In the nuclear fuel cycle (NFC) facilities, the failure of Heating Ventilation and Air Conditioning (HVAC) system starts with minor component failures and can escalate to affecting the entire system, ultimately resulting in radiological consequences to workers. In the field of air-conditioning and refrigerating engineering, the fault detection and diagnosis (FDD) of HVAC systems have been studied since faults occurring in improper routine operations and poor preventive maintenance of HVAC systems result in excessive energy consumption. This paper aims to provide a systematic review of existing FDD methods for HVAC systems therefore explore its potential application in nuclear field. For this goal, typical faults and FDD methods are investigated. The commonly occurring faults of HVAC are identified through various literature including publications from International Energy Agency (IEA) and American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE). However, most literature does not explicitly addresses anomalies related to pressure, even though in nuclear facilities, abnormal pressure condition need to be carefully managed, particularly for maintaining radiological contamination differently within each zone. To build simulation model for FDD, the whole-building energy system modeling is needed because HVAC systems are major contributors to the whole building’s energy and thermal comfort, keeping the desired environment for occupants and other purposes. The whole-building energy modeling can be grouped into three categories: physics-based modeling (i.e., white-box models), hybrid modeling (i.e., grey-box models), and data-driven modeling (i.e., black-box models). To create a white-box FDD model, specialized tools such as EnergyPlus for modeling can be used. The EnergyPlus is open source program developed by US-DOE, and features heat balance calculation, enabling the dynamic simulation in transient state by heat balance calculation. The physics based modeling has the advantage of explaining clear cause-and-effect relationships between inputs and outputs based on heat and mass transfer equations, while creating accurate models requires time and effort. Creating a black-box FDD model requires a sufficient quantity and diverse types of operational data for machine learning. Since operation data for HVAC systems in existing nuclear cycle facilities are not fully available, so efforts to establish a monitoring system enabling the collection, storage, and management of sensor data indicating the status of HVAC systems and buildings should be prioritized. Once operational data are available, well-known machine learning methods such as linear regression, support vector machines, random forests, artificial neural networks, and recurrent neural networks (RNNs) can be used to classify and diagnose failures. The challenge with black-box models is the lack of access to failure data from operating facilities. To address this, one can consider developing black-box models using reference failure data provided by IEA or ASHRAE. Given the unavailability of operation data from the operating NFC facilities, there is a need for a short to medium-term plan for the development of a physics-based FDD model. Additionally, the development of a monitoring system to gather useful operation data is essential, which could serve both as a means to validate the physics-based model and as a potential foundation for building data-driven model in the long term.

A lot of sensor and control signals is generated by an industrial controller and related internet-of-things in discrete manufacturing system. The acquired signals are such records indicating whether several process operations have been correctly conducted or not in the system, therefore they are usually composed of binary numbers. For example, once a certain sensor turns on, the corresponding value is changed from 0 to 1, and it means the process is finished the previous operation and ready to conduct next operation. If an actuator starts to move, the corresponding value is changed from 0 to 1 and it indicates the corresponding operation is been conducting. Because traditional fault detection approaches are generally conducted with analog sensor signals and the signals show stationary during normal operation states, it is not simple to identify whether the manufacturing process works properly via conventional fault detection methods. However, digital control signals collected from a programmable logic controller continuously vary during normal process operation in order to show inherent sequence information which indicates the conducting operation tasks. Therefore, in this research, it is proposed to a recurrent neural network-based fault detection approach for considering sequential patterns in normal states of the manufacturing process. Using the constructed long short-term memory based fault detection, it is possible to predict the next control signals and detect faulty states by compared the predicted and real control signals in real-time. We validated and verified the proposed fault detection methods using digital control signals which are collected from a laser marking process, and the method provide good detection performance only using binary values.

To identify the cause of the error and maintain the health of system, an administrator usually analyzes event log data since it contains useful information to infer the cause of the error. However, because today’s systems are huge and complex, it is almost impossible for administrators to manually analyze event log files to identify the cause of an error. In particular, as OpenStack, which is being widely used as cloud management system, operates with various service modules being linked to multiple servers, it is hard to access each node and analyze event log messages for each service module in the case of an error. For this, in this paper, we propose a novel message-based log analysis method that enables the administrator to find the cause of an error quickly. Specifically, the proposed method 1) consolidates event log data generated from system level and application service level, 2) clusters the consolidated data based on messages, and 3) analyzes interrelations among message groups in order to promptly identify the cause of a system error. This study has great significance in the following three aspects. First, the root cause of the error can be identified by collecting event logs of both system level and application service level and analyzing interrelations among the logs. Second, administrators do not need to classify messages for training since unsupervised learning of event log messages is applied. Third, using Dynamic Time Warping, an algorithm for measuring similarity of dynamic patterns over time increases accuracy of analysis on patterns generated from distributed system in which time synchronization is not exactly consistent.

