Real Accident Data Research Articles

Enhancing maritime transportation security: A data-driven Bayesian network analysis of terrorist attack risks.

Maritime terrorist accidents have a significant low-frequency-high-consequence feature and, thus, require new research to address the associated inherent uncertainty and the scarce literature in the field. This article aims to develop a novel method for maritime security risk analysis. It employs real accident data from maritime terrorist attacks over the past two decades to train a data-driven Bayesian network (DDBN) model. The findings help pinpoint key contributing factors, scrutinize their interdependencies, ascertain the probability of different terrorist scenarios, and describe their impact on different manifestations of maritime terrorism. The established DDBN model undergoes a thorough verification and validation process employing various techniques, such as sensitivity, metrics, and comparative analyses. Additionally, it is tested against recent real-world cases to demonstrate its effectiveness in both retrospective and prospective risk propagation, encompassing both diagnostic and predictive capabilities. These findings provide valuable insights for the various stakeholders, including companies and government bodies, fostering comprehension of maritime terrorism and potentially fortifying preventive measures and emergency management.

Applicability of an Ionising Radiation Measuring System for Real-Time Effective-Dose-Optimised Route Finding Solution during Nuclear Accidents

The reduction in the effective dose of evacuated injured persons through contaminated areas of nuclear accidents is an essential emergency services requirement. In this context, there appeared a need to develop a dose-optimised route finding method for firefighting rescue vehicles, which includes the development of a real-time decision support measurement and evaluation system. This determines and visualises the radiation exposure of possible routes in a tested area. The system inside and outside of the vehicle measures the ambient dose equivalent rate, the gamma spectra, and also the airborne radioactive aerosol and iodine levels. The method uses gamma radiation measuring NaI(Tl) scintillation detectors mounted on the outside of the vehicle, to determine the dose rate inside the vehicle using the previously recorded attenuation conversation function, while continuously collecting the air through a filter and using an alpha, beta, and gamma radiation measuring NaI(Tl)+ PVT + ZnS(Ag) scintillator to determine the activity concentration in the air, using these measured values to determine the effective dose for all routes and all kinds of vehicles. The energy-dependent shielding effect of the vehicle, the filtering efficiency of the collective protection equipment, and the vehicle’s speed and travel time were taken into account. The results were validated by using gamma point sources with different activity and energy levels. The measurement results under real conditions and available real accident data used in our simulations for three different vehicles and pedestrians proved the applicability of the system. During a nuclear accident based on our model calculations, the inhalation of radioactive aerosols causes a dose almost an order of magnitude higher than the external gamma radiation caused by the fallout contamination. The selection of the appropriate vehicle and its route is determined by the spectrum that can be measured at the accident site but especially by the radioactive aerosol concentration in the air that can be measured in the area. In the case of radiation measuring detectors, the shielding effect of the carrier vehicle must be taken into account, especially in the case of heavy shielding vehicles. The method provides an excellent opportunity to reduce the damage to the health of accident victims and first responders during rescue operations.

A study on the cyclist lower extremity injury risk by using German in-depth accident data

The lower extremities of cyclists are the most vulnerable body parts in vehicle-bicycle traffic accidents. However, few studies have been conducted on the effects of vehicle front-end profile parameters on cyclist lower extremity injuries by using real accident data. This paper aimed to explore the correlation between the cyclist’s severe lower extremity injury (AIS2+) risk and impact speed, vehicle front-end profile variables, and cyclist parameters via using real cyclist accidents. A subsample of 390 cyclist accident cases with lower extremity injuries was selected from the German In-Depth Accident Study (GIDAS) to conduct statistical analysis. Furthermore, an independent samples t-test was performed to analyse the associations between the selected predictive parameters and the risk of cyclists with AIS2+ lower extremity injuries. Finally, the risk function for cyclists with AIS2+ lower extremity injuries was developed by using the logistic regression method. The research results indicated that vehicle impact speed, lower bumper height (LBH), bumper depth (BD), cyclist age, and weight were statistically significant factors for cyclists with AIS2+ lower extremity injuries. The higher vehicle speed, higher LBH and BD corresponded to a greater likelihood of the cyclist suffering an AIS2+ lower extremity injury. These findings contribute to developing more effective safety measures for vehicle front designs and establishing regulations to protect the lower extremities of cyclists.

