Article
Engineering, Industrial
Rodrigo L. Rose, Tejas G. Puranik, Dimitri N. Mavris, Arjun H. Rao
Summary: This paper presents an application of structural topic modeling (STM) to analyze aviation safety data. The results show that STM has the potential to identify themes within technical datasets, with improved performance for more specific corpora that use precise and unique language.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2022)
Article
Engineering, Marine
Lucjan Gucma, Andrej Androjna, Kinga Lazuga, Peter Vidmar, Marko Perkovic
Summary: Despite advances in navigation, accidents at sea remain challenging to prevent. Improvement in accident models may enhance future prevention efforts, but the traditional models lack coverage of organizational management and design phase errors, as well as changes in systems over time.
JOURNAL OF MARINE SCIENCE AND ENGINEERING
(2021)
Article
Computer Science, Artificial Intelligence
Luca Farnia, Mia Alibegovic, Edward Cruickshank
Summary: This study evaluates the performance of various causal structure learning algorithms in detecting true causal relations among variables. Constraint-based, score-based, and hybrid algorithms are compared and ranked based on effectiveness and efficiency in directed or undirected acyclic graphs. Monte Carlo simulations are conducted to create linear causal effects among variables in different sample sizes and causal network properties. The presence of latent confounding variables is found to be the main limitation of the algorithms, regardless of the sample size.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Ergonomics
Su-Xia Liu, Hua-Zhong Chen, Qiang Mei, Ying Zhou, Nkrumah Nana Kwame Edmund
Summary: Employee safety behavior is crucial for work safety in enterprises. This study investigated the relationship between the behavior of front-line managers and employee safety behavior using a combination of qualitative (interpretive structural model) and quantitative (Bayesian network) methods. The results revealed that safety incentives and safety communication were the main predictors of safety participation, while safety supervision and safety control were the main predictors of safety compliance. Furthermore, the study found that instant improvements in safety communication, safety incentives, safety supervision, and safety guidance were the most effective measures to achieve a high level of employee safety behavior in the workplace.
INTERNATIONAL JOURNAL OF OCCUPATIONAL SAFETY AND ERGONOMICS
(2022)
Article
Engineering, Industrial
S. T. Ung
Summary: The study uses a Bayesian Network to estimate shipping accident risk in Taiwan, revealing variations in accident risk among different types and tonnages of vessels warranting global attention and further research. The rationality of the Bayesian Network model is validated through sensitivity analysis, providing an effective tool for navigation risk prediction.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2021)
Article
Environmental Sciences
Guyuan Li, Xiaofeng Wu, Jing-Cheng Han, Bing Li, Yuefei Huang, Yongqiang Wang
Summary: This paper proposes a risk assessment model based on the Bayesian network and Interpretative Structural Modeling (ISM) method for reliable flood assessments. The model incorporates ISM to identify the relations among multiple risk factors and configure the BN structure for risk inference. The model was demonstrated in Shenzhen city, China, and achieved an accuracy of 76% with an Area Under ROC Curve (AUC) of 0.82. The spatial distribution of risk levels showed satisfactory performance, and the maximum daily rainfall, elevation, and storm frequency were identified as key indicators for flood occurrence in the study area. The spatial flood risk map generated under various design rainfall scenarios can help identify areas that require particular attention. Thus, the developed assessment model is a useful tool for supporting flood risk governance and achieving reliable urban water security.
JOURNAL OF ENVIRONMENTAL MANAGEMENT
(2023)
Article
Energy & Fuels
Wanguan Qiao, Xue Chen, Wenxin Xia
Summary: This paper employs the STAMP model to analyze coal mine accidents and finds that it can lead to more comprehensive conclusions compared to traditional models. Apart from equipment defects and illegal operations by workers, the fundamental reason behind accidents lies in the failure of feedback loops in the layered control structure of the system.
Article
Automation & Control Systems
Diviyan Kalainathan, Olivier Goudet, Isabelle Guyon Guyon, David Lopez-Paz, Michele Sebag
Summary: SAM is a causal discovery method that utilizes both conditional independencies and distributional asymmetries to infer the underlying causal structure. It involves a game between players estimating variable distributions conditionally and an adversary discriminating generated data from original data.
JOURNAL OF MACHINE LEARNING RESEARCH
(2022)
Article
Engineering, Marine
Shiguan Liao, Jinxian Weng, Zhaomin Zhang, Zhuang Li, Fang Li
Summary: This paper presents a scenario evolution model for maritime accidents using Bayesian networks to predict the most probable causes of different types of incidents. The model considers factors such as accident type, life loss contingency, accident severity, quarter and time period of the accident, and type and gross tonnage of the involved ships. An analysis of 5660 global maritime accidents from 2005 to 2020 was conducted, and a tree augmented network (TAN) model was constructed using Netica software to account for interdependencies among risk-influencing factors. The findings demonstrate the model's competence in predicting accident scenarios and extrapolating accident consequences.
JOURNAL OF MARINE SCIENCE AND ENGINEERING
(2023)
Article
Engineering, Industrial
Andres Ruiz-Tagle, Enrique Lopez-Droguett, Katrina M. Groth
Summary: This paper presents a novel probabilistic approach to counterfactual reasoning, allowing rigorous evaluation of counterfactual hypotheses in system safety through causally-sound probabilistic methods. The capabilities of this approach are demonstrated through a real-world case study on the 2018 Sun Prairie gas explosion.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2022)
Article
Engineering, Chemical
Changfeng Yuan, Yichao Hu, Yulong Zhang, Tