 IoT 자원은 머신러닝 컨볼루션을 활용하여 항상 변화하는 IoT 에지 자원을 안정적으로 유지한다, 제안 기법은 각각의 머신러닝 기반 IoT 에지 자원을 이전 패턴의 자원과 함께 해시값으로 저장함으로써 분산된 AIoT 맥락에서 공격 패턴으로 자원을 효과적으로 검증한다. 실험 결과에서는 IoT Edge 리소스의 무결성을 검증하기 위해서 이질적인 계산 하드웨어가 있는 복잡한 환경에서 잘 동작하는지 세 가지 다른 테스트 시나리오에서 에너지 효율성을 평가하였다.

• 대한인간공학회:학술대회논문집
• /
• 대한인간공학회 1997년도 춘계학술대회논문집
• /
• pp.56-60
• /
• 1997

The study of human erroneous actions has traditionally taken place along two different lines of approach. One has been concerned with finding and explaining the causes of erroneous actions, such as studies in the psychology of "error". The other has been concerned with the qualitative and quantitative prediction of possible erroneous actions, exemplified by the field of human reliability analysis (HRA). Another distinction is also that the former approach has been dominated by an academic point of view, hence emphasising theories, models, and experiments, while the latter has been of a more pragmatic nature, hence putting greater emphasis on data and methods. We have been developing a method to make predictions about error modes. The input to the method is a detailed task description of a set of scenarios for an experiment. This description is then analysed to characterise thd nature of the individual task steps, as well as the conditions under which they must be carried out. The task steps are expressed in terms of a predefined set of cognitive activity types. Following that each task step is examined in terms of a systematic classification of possible error modes and the likely error modes are identified. This effectively constitutes a qualitative analysis of the possibilities for erroneous action in a given task. In order to evaluate the accuracy of the predictions, the data from a large scale experiment were analysed. The experiment used the full-scale nuclear power plant simulator in the Halden Man-Machine Systems Laboratory (HAMMLAB) and used six crews of systematic performance observations by experts using a pre-defined task description, as well as audio and video recordings. The purpose of the analysis was to determine how well the predictions matiched the actually observed performance failures. The results indicated a very acceptable rate of accuracy. The emphasis in this experiment has been to develop a practical method for qualitative performance prediction, i.e., a method that did not require too many resources or specialised human factors knowledge. If such methods are to become practical tools, it is important that they are valid, reliable, and robust.