Adversarial learning for robust battery thermal runaway prognostic of electric vehicles

The thermal runaway risk in batteries is a critical issue affecting their safety. Although considerable research is dedicated towards the thermal runaway prognostic of batteries, a common challenge remains that how to reduce the false positive rate to enable practical real world application. To address this problem, this paper proposes a robust thermal runaway prognostic method for Li-ion batteries using a generative adversarial network. Specifically, the model utilizes the charging curves of batteries as features during training to reduce the randomness of sparse data and enhance the robustness of the model. A resampling algorithm is proposed to realize the standardized conversion of the charging curve to retain the complete charging information. By employing a one-class generative adversarial network, the method generates a detection model that provides a reference normal curve for the original battery charging process, detecting potential anomalies. Real accident data of electric vehicles that experienced thermal runaway and subsequent fires is used to verify the performance of the model. Experimental results show that the proposed method can identify all abnormal cells with a false positive rate of 1.75% before thermal runaway occurs, reducing 7.54% to 31.18% false positive rate compared to existing methods.

Carbon fiber composite core conductor operation risk assessment model based on various types of monitoring data

Abstract As the national economy grows gradually and the scale of urbanization expands, the standard for power transmission is also continuously increasing. Currently, the existing transmission lines can no longer meet the rising power demand, thus urgently requiring the construction of power transmission capacity expansion. Although there are many strategies for increasing capacity, the development of enhanced conductor material is indeed its greatest advantage. However, among the various options for increasing capacity, the development of capacity-enhancing wires is also the most advantageous. Carbon fiber conductors, which are known as “black gold,” are the perfect material for this purpose. Given the inherent economic value, high brittleness, and complexity of the installation process of carbon fiber materials, this study proposes an operational risk assessment model for carbon fiber composite core conductors based on multiple monitoring data. For the first time, the influence between quarterly lightning strike probability and arc droop as well as the temperature drop effect on overhead transmission conductors by convective heat dissipation from natural wind are taken into account. The expected arc sag size of the conductor in the future period can be predicted according to the meteorological forecast, which can effectively prevent the impact of flashovers and wire breakage on the operation of the power grid. In addition, the practicality and functionality of the assessment model have been verified by real accident data.

Causative analysis of freight railway accident in specific scenes using a data-driven Bayesian network

As the freight railway system is a typical complex system, freight railway accidents have various and complex accident scenes. A Data-Driven Bayesian Network (DDBN) with random variables representing scene elements, accident causes, and accident consequences as nodes, and conditional dependencies between nodes as edges is proposed to identify the most significant accident causes in various and complex specific accident scenes. First, an unsupervised-supervised method is designed to define the states of nodes in the DDBN, considering the characteristics of the data involving both continuous and discrete states. Second, a greedy algorithm is proposed to mine the causal sequence between nodes, and the direction of edges in DDBN is established accordingly. Then, an NB-K2-MLE approach is proposed to generate the structure and parameters of the DDBN from data. Finally, a risk calculation function based on DDBN is proposed to calculate the risk of various accident causes in specific scenes. In empirical analysis based on real accident data, the inference accuracy of DDBN reached 87.92 %, with precision and recall exceeding 70 %. More importantly, the research results indicate that the distribution of accident causes will concentrate with the refinement and specific of the scene, and the main accident causes at the macro level cannot be fully applicable to accident prevention in specific scenes. The DDBN constructed in this study can provide data support for the determination of the significant accident causes and the development of targeted accident prevention strategies in various and complex specific scenes.

An integrated model for evaluating the leakage risk of urban gas pipe: a case study based on Chinese real accident data

An integrated model for evaluating the leakage risk of urban gas pipe: a case study based on Chinese real accident data

Deep representation of imbalanced spatio‐temporal traffic flow data for traffic accident detection

AbstractAutomatic detection of traffic accidents has a crucial effect on improving transportation, public safety, and path planning. Many lives can be saved by the consequent decrease in the time between when the accidents occur and when rescue teams are dispatched, and much travelling time can be saved by notifying drivers to select alternative routes. This problem is challenging mainly because of the rareness of accidents and spatial heterogeneity of the environment. This paper studies deep representation of loop detector data using long‐short term memory (LSTM) network for automatic detection of freeway accidents. The LSTM‐based framework increases class separability in the encoded feature space while reducing the dimension of data. The experiments on real accident and loop detector data collected from the Twin Cities Metro freeways of Minnesota demonstrate that deep representation of traffic flow data using LSTM network has the potential to detect freeway accidents in less than 18 min with a true positive rate of 0.71 and a false positive rate of 0.25 which outperforms other competing methods in the same arrangement.