본 연구에서는 와이어로프의 국부손상 검색을 위해 누설자속기법을 적용하였다. 와이어로프 구조물에 적용하기 위해 리프트오프의 발생을 최소화한 4채널 누설자속 센서헤드를 제작하였고, 이를 사용하여 와이어로프의 국부손상 검색실험을 수행하였다. 국부손상 검색실험을 위해 와이어로프를 준비하였고, 다양한 원주방향을 가지는 부분 단선 손상들을 발생시켰다. 제작된 자속누설 센서헤드를 이용하여 와이어로프 시편의 자속신호를 스캔하였고, 노이즈의 영향을 최소화하고 자속신호의 해상도를 향상시키고자 자속 신호를 미분하여 순간변화량을 손상 검색에 활용하였다. 객관적인 손상 판단을 위해 각 채널에서 계측된 자속신호를 GEV분포를 이용해 설정된 임계값과 비교하였다. 최종적으로 임계값을 초과한 부분의 길이방향 및 원주방향 위치를 실제 손상과 비교함으로써 본 기법의 국부손상 검색 가능성을 살펴보았다.

Multivariate control charts are widely needed to monitor the production processes in various industry. Among the several multivariate control charts, control chart have been used of the typical technique. The control chart shows a statistic that represents observed variables and monitors the process through the statistic. In this case, the statistic generally have the limit that any variables affect to that statistic. To solve this problem, some studies have been progressed in the meantime. The representative method is to disassemble total statistic into each of the variable value and make a decision the parameters with large values than threshold value as a main cause. However, the means is requested to follow the normal distribution. To settle this problem, the bootstrap technique that don't be needed the probability distribution was introduced in 2011. In this paper, I introduced the detection technique of the fault variables using multiple regression analysis. There are two advantages; First, it is possible to use less samples than the ascertainment technique applying to bootstrap. Second, the technique using the regression analysis is easy to apply to the actual environment because the global threshold value is used.

본 연구에서는 다채널 자속누설 센서를 이용하여 강케이블의 국부손상을 검색하였다. 먼저 자속누설 기법을 고정된 케 이블 구조물에 적용하기 위해 프로토타입의 8채널 자속누설 센서헤드를 제작하였고, 국부손상이 발생한 케이블을 구현하 기 위하여 PVC 파이프에 강케이블을 채워 강케이블 다발 시편을 제작하였고, 케이블 시편 외부 및 내부에 다양한 크기 및 방향을 가지는 국부손상을 단계적으로 발생시켰다. 이와 같이 제작된 강케이블 시편을 대상으로 각 손상단계에서 자속누설 센서헤드를 이용하여 자속신호를 스캔하고 출력전압으로 표현하였다. 이어서 일반극치분포를 이용해 손상유무를 판단할 수 있는 기준이 되어줄 임계값을 설정하였고, 이를 각 채널에서 계측된 자속신호와 비교하여 객관적인 손상판단을 수행하 였다. 또한 케이블 모니터링에 있어 가장 중요한 정보인 손상의 길이방향 위치를 효과적으로 검색하기 위해 모든 채널의 자속값을 합하여 총합값의 형태로 임계값과 함께 나타내었다. 최종적으로 임계값을 초과한 부분의 길이방향 및 원주방향 위치를 실제 손상과 비교함으로써 본 기법의 국부손상 검색 가능성을 살펴보았다.

In order to supply the high-quality electric power, several researches have been conducted. For the high-quality power, it is necessary to inspect the power lines and insulators before the lines or insulators are disconnected or damaged. However, it is not enough to inspect all the power lines by human inspectors. In previous study, a power line inspection robot was developed to investigate the power lines and insulators. According to replace the human operators by the robot, the inspection robot has several advantages: the improved working conditions, machine accuracy, and the prevention of accident. However, the robot has some defects in its heavyweight. In this study, a lightweight inspection robot has been developed using a RP and a vacuum casting. And, this study developed a fault detection monitoring program for the high voltage equipment using a microphone which detects the location of fault and the thermal imaging and CCD cameras which verifies the fault and stores the image, respectively.

