• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.4
• /
• pp.14-25
• /
• 2018

In recent years, it has become technically easier to explain factors related with traffic accidents in the Big Data era. Therefore, it is necessary to apply the latest analysis techniques to analyze the traffic accident data and to seek for new findings. The purpose of this study is to compare the predictive performance of the negative binomial regression model and the deep learning method developed in this study to predict the frequency of traffic accidents in expressways. As a result, the MOEs of the deep learning model are somewhat superior to those of the negative binomial regression model in terms of prediction performance. However, using a deep learning model could increase the predictive reliability. However, it is easy to add other independent variables when using deep learning, and it can be expected to increase the predictive reliability even if the model structure is changed.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.8
• /
• pp.301-310
• /
• 2021

The TATI model is a Traffic Accident Text to RGB Image model, which is a methodology proposed in this paper for predicting the severity of traffic accidents. Traffic fatalities are decreasing every year, but they are among the low in the OECD members. Many studies have been conducted to reduce the death rate of traffic accidents, and among them, studies have been steadily conducted to reduce the incidence and mortality rate by predicting the severity of traffic accidents. In this regard, research has recently been active to predict the severity of traffic accidents by utilizing statistical models and deep learning models. In this paper, traffic accident dataset is converted to color images to predict the severity of traffic accidents, and this is done via CNN models. For performance comparison, we experiment that train the same data and compare the prediction results with the proposed model and other models. Through 10 experiments, we compare the accuracy and error range of four deep learning models. Experimental results show that the accuracy of the proposed model was the highest at 0.85, and the second lowest error range at 0.03 was shown to confirm the superiority of the performance.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.29 no.1
• /
• pp.72-75
• /
• 2006

Event trees are widely used graphical tool to denote the accident inintiation and escalation to more severe accident. But they have some drawbacks in that they do not have efficient way of updating model parameters and also they can not contain the information about dependency or independency among model parameters. A tool that can cure such drawbacks is an influence diagram. We introduce influence diagrams and explain how to update model parameters and obtain predictive distributions. We show that an event tree can be converted to a statistically equivalent influence diagram, and bayesian prediction can be made more effectively through the use of influence diagrams.

• Journal of Ocean Engineering and Technology
• /
• v.36 no.5
• /
• pp.295-302
• /
• 2022

The initial response to a marine accident can play a key role to minimize the accident. Therefore, various decision support systems have been developed using sensors, simulations, and active response equipment. In this study, we developed an algorithm to predict damage locations using ship motion data with bidirectional long short-term memory (BiLSTM), a type of recurrent neural network. To reflect the low frequency ship motion characteristics, 200 time-series data collected for 100 s were considered as input values. Heave, roll, and pitch were used as features for the prediction model. The F1-score of the BiLSTM model was 0.92; this was an improvement over the F1-score of 0.90 of a prior model. Furthermore, 53 of 75 locations of damage had an F1-score above 0.90. The model predicted the damage location with high accuracy, allowing for a quick initial response even if the ship did not have flood sensors. The model can be used as input data with high accuracy for a real-time progressive flooding simulator on board.

• Journal of Electrical Engineering and Technology
• /
• v.12 no.2
• /
• pp.969-980
• /
• 2017

Galloping is one of the most serious vibration problems in transmission lines. Power lines can be extensively damaged owing to aerodynamic instabilities caused by ice accretion. In this study, the accident probability induced by galloping phenomenon was analyzed using logistic regression analysis. As former studies have generally concluded, main factors considered were local weather factors and physical factors of power delivery systems. Since the number of transmission towers outnumbers the number of weather observatories, interpolation of weather factors, Kriging to be more specific, has been conducted in prior to forming galloping accident estimation model. Physical factors have been provided by Korea Electric Power Corporation, however because of the large number of explanatory variables, variable selection has been conducted, leaving total 11 variables. Before forming estimation model, with 84 provided galloping cases, 840 non-galloped cases were chosen out of 13 billion cases. Prediction model for accidents by galloping has been formed with logistic regression model and validated with 4-fold validation method, corresponding AUC value of ROC curve has been used to assess the discrimination level of estimation models. As the result, logistic regression analysis effectively discriminated the power lines that experienced galloping accidents from those that did not.