Adverse weather impact on driver performance in the UAE

Road traffic accident is a significant cause of death around the world. Many factors cause accidents, such as demographic factors, humane factors, road design factors, and weather factors. This paper will investigate the adverse weather impact on driver performance by analyzing the impact of weather conditions on drivers' behaviors. Statistical analysis was conducted using one-way ANOVA test, correlation analysis, and Artificial Neural Network. The data was collected using a questionnaire for the people live in the United Arab Emirates. The survey includes two parts; the first part is about driver characteristics (Gender, Age, Nationality, Marital status, Education level, Monthly income, and Driver experience). The second part contains three questions related to the reaction and attitude of drivers due to bad weather conditions. The results showed that 45% of drivers involved in accidents during bad weather conditions although the majority (more than 75%) of them reduced the speed during the bad weather conditions. Overall, more research is needed to be done to evaluate the dust impact on driver performance as well as using real traffic accident data and compare it with this study to reach a solid conclusion and recommend the best solutions to improve traffic safety.

A Data-Driven Method Towards Minimizing Collision Severity for Highly Automated Vehicles

The deployment of autonomous vehicles on public roads calls for the development of methods that are reliably able to mitigate injury severity in case of unavoidable collisions. This study proposes a data-driven motion planning method capable of minimizing injury severity for vehicle occupants in unavoidable collisions. The method is based on establishing a metric that models the relationship between impact location and injury severity using real accident data, and subsequently including it in the cost function of a motion planning framework. The vehicle dynamics and associated constraints are considered through a precomputed trajectory library, which is generated by solving an optimal control problem. This allows for efficient computation as well as an accurate representation of the vehicle. The proposed motion planning approach is evaluated by simulation, and it is shown that the trajectory associated with the minimum cost mitigates the collision severity for occupants of passenger vehicles involved in the collision.

Probability-based extensive quantitative risk analysis: collision and grounding case studies for bulk carrier and general cargo ships

ABSTRACT In this study, the risk of collision and grounding accidents on bulk carrier and general cargo ships was investigated. A comprehensive risk analysis consisting of past accident reports, fault tree analysis (FTA) and Cut sets importance measure (CS-I) methods was carried out. The failures effective in the occurrence of collision and grounding accidents were determined by examining past accident reports. In the study, the probability for each failure was derived by using real accident data. Using FTA, various potential accident occurrence combinations/scenarios, also known as minimal cut sets (MCSs), for both accident types were obtained. In addition, prioritization was conducted for the MCSs of both types of accidents using CS-I sensitivity analyses. The faults that increase the risk of collision accidents the most were found to be interpretation failure of the Convention on the International Regulations for Preventing Collisions (COLREG) and lack of situational awareness. In addition, the top failures that lead to groundings were determined to be lack of communication in bridge resource management (BRM), fatigue/sleeplessness and inappropriate voyage plan. Recommendations are made for eliminating potential risk factors, preventing the recurrence of such accidents and establishing safer maritime transport.

Prediction of Occupant Injuries for Improved eCall Efficiency

Especially in emerging markets, no emergency infrastructure is established. In case of an accident, passing road users need to call for help or the ambulance. An automated eCall to friends, family or in future directly to the rescue control with transmission of GPS position, probable injury pattern and injury severity could significantly improve the rescue chain and would save lives in large scale. Initial approaches to the prediction of an injury pattern were investigated in this paper. A simulation model of a reference vehicle and with a hybrid III 50 percentile dummy was built by using the program MADYMO. A large amount of real accident data was processed using an algorithm created in MATLAB. This made it possible to adapt the acceleration data with identical rules and to determine the different restraint system ignition times. The algorithm allows the data to be used in the simulation model. Furthermore, the corresponding accident descriptions were analysed and translated into protection criteria level by means of literature. For the subsequent evaluation, the assessment protocol (Version 8.0.3) of the European New Car Assessment Programme (EuroNCAP) was used. To validate the model, an ordinal rating system was created to compare the ratings of the simulation and the injury descriptions. The results indicate a good prediction of the injury patterns with an agreement of 39.56 percent. Especially considering that a large number of influencing parameters are unknown.

Accident factors per construction type and stage: a synthesis of scientific research and professional experience

Following manufacturing, construction is the industry that is responsible for the majority of work-related accidents due to the complexity and uniqueness of each construction site. The accidents that occur on construction sites reflect the safety culture and the general health and safety (H&S) policy in the construction field. The purpose of this paper is to investigate the frequency of accidents and their contributing factors in each stage of 9 different types of construction. This paper utilizes data from published scientific literature to extract the universal contributing factors to be evaluated in the Greek construction industry and to compare results with documented international accidents. The riskiest construction stages per construction type and the main factors behind them are determined. The results show contradictions regarding the contributing factors as perceived by Greek professionals and those verified by international real accident data.

