• Spatial Information Research
• /
• v.12 no.1
• /
• pp.89-100
• /
• 2004

Nowadays, the gas utilities have been increasing constantly due to the expansion of the urban areas. Using computerized information database, the gas companies have developed a gas management system in order to maintain the current status. However, this system can only give basic functions of the maintenance and management of the gas facilities and it has no proper utilities to provide information against accidents from gas leaks. Therefore, a gas accident management system has been developed in this study. Through primary and secondary pipe searching algorithm realtime based management system was devised against gas leaks to propose proper actions. In addition, supporting decision making has been enabled providing estimated maximum amount of gas leaks. Furthermore, all the residential units could be identified thereby minimizing damages through early warning. This system can be expected to contribute to enhance the efficiency of the gas management not to mention of protecting human lives and properties of the nation.

• Journal of Korean Society of Transportation
• /
• v.25 no.5
• /
• pp.149-160
• /
• 2007

Automobile black boxes store and provide accident and driving information. The accident and driving information can be utilized to build scientific traffic-event database and can be applied in various industries. The objective of this study is to develop a recognition system of dangerous driving through analyzing the driving characteristic patterns. In this paper, possible dangerous driving models are classified into four models on the basis of vehicle behaviors(acceleration, deceleration, rotation) and accident types from existing statistical data. Dangerous driving data have been acquired through vehicle tests using automobile black boxes. Characteristics of driving patterns have been analyzed in order to classify dangerous driving models. For the recognition of dangerous driving, this study selected critical value of each dangerous driving model and developed the recognition algorithm of dangerous driving. The study has been verified by the application of recognition algorithm of dangerous driving and vehicle tests using automobile black boxes. The presented recognition methods of dangerous driving can be used for on-line/off-line management of drivers and vehicles.

• International conference on construction engineering and project management
• /
• 2022.06a
• /
• pp.128-135
• /
• 2022

Because construction work is conducted outdoors, construction workers are affected by harmful environmental factor. Especially, Particulate Matter (PM₁₀) is one of the harmful environmental factors with a diameter of 10㎍/m³ or less. When PM₁₀ is inhaled by human, it can cause fatal impact on the human. Contrary to the various analyses of health impact on PM₁₀, the research on the relationship between construction accidents and PM₁₀ are few. Therefore, this study aims to conduct the relative frequency analysis which find out the correlation between construction accidents and PM₁₀, and the modified PM₁₀ grade is suggested to expect accidents probability caused by PM₁₀ in the construction industry. This study is conducted by four steps. i) Establishment of the database; ii) Classification of data; iii) Analysis of the Relative Frequency of accidents in the construction industry by PM₁₀ concentration; iv) Modified PM₁₀ groups to classify the impact of PM₁₀ on accident. In terms of frequency analysis, the most accidents were occurred in the average concentration of PM₁₀ (32㎍/m³). However, we found that the relative frequency of accident was increased as the concentration of PM₁₀ increased. This means the higher PM₁₀ concentration can cause more accidents during construction. In addition, PM₁₀ concentration was divided as 6 groups by the WHO, but the modified PM₁₀ grade by the relative frequency on accident was suggested as 3 groups.

• International conference on construction engineering and project management
• /
• 2022.06a
• /
• pp.579-586
• /
• 2022

Modular building is a fast-growing construction method, mainly due to its ability to drastically reduce the amount of time it takes to construct a building and produce higher-quality buildings at a more consistent rate. However, while modular construction is relatively safer than traditional construction methods, workers are still exposed to hazards that lead to injuries and fatalities, and these hazards could be controlled using emerging smart technologies. Currently, limited information is available at the intersection of modular construction, safety risk, and smart safety technologies. This paper aims to investigate what aspects of modular construction are most dangerous for its workers, highlight specific risks in its processes, and propose ways to utilize smart technologies to mitigate these safety risks. Findings from the archival analysis of accident reports in Occupational Safety and Health Administration (OSHA) Fatality and Catastrophe Investigation Summaries indicate that 114 significant injuries were reported between 2002 and 2021, of which 67 were fatalities. About 72% of fatalities occurred during the installation phase, while 57% were caused by crushing and 85% of crash-related incidents were caused by jack failure/slippage. IoT-enabled wearable sensing devices, computer vision, smart safety harness, and Augment and Virtual Reality were identified as potential solutions for mitigating identified safety risks. The present study contributes to knowledge by identifying important safety trends, critical safety risk factors and proposing practical emerging methods for controlling these risks.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.22 no.2
• /
• pp.228-232
• /
• 2016