There has been a huge progress in semiconductor manufacturing processes such as reduction of the design rule(DR) and development of multi-processes. And, semiconductor industries have steadily extended its business and market share by reducing the design rule(DR) and enlarging the wafer size as well as by resolving many difficult technical problems through various noble approaches in order to reduce the production cost and to improve the yield. In semiconductor manufacturing, there is a significant difference in the number of memory chips produced according to the wafer size, wafer yield, and the level of the design rule even though the same number of wafers were put to the manufacturing process. So, almost all semiconductor manufacturing companies reach the conclusion that the enlarged size of wafer should be adopted in order to enhance the productivity and reduce the production cost. Thus, in this study, we investigate the specifications of the key functions and capabilities of the necessary modules in the yield analysis and improvement system required to acquire the stationary wafer yield with considering the 450mm wafer manufacturing system. Then the results of this research will be helpful for constructing the advanced yield analysis and improvement system called Real-Time Fault Monitoring and Detection (RTFMD) system.

본 논문은 상대적으로 새로운 기법인 Parzen Density Estimation과 Multi-class SVM을 이용한 지능형 고장 탐색과 진단 방법을 제안하고 있다. 본 연구에서는 롤링 베어링을 대상으로 고장을 탐색하고 진단하기 위한 방법을 제안하는데 Parzen Density Estimation과 Multi-class SVM은 고장 클래스를 잘 표현할 수 있다. Parzen Density Estimation은 새로운 패턴 데이터의 거절과 알려진 데

The widespread sensors in a structural monitoring system provide vital support to its operation. Data is obtainedf rom sensors in a structural health monitoring system for integrity assessment of the structure, and false alarm will be frequently triggered if a faulty sensor is detected. In this study, a proposed method based on machine learning algorithm and Gaussian distribution is present to identify sensor fault.

열차운행에 따라 철도레일에 반복적으로 가해지는 높은 하중은 레일에 결함을 발생시키게 되며 결함이 진전될 경우 궁극적으로 레일의 파손을 유발할 수 있다. 레일의 파손은 많은 유지보수 비용을 유발시키며 나아가 열차탈선이라는 안전문제와 직결되어 있어 레일의 결함을 조기에 검출할 수 있는 효율적인 비파괴 검사법(NDT)이 필요하다. 특히, 레일과 같이 길이가 매우 긴 연속체 구조물을 효과적으로 검사하기 위해서는 고속탐상을 통한 시간단축이 실용화를 위한 주요 요구조건이 된다. 이에 본 연구에서는 비접촉으로 고속탐상이 가능한 비파괴 검사기술인 누설자속탐상 (Magnetic Flux Leakage, MFL) 기술의 기반으로 철도레일 결함 탐상을 위한 적용 가능성을 검증하였다. 검증은 다양한 위치와 크기를 갖는 5가지 종류의 결함레일 시편에 자체 제작한 MFL 센서를 적용하여 결함검출 가능여부를 확인하는 방법으로 진행하였다. 시험결과 수직 및 대각선 방향의 국부손상에 대해 자속누설 신호가 잘 검출 되었으며 이를 통해 누설자속탐상 기술의 철도레일 적용 가능성을 확인할 수 있었다. 향후 연구를 통해, 철도레일 단면에 최적화된 센서부 제작 및 고속탐상이 가능하도록 신호처리가 이루어진다면 철도레일 국부결함 검출을 위한 매우 효과적인 비파괴 검사법이 될 것으로 기대된다.

현재 진동 정보를 통해 기계 설비의 상태나 고장 유무를 판단하는 연구들이 다수 진행 중에 있는데, 대부분의 연구에서는 설비에 대한 진동을 모니터링하거나 고장 유무를 판별하여 사용자에게 알리는 수준이다. 본 논문에서는 진동 정보 적용 대상을 선박으로 정하고, 진동에 의한 고장 진단과 판별을 보다 정교하게 수행하는 선박 엔진 감지 기법과 시스템을 제안하였다. 일차적으로 이중화된 진동 정보 판별 기법을 적용하여 진동 정보를 확인한 다음에 고장 유무를 검사한다. 만일 고장이 발생한 경우에는 적분을 이용하여 고장 진동 파형에 대한 넓이를 기준으로 어떤 유형의 고장인지를 판별할 수 있는 기법을 적용하였다. 또한 선박의 진동 경향 분석과 엔진 안전 보존을 목적으로 진동 정보를 데이터베이스에 저장하고 추적할 수 있도록 시스템을 구현하였다. 제안 시스템을 선박 엔진의 고장 판별 유무와 고장 진동 파형 감별 인자에 대해 실험을 수행한 결과 고장 판별은 약 98% 정확성을 가졌고 고장 진동 파형 감별에서는 약 72% 정확성을 가졌다.