Crash Prediction on Expressway Incorporating Traffic Flow Continuity Parameters Based on Machine Learning Approach

Real-time crash prediction helps identify and prevent the occurrence of traffic crash. For years, various real-time crash prediction models have been investigated to provide effective information for proactive traffic management. When building real-time crash prediction model, a suitable variable space together with a specific time interval for traffic data aggregation and an appropriate modelling algorithm should be applied. Regarding the intercorrelation problem with variable space, comprehensive real-time crash prediction model considering available traffic data characteristics in applicable circumstances needs to be explored. Taking Xi’an G3001 Expressway as study area, real road traffic and accident data during the period from January 2014 to January 2019 on this expressway are applied for real-time crash prediction. To better capture traffic flow characteristics on expressway and improve the practicality of real-time crash prediction model, two new variables (segment difference coefficient and lane difference coefficient) describing the smoothness and continuity of traffic flow in spatial dimension are developed and incorporated in building the crash prediction model to solve the intercorrelation problem with variable space. Random forest (RF) is then adopted to specify the quantitative relationship between specific variable and crash risk. Real-time crash prediction model based on support vector machine (SVM) using new composed variable space is built. The results show that simplified variable space could contribute to the same classification power in currently used real-time crash prediction models compared with traditional variable space. Moreover, the prediction model based on SVM reaches an accuracy level of 0.9, which performs better than other currently used prediction models.

Methodology for Risk Ranking of At-Grade Intersections on Existing Rural Two-Lane Highways

The objective of this paper is the creation of a suitable methodology capable of deploying elements that characterize the existing infrastructure on rural two-lane highways regarding the location and geometric features of intersections and evaluating the built-in road safety. The target is the utilization of these elements to lead to evaluation coefficients that show the hazard level of every intersection and every road section, highlighting the intersections of the road network with reduced safety level. In this context, a hierarchy of critical parameters influencing the risk of an intersection was made, along with appropriate mathematical equations based on the current regulations and corresponding crash reduction factors (CRFs) that have been presented in the literature. Initially, case studies were conducted, and special diagrams were produced to assess the hazard level of each parameter. Then, basic statistical analysis of the results was made using regression factors to develop an appropriate mathematical model for the estimation of the hazard level. A corresponding procedure was performed for each parameter separately, and appropriate coefficients were sought. These coefficients were summed up, and a total score characterizing each intersection has been calculated, leading to a ranking in respect with the hazard level. Finally, data were collected on a significant sample of the existing road network of Greece, about 1,000 km of road network with more than 4,000 intersections (unpaved road intersections are also included), in order to evaluate the results of the proposed methodology. Part of the results has been interpreted with real accident data collected from the police department, showing that there is a correlation between them. This correlation must be confirmed in further research by a statistical analysis between the results of the present methodology and the actual recorded accidents.

Near Miss Detection for Encountering Ships in Sunda Strait

Sunda Strait has played an essential role to the Indonesian marine transportation. Safety of navigation is one issue that often leads to an accident, especially ship collision. A collision risk assessment in a waterway is critical not only to predict the probability of collision accident but also to understand its consequences to human life, shipping company, as well as to the environment. However, it could be hard to evaluate the accident due to the scarcity of real accident data. Thus, a near miss analysis can be considered as an equivalent substitute. A one-week of Automatic Identification System (AIS) data is taken as the sample to test the methodology proposed in this paper. The target area of the near miss analysis discussed in this paper is TSS area in Sunda Strait. To categorise the encounter as a safe or unsafe encounter, this paper utilises the closest point of approach (CPA) as the indicator and it contains two parameters, the distance to CPA (DCPA) and the time to CPA (TCPA). Four values of DCPA and TCPA are used as the threshold, those are introduced by Park, Fukuto and Imazu, Goodwin, and Limbach. In this time range, 135929 sample data are collected, but only 92,379 points from 276 ships are found inside the target area. The number of situations that are categorized as near miss encounters (NNM) for each threshold are 1668, 5870, 9370, and 13549, respectively. The combined threshold is proposed to adjust with the condition of Sunda Strait and produce a result with 9,946 points detected as near miss encounter.