• Nuclear Engineering and Technology
• /
• v.53 no.8
• /
• pp.2547-2555
• /
• 2021

The frequency of reactor coolant leakage is expected to increase over the lifetime of a nuclear power plant owing to degradation mechanisms, such as flow-acceleration corrosion and stress corrosion cracking. When loss of coolant accidents (LOCAs) occur, several parameters change rapidly depending on the size and location of the cracks. In this study, leak flow during LOCAs is predicted using a deep fuzzy neural network (DFNN) model. The DFNN model is based on fuzzy neural network (FNN) modules and has a structure where the FNN modules are sequentially connected. Because the DFNN model is based on the FNN modules, the performance factors are the number of FNN modules and the parameters of the FNN module. These parameters are determined by a least-squares method combined with a genetic algorithm; the number of FNN modules is determined automatically by cross checking a fitness function using the verification dataset output to prevent an overfitting problem. To acquire the data of LOCAs, an optimized power reactor-1000 was simulated using a modular accident analysis program code. The predicted results of the DFNN model are found to be superior to those predicted in previous works. The leak flow prediction results obtained in this study will be useful to check the core integrity in nuclear power plant during LOCAs. This information is also expected to reduce the workload of the operators.

• International Journal of Highway Engineering
• /
• v.15 no.6
• /
• pp.113-123
• /
• 2013

PURPOSES : Although numerous researches have been studied to reveal accident causations for road intersections, there are still many research gaps for road segments. It is mainly because of difficulty of data and lack of analytical method. This study aims to study accident causations for rural road segments and develop accident modification factors for safety evaluation. The accident modification factors can be used to improve road safety. METHODS : Methods for developing AMF are diverse. This study developed AMFs using accident prediction models and selected explanatory variables from the accident models. In order to select final AMFs, three different methods were applied in the study. RESULTS : As a result of the study, many AMFs such as horizontal curves or vertical curves were developed and explained the meanings of the results. CONCLUSIONS : This study introduced meaningful methods for developing significant AMFs and also showed several AMFs. It is expected that traffic or road engineers will be able to use the AMFs to improve road segment safety.

• Journal of the Korean Society of Safety
• /
• v.34 no.5
• /
• pp.144-152
• /
• 2019

According to the National Police Agency, point-to-point speed enforcement system is being installed and operated in 97 sections across the country. It is more effective than other enforcement systems in terms of stabilizing the traffic flow and inhibiting the kangaroo effect. But it is only 5.1% of the total enforcement systems. The National Police Agency is also aware that its operation ratio is very low and it is necessary to expand point-to-point speed enforcement system. Hence, this study aims to provide the expansion basis of the point-to-point speed enforcement operation through analysis of the quantitative effects and development the accident prediction model. Firstly, this study analyzed the effectiveness of point-to-point speed enforcement system. Naive before-after study and comparison group method(C-G Method) were used as methodologies of analyzing the effectiveness. The result of using the naive before-after study was significant. Total accidents, EPDOs and casualty crashes decreased by 42.15%, 70.64% and 45.30% respectively. And average speed and the ratio of exceeding speed limit decreased by 6.92% and 20.50%p respectively. Moreover, using the C-G method total accidents, EPDOs and casualty crashes decreased by 31.35%, 66.62% and 10.04% respectively. And average speed and the ratio of exceeding speed limit decreased by 3.49% and 56.65%p respectively. Secondly, this study developed a prediction model for the probability of casualty crash. It was dependant on factors of traffic volume, ratio of exceeding speed limit, ratio of heavy vehicle, ratio of curve section, and presence of point-to-point speed enforcement. Finally, this study selected the most danger sections to the major highway and evaluated proper installation sections to the recent installation section by applying the accident prediction model. The results of this study are expected to be useful in establishing the installation standards for the point-to-point speed enforcement system.

• Journal of the Korean Institute of Gas
• /
• v.18 no.1
• /
• pp.8-16
• /
• 2014

In this study, the number of gas accidents prediction model was suggested by analyzing the gas accidents occurred in Korea. In order to predict the number of gas accidents, simple moving average method (3, 4, 5 period), weighted average method and exponential smoothing method were applied. Study results of the sum of mean-square error acquired by the models of moving average method for 4 periods and weighted moving average method showed the highest value of 44.4 and 43 respectively. By developing the number of gas accidents prediction model, it could be actively utilized for gas accident prevention activities.

• Journal of the Korean Institute of Gas
• /
• v.22 no.2
• /
• pp.29-33
• /
• 2018

Accident likelihood is growing due to a correlation for gas and electricity installed in the area of dense energy consumption like traditional market and underground shopping center. In order to prevent and respond accident risks related to gas and electricity in this area, it should be monitored and predicted for factors of gas leak or electricity by developing safety management system. This study is about accident prediction model development considering fire risk factor related to gas accident. The temperature variation characteristic near a gas burner was analyzed. Also, accident prediction algorithm and related module were developed to prevent fire in the area of dense energy consumption.