• Journal of the Korean Society of Safety
• /
• v.30 no.1
• /
• pp.87-93
• /
• 2015

Most incidents and accidents involved human during operating NPPs have a tendency to be structured by complicated and various organizational, individual, and environmental factors. The salient feature of the human error in NPP was extremely low frequency, extremely high complicated and extremely serious damage of human life and property. Our research team defined as 'rare human errors'. To prevent the rare human errors, the most researchers and analysts insist invariably that the root causes be made clear. The making them clear, however, is difficult because their root causes are very various and uncertain. However, These tools have limits that they do not adapt all operating situations and circumstances such as design base events. The purpose of this study is to improve the rare human error hazards consider the situational contex. Through this challenging try based on evidences to the human errors could be useful to prevent rare and critical events can occur in the future.

• Nuclear Engineering and Technology
• /
• v.55 no.2
• /
• pp.603-622
• /
• 2023

Human error (HE) is an important concern in safety-critical systems such as nuclear power plants (NPPs). HE has played a role in many accidents and outage incidents in NPPs. Despite the increased automation in NPPs, HE remains unavoidable. Hence, the need for HE detection is as important as HE prevention efforts. In NPPs, HE is rather rare. Hence, anomaly detection, a widely used machine learning technique for detecting rare anomalous instances, can be repurposed to detect potential HE. In this study, we develop an unsupervised anomaly detection technique based on generative adversarial networks (GANs) to detect anomalies in manually collected surveillance data in NPPs. More specifically, our GAN is trained to detect mismatches between automatically recorded sensor data and manually collected surveillance data, and hence, identify anomalous instances that can be attributed to HE. We test our GAN on both a real-world dataset and an external dataset obtained from a testbed, and we benchmark our results against state-of-the-art unsupervised anomaly detection algorithms, including one-class support vector machine and isolation forest. Our results show that the proposed GAN provides improved anomaly detection performance. Our study is promising for the future development of artificial intelligence based HE detection systems.

• Nuclear Engineering and Technology
• /
• v.37 no.2
• /
• pp.151-158
• /
• 2005

Advanced nuclear power plants are generally large complex systems automated by computers. Whenever a rare plant emergency occurs the plant operators must cope with the emergency under severe mental stress without committing any fatal errors. Furthermore, The operators must train to improve and maintain their ability to cope with every conceivable situation, though it is almost impossible to be fully prepared for an infinite variety of situations. In view of the limited capability of operators in emergency situations, there has been a new approach to preventing the human error caused by improper human-machine interaction. The new approach has been triggered by the introduction of advanced information systems that help operators recognize and counteract plant emergencies. In this paper, the adverse effect of automation in human-machine systems is explained. The discussion then focuses on how to configure a joint human-machine system for ideal human-machine interaction. Finally, there is a new proposal on how to organize technologies that recognize the different states of such a joint human-machine system.

• International Journal of Contents
• /
• v.13 no.4
• /
• pp.23-28
• /
• 2017

Of the total economic loss caused by disasters, 40% are due to floods and floods have a severe impact on human health and life. So, it is important to monitor the water level of a river and to issue a flood warning during unfavorable circumstances. In this paper, we propose a modified error function to improve a hydrological modeling using a multi-layer perceptron (MLP) neural network. When MLP's are trained to minimize the conventional mean-squared error function, the prediction performance is poor because MLP's are highly tunned to training data. Our goal is achieved by preventing overspecialization to training data, which is the main reason for performance degradation for rare or test data. Based on the modified error function, an MLP is trained to predict the water level with rainfall data at upper reaches. Through simulations to predict the water level of Nakdong River near a UNESCO World Heritage Site "Hahoe Village," we verified that the prediction performance of MLP with the modified error function is superior to that with the conventional mean-squared error function, especially maximum error of 40.85cm vs. 55.51cm.

• Journal of the Korean Association of Oral and Maxillofacial Surgeons
• /
• v.25 no.4
• /
• pp.295-303
• /
• 1999

Presented in this paper are the experimental results that measure rapid prototyping (RP) errors in 3D medical models. We identified various factors that can cause dimensional errors when producing RP models, specifically in maxillofacial areas. For the experiment, we used a human dry skull. A number of linear measurements based on landmarks were first obtained on the skull. This was followed by CT scanning, 3D model reconstruction, and RP model fabrication. The landmarks were measured again on both the reconstructed models and the physical RP models, and these were compared with those on dry skull. We focused on major sources of errors, such as CT scanning, conversion from CT data to STL models, and RP model fabrication. The results show that the overall error from skull to RP is $0.64{\times}0.36mm(0.71{\times}0.66%)$ in absolute value. This indicates that the RP technology can be acceptable in the real clinical applications. A clinical case that has applied RP models successfully for treatment planning and surgical rehearsal is presented. Although the use of RP models is rare in the medical area yet, we believe RP is promising in that it has a great potential in developing new tools which can aid diagnosis, treatment planning, surgical rehearsal, education, and so on.

• Journal of the Korean Institute of Oriental Medical Informatics
• /
• v.20 no.1
• /
• pp.25-156
• /
• 2014

This thesis is consisted of studying of the medical literature about Five sense of organs. Five sense of organs are the eyes, tongue, mouth, nose and ears. Five sense of organs are performed human senses which external sensory information by accepting an important feature for maintaining the biological activity to be performed. The contents was compiled up to the Donguibogam to Chinese literature and documents encompass the Korea medical literature, Donguibogam related to the senses to identify the citation of each chapter, the actual quotation through doctrine and other publications revealed that the citation is to investigate how accurately identified through studying the analysis and observation. It is as following as I observed carefully the senses of Donguibogam quotations related to each other through doctrine and publishment institution follows in order of dynasties. There are four volumes of Han-dynasty, one volume of Weijinnanbei-Era, two volumes of Tang-dynasty, nineteen volumes of SongJinYuan-dynasty, seven volumes of Ming-dynasty as Chinese medical literature. There are four volumes of Chosun-dynasty as Korean medical literature. It is the most quotation publishment that the books of SongJinYuan-dynasty of above thirty-six-volume. It is the latest quotation book that is Gujinyigan in Chinese medical literature and Euirimchwalyo in Korean medical literature. It is very positive quotation considering even Donguibogam publishment year in 1613. The reference books are four volumes of Chosun-dynasty as Korean medical literature and thirty-two-volume of Chinese medical literature. By observing the quotation frequency, 157 times in Sheyideaiofang, 115 times in Yixuerumen, 74 times in Yixuegangmu, 39 times in Wanbinghuichun, 31 times in Euibangryuchwi, 30 times in Renzhezhizhifang and Gujinyigan, 28 times in Danxixinfafuyu, 23 times Hwangdineijing, 17 times in Nanshibizang and Yixuezhengchuan. Other else books have been cited less than 10 times. It might be made error that did not find the source of the books even though cited reference, also even though defining the source of reference it is only rare reference book. As mention above, there are a lot of discovering as the feature of reference Publications. Most of all we could find out the reference literature cited in Donguibogam, however we couldn't clarify other books in original books. Thus, we should remember that it did not coincide with cited marks when studying the Donguibogam.