• Nuclear Engineering and Technology
• /
• v.55 no.4
• /
• pp.1528-1539
• /
• 2023

Since the leakage of sodium in an SFR (sodium-cooled fast reactor) causes an explosion upon reaction with air and water, sodium leakages represent an important safety issue. In this study, a novel technique for improving the reliability of sodium leakage detection applying DDVR (dynamic data validation and reconciliation) is proposed and verified to resolve this technical issue. DDVR is an approach that aims to improve the accuracy of a target system in a dynamic state by minimizing random errors, such as from the uncertainty of instruments and the surrounding environment, and by eliminating gross errors, such as instrument failure, miscalibration, or aging, using the spatial redundancy of measurements in a physical model and the reliability information of the instruments. DDVR also makes it possible to estimate the state of unmeasured points. To validate this approach for supporting sodium leakage detection, this study applies experimental data from a sodium leakage detection experiment performed by the Korea Atomic Energy Research Institute. The validation results show that the reliability of sodium leakage detection is improved by cooperation between DDVR and hardware measurements. Based on these findings, technology integrating software and hardware approaches is suggested to improve the reliability of sodium leakage detection by presenting the expected true state of the system.

• Transactions of the Korean Society of Pressure Vessels and Piping
• /
• v.19 no.2
• /
• pp.93-101
• /
• 2023

In this study, a deep learning algorithm was used to diagnose electric potential signals obtained through CIPS and DCVG, used indirect inspection methods to confirm the soundness of buried pipes. The deep learning algorithm consisted of CNN(Convolutional Neural Network) model for diagnosing the electric potential signal and Grad CAM(Gradient-weighted Class Activation Mapping) for showing the flaw prediction point. The CNN model for diagnosing electric potential signals classifies input data as normal/abnormal according to the presence or absence of flaw in the buried pipe, and for abnormal data, Grad CAM generates a heat map that visualizes the flaw prediction part of the buried pipe. The CIPS/DCVG signal and piping layout obtained from the 3D finite element model were used as input data for learning the CNN. The trained CNN classified the normal/abnormal data with 93% accuracy, and the Grad-CAM predicted flaws point with an average error of 2m. As a result, it confirmed that the electric potential signal of buried pipe can be diagnosed using a CNN-based deep learning algorithm.

• Journal of Advanced Marine Engineering and Technology
• /
• v.27 no.1
• /
• pp.42-48
• /
• 2003

In general many engineers can diagnose the fault condition using the abnormal ones among data monitored from a diesel engine, but they don't need the system modelling or identification for the work. They check the abnormal data and the relationship and then catch the fault condition of the engine. This paper proposes the construction of a fault diagnosis engine through malfunction data gained from the data fault detection system of neural networks for diesel generator engine, and the rule inference method to induce the rule for fuzzy inference from the malfunction data of diesel engine like a site engineer with a fuzzy system. The proposed fault diagnosis system is constructed in the sense of the Malfunction Diagnosis Engine(MDE) and Hierarchy of Malfunction Hypotheses(HMH). The system is concerned with the rule reduction method of knowledge base for related data among the various interactive data.

• Journal of information and communication convergence engineering
• /
• v.20 no.4
• /
• pp.295-302
• /
• 2022

This paper contains the development of a smart power device designed to collect load power data from industrial manufacturing machines, predict future variations in load power data, and detect abnormal data in advance by applying a machine learning-based prediction algorithm. The proposed load power data prediction model is implemented using a Long Short-Term Memory (LSTM) algorithm with high accuracy and relatively low complexity. The Flask and REST API are used to provide prediction results to users in a graphical interface. In addition, we present the results of experiments conducted to evaluate the performance of the proposed approach, which show that our model exhibited the highest accuracy compared with Multilayer Perceptron (MLP), Random Forest (RF), and Support Vector Machine (SVM) models. Moreover, we expect our method's accuracy could be improved by further optimizing the hyperparameter values and training the model for a longer period of time using a larger amount of data.

• Ocean and Polar Research
• /
• v.31 no.4
• /
• pp.339-347
• /
• 2009

It is possible to obtain accurate temperature and salinity profiles of the oceans using a SBE 911plus CTD and accompanying data conversion packages. To obtain highly accurate results, CTD data needs to be carefully processed in addition to proper and regular maintenance of the CTD itself. Since the manufacturer of the CTD provides tools that are necessary for post processing, it is possible to conduct proper processing without too much effort. Some users, however, are not familiar with all of the processes and inadvertently ignore some of these processes at the expense of data quality. To draw attention to these and other similar issues, we show how it is possible to improve data quality by utilizing a few extra processes to the standard or default data process procedures with CTD data obtained from the equatorial Eastern Pacific between 2001 and 2005, and 2007. One easy step that is often ignored in the standard data process procedure is "wild edit", which removes abnormal values from the raw data. If those abnormal values are not removed, the abnormality could spread vertically during subsequent processes and produce abnormal salinity in a range much wider than that of the raw data. To remove spikes in salinity profiles the "align CTD" procedure must be carried out not with the default values included in the data processing software but with a proper time constant. Only when "cell thermal mass" correction is conducted with optimal parameters, we can reduce the difference between upcast and downcast, and obtain results that can satisfy the nominal accuracy of the CTD.