A Joint Stochastic/Deterministic Process with Multi-Objective Decision Making Risk-Assessment Framework for Sustainable Constructions Engineering Projects—A Case Study

This study, on the one hand, develops a newfangled risk assessment and analysis (RAA) methodological approach (the MCDM-STO/DET one) for sustainable engineering projects by the amalgamation of a multicriteria decision-making (MCDM) process with the joint-collaboration of a deterministic (DET) and a stochastic (STO) process. On the other hand, proceeds to the application of MCDM-STO/DET at the workplaces of the Greek construction sector and also of the fixed-telecommunications technical projects of OTE SA (that is, the Greek Telecommunications Organization S.A.) by means of real accident data coming from two official State databases, namely of “SEPE” (Labor Inspectorate, Hellenic Ministry of Employment) and of “IKA” (Social Insurance Institution, Hellenic Ministry of Health), all the way through the period of the years2009–2016.Consequently, the article’s objectives are the following: (i) The implementation and execution of the joint MCDM-STO/DET framework, and (ii) to make known that the proposed MCDM-STO/DET algorithm can be a precious method for safety managers (and/or decision-makers) to ameliorate occupational safety and health (OSH) and to endorse the sustainable operation of technical or engineering projects as well. Mainly, we mingle two different configurations of the MCDM method, initially the Analytical Hierarchy-Process (the typical-AHP), and afterwards the Fuzzy-Extended AHP (the FEAHP) one, along with the Proportional Risk Assessment Technique (PRAT) and the analysis of Time-Series Processes (TSP), and finally with the Fault-Tree Analysis (FTA).

Validation of Simulated Safety Indicators with Traffic Crash Data

The purpose of this document is to validate a new methodology useful for the estimation of road accidents resulting from possible driver distractions. This was possible through a statistical comparison made between real accident data between 2016 and 2018 in the city of Santander (Spain) and simulated data resulting from the application of the methodology on two areas of study. The methodology allows us to evaluate possible collisions starting from the knowledge of vehicular trajectories extrapolated from microsimulation. Studies show that there are good correlations between the real data and the simulated data. The results obtained show that the proposed methodology can be considered reliable and, therefore, it could be of fundamental importance for designers, since it would simplify the choice between different possible intervention scenarios, determining which is the least risky in terms of road safety.

Modeling of boilover phenomenon consequences: Computational fluid dynamics (CFD) and empirical correlations

In the present study, empirical correlations were applied to predict the time to boilover and its consequences. Next, a computational fluid dynamics (CFD) model was applied for the simulation of boilover consequences based on the resulting incident radiative heat flux. Empirical correlations, and the Fire Dynamic Simulator (FDS) were compared with a crude oil boilover experiment carried out in a 1m diameter pan (small-scale). FDS predictions were consistent with experimental results. A case study concerning the boilover accident in Milford Haven with a 78m diameter tank (large-scale) was performed with the CFD model and empirical correlations. A comparison of the simulation results and accident data demonstrated that FDS was able to predict the radiation heat flux of the boilover in the target area in agreement with real accident data. The time to boilover correlations gave results in agreement with experimental and real accident data while consequence prediction correlations exaggerated the radiation heat flux and fireball height and diameter. It is concluded that an integrated approach including empirical correlations for the prediction of time to boilover and FDS for the prediction of the radiation heat flux in the target area, provide realistic results for prediction of time to boilover and its consequences.

Prediction of injury risks and features among scooter riders through MADYMO reconstruction of a scooter-microvan accident: Identifying the driver and passengers

In dealing with a scooter-related traffic accident with rider death, it is necessary to identify the driver responsible for the accident. This study aimed to reconstruct the kinematics of a scooter–microvan accident involving three riders and explored the differences in injury risks and characteristics of the scooter driver and passengers. We reconstructed a real accident by using MADYMO multi-body simulation software. Moreover, we designed two-variable simulation experiments to analyze how the velocity and impact angle of the microvan are related to the injuries of the three riders. When the microvan speed is set at 18 km/h and that of the scooter is set at 28.8 km/h, the simulated kinematics correlates well with real accident data, and the impact positions and injury parameters correlate well with the actual injuries. When the impact angle is smaller than 30° and the microvan impact velocity is lower than 40 km/h, the head injury of the driver is more life-threatening than the corresponding injuries of the rear passengers. When the impact angle is 15° and the microvan impact velocity is in the range of 0–20 km/h, the femur fracture risk is higher for the driver than for passengers. As the impact angle increases to 45°, passengers have a higher risk of femur fracture than the driver in the velocity range of 0–10 km/h. This impact velocity range becomes 0–30 km/h at an impact angle of 60° and then 40–70 km/h at an impact angle of 90°. Our study shows that the multibody method can reconstruct accidents and predict the different injury features and risks between the driver and passengers, which is useful in identifying the driver.