As the quantity of HNS sea trasport and the number of HNS accidents at sea are increasing recently, the importance of HNS management is emphasized so that we try to develop marine accident case standard code for making HNS accidents at sea databased systemically in this study. First and foremost, we draw the related requisites of essential accident reports along with internal and external decrees and established statistics of classified items for conducting study, and we referred to analogous standard codes obtained from developed countries in order to research code design. Code design is set like 'Accident occurrence ${\rightarrow}$ The initial accident information ${\rightarrow}$ Accident response ${\rightarrow}$ Accident investigation' in accordance with the general flow of marine HNS accidents of in which the accident information is input and queried. We classified initial accident information into the items of five categories and constructed "Preliminary Information Code(P.I.C.)". In addition we constructed accident response in two categories and accident investigation in three categories that get possible after the accident occurrence as called "Full Information(F.I.C.)", including the P.I.C. It is represented in 3 kinds of steps on each topic by departmentalizing the classified majority as classified middle class and classified minority. As a result of coding marine HNS accident and of the code to a typical example of marine HNS accident, HNS accident was ascertained to be represented sufficiently well. We expect that it is feasible to predict possible trouble or accident henceforward by applying code, and also consider that it is valuable to the preparedness, response and restoration in relation to HNS accidents at sea by managing systemically the data of marine HNS accidents which will occur in the future.

• Journal of Cerebrovascular and Endovascular Neurosurgery
• /
• v.25 no.3
• /
• pp.275-287
• /
• 2023

Objective: Flow diverting stents (FDS) are a validated device in the treatment of intracranial aneurysms, allowing for minimally invasive intervention. However, after its approval for use in the United States in 2011, post-market surveillance of adverse events is limited. This study aims to address this critical knowledge gap by analyzing the FDA Manufacturer and User Facility Device Experience (MAUDE) database for patient and device related (PR and DR) reports of adverse events and malfunctions. Methods: Using post-market surveillance data from the MAUDE database, PR and DR reports from January 2012-December 2021 were extracted, compiled, and analyzed with R-Studio version 2021.09.2. PR and DR reports with insufficient information were excluded. Raw information was organized, and further author generated classifications were created for both PR and DR reports. Results: A total of 2203 PR and 4017 DR events were recorded. The most frequently reported PR adverse event categories were cerebrovascular (60%), death (11%), and neurological (8%). The most frequent PR adverse event reports were death (11%), thrombosis/thrombus (9%) cerebral infarction (8%), decreased therapeutic response (7%), stroke/cerebrovascular accident (6%), intracranial hemorrhage (5%), aneurysm (4%), occlusion (4%), headache (4%), neurological deficit/dysfunction (3%). The most frequent DR reports were activation/positioning/separation problems (52%), break (9%), device operates differently than expected (4%), difficult to open or close (4%), material deformation (3%), migration or expulsion of device (3%), detachment of device or device component (2%). Conclusions: Post-market surveillance is important to guide patient counselling and identify adverse events and device problems that were not identified in initial trials. We present frequent reports of several types of cerebrovascular and neurological adverse events as well as the most common device shortcomings that should be explored by manufacturers and future studies. Although inherent limitations to the MAUDE database are present, our results highlight important PR and DR complications that can help optimize patient counseling and device development.

• Journal of the Korean Society of Hazard Mitigation
• /
• v.5 no.4 s.19
• /
• pp.9-15
• /
• 2005

This research developed the RMIS(Risk Management Information System) which focus on works of risk management fields required of apply of a space information, and focus on the DB to establish and apply the space information efficiently with research scope on the LPG refueling station in city. On the basis of the RMIS, this research provides the baseline to lead on an efficiency of safety inspection of LPG refueling station, advance risk assessment, and efficient making decision of an accident correspondence assessment with interlocking the GIS representing risk through the automation of a quantitative risk assessment standardize requirement to control at real-time. The RMIS development process is as follows. firstly, Relational Database(RDB) was developed by using fundamental data both On-site and Off-site relating data as peforming risk assessment on the LPG refueling station in city. Second, the risk management integral database system was developed to monitor and control the risk efficiently for user with using the Visual Basic Program. Third, through interlocking the risk management integral database system and the GIS(Falcon-map) was suggested the decision making method. Represented results through out the RMIS program development are as follows. Firstly, the RMIS was established the mutual information to advance management the risk efficiently for user and inspector with using the risk management data. Second, as this study managed risk for on-site and off-site separately and considered effect for inside and outside of facility, constructed the basis on safety management which can respond to major accident. Third, it was composed the baseline to making decision that on the basis of user interface.

• The KIPS Transactions:PartD
• /
• v.8D no.2
• /
• pp.125-131
• /
• 2001

Various kind of disasters happens in our society. Most of them require immediate treatment to save life or to protect valuable products. When an accident happens in a place, it is reported to the headquarter of emergency measures system. According to the nature of accident several treatments orders are transmitted to the related authorities. In this paper, we introduce an intelligent emergency measures system that uses trigger mechanism of active databases. The system responds to various events spontaneously without intervention of mankind by triggering proper rules. The most important part of an action in the system is the capability of searching places to apply adequate treatments quickly. We have developed a new method for range queries and nearest neighbor queries which utilize the z-ordering technique to get fast responses. Those new methods are further extended to handle more realistic actual distance of road among positions.

• Journal of Nuclear Fuel Cycle and Waste Technology(JNFCWT)
• /
• v.17 no.2
• /
• pp.235-243
• /
• 2019

Radioactive materials are released into the air and deposited on the surface soil after a nuclear accident. Radionuclides deposited in soil are transported by precipitation to nearby environments and contaminate the surface water system. Basic data on surface watershed and soil erosion models have been collected and analyzed to evaluate the behavior of radionuclides deposited on surface soil after a nuclear accident. Data acquisition and analysis in aquatic environment were performed to investigate the physical characteristics and variation of biota in rivers and lakes of the Nakdong river area near the Wolsong nuclear power plant. For these purposes, a digital map, and hydrological, water quality and biota data were gathered and a systematic database (DB) was constructed in connection with them. Constructed aquatic DB will be supplied and used in surface watershed and soil erosion models for investigation of long-term movement of radionuclides in adsorptive form in surface soil. Finally, basic data and established models will be utilized for general radiological impact assessment in aquatic environment.

• Journal of the Korean Society of Systems Engineering
• /
• v.18 no.2
• /
• pp.75-93
• /
• 2022

Machine learning (ML) data-driven meta-model is proposed as a surrogate model to reduce the excessive computational cost of the physics-based model and facilitate the real-time prediction of a nuclear power plant's transient response. To forecast the transient response three machine learning (ML) meta-models based on recurrent neural networks (RNNs); specifically, Long Short Term Memory (LSTM), Gated Recurrent Unit (GRU), and a sequence combination of Convolutional Neural Network (CNN) and LSTM are developed. The chosen accident scenario is a control element assembly withdrawal at power concurrent with the Loss Of Offsite Power (LOOP). The transient response was obtained using the best estimate thermal hydraulics code, MARS-KS, and cross-validated against the Design and control document (DCD). DAKOTA software is loosely coupled with MARS-KS code via a python interface to perform the Best Estimate Plus Uncertainty Quantification (BEPU) analysis and generate a time series database of the system response to train, test and validate the ML meta-models. Key uncertain parameters identified as required by the CASU methodology were propagated using the non-parametric Monte-Carlo (MC) random propagation and Latin Hypercube Sampling technique until a statistically significant database (181 samples) as required by Wilk's fifth order is achieved with 95% probability and 95% confidence level. The three ML RNN models were built and optimized with the help of the Talos tool and demonstrated excellent performance in forecasting the most probable NPP transient response. This research was guided by the Systems Engineering (SE) approach for the systematic and efficient planning and execution of the research.