• Journal of the Korean Society for Precision Engineering
• /
• v.22 no.3 s.168
• /
• pp.123-129
• /
• 2005

The unbalance of the power output, noise, and vibration is happened by the disproportionate pressure variation in the cylinder. For this reason, decrease of the pressure in the cylinder and increase of the residual gas effect on the engine performance. If the abnormal combustion is continued, the crack would be occurred in the engine block. And it could be broken down. For the normal combustion of the SI engine, it is important to supply the balanced mixture by each operating condition. In this study, it was tested the combustion characteristics in the cylinder according to the abnormal adjusting of intake & exhaust valve. This test is willing to set a basic data's analysis fur developing an automotive diagnosis system by analyzing the pressure in the cylinder, the output signal of MAP sensor, the exhaust gas, etc.

• Proceedings of the IEEK Conference
• /
• 2008.06a
• /
• pp.669-670
• /
• 2008

In this paper, we present a real-time system to detect abnormal events on gas pipes, based on the signals which are observed through the audio sensors attached on them. First, features are extracted from these signals so that they are robust to noise and invariant to the distance between a sensor and a spot at which an abnormal event like an attack on the gas pipes occurs. Then, a classifier is constructed to detect abnormal events using neural networks. It is a combination of two neural network models, a Gaussian mixture model and a multi-layer perceptron, for the reduction of miss and false alarms. The former works for miss alarm prevention and the latter for false alarm prevention. The experimental result with real data from the actual gas system shows that the proposed system is effective in detecting the dangerous events in real-time with an accuracy of 92.9%.

• Journal of the Korean Society of Manufacturing Technology Engineers
• /
• v.8 no.3
• /
• pp.76-82
• /
• 1999

This study was carried out to verify finding performance of dental diamond bur and investigate the possibility of AE application in density field. Work pieces were made of acryl and bovine respectively for the experiments in this study. Grinding test was conducted to get the data of grinding resistance and specific finding energy of low different types of diamond bur by using tool dynamometer. AE signal was acquired to verify grinding process in the AE measuring system. AErms value was increased as the grinding velocity and depth were increasing, but it decreased as the feed rate was increasing. The case of the small value of AE signal is due to abnormal grinding in D type diamond bur. By analyzing AErms start and finish time of grinding working, abnormal grinding state can be confined. Abnormal state can be found through the behavior of AE signal in the finding working. As a result, it is expected that forecast of abnormal state is possible using AE equipments under real time process.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.5 no.2
• /
• pp.313-329
• /
• 2011

To provide a seamless network to customers, Internet service providers must promptly detect and control abnormal traffic. One approach is to shorten the traffic information measurement cycle. However, performance degradation is inevitable if traffic measurement servers merely shorten the cycle and measure all traffic. This paper presents a software architecture that can measure traffic more frequently without degrading performance by estimating the level of abnormal traffic. The algorithm in the architecture estimates the values of the interface group objects in MIB by using the IP group objects thereby reducing the number of measurements and the size of measured data. We evaluated this architecture on part of Internet service provider's IP network. When the traffic was measured 5 times more than before, the CPU usage and TPS of the proposed scheme was 7% and 41% less than that of the original scheme while the false positive rate and false negative rate were 3.2% and 2.7% respectively.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.634-636
• /
• 2023

Surveillance systems play a pivotal role in ensuring the safety and security of various environments, including public spaces, critical infrastructure, and private properties. However, detecting abnormal human behavior in lowlight conditions is a critical yet challenging task due to the inherent limitations of visual data acquisition in such scenarios. This paper introduces a spatiotemporal framework designed to address the unique challenges posed by low-light environments, enhancing the accuracy and efficiency of human abnormality detection in surveillance camera systems. We proposed the pre-processing using lightweight exposure correction, patched frames pose estimation, and optical flow to extract the human behavior flow through t-seconds of frames. After that, we train the estimated-action-flow into autoencoder for abnormal behavior classification to get normal loss as metrics decision for normal/abnormal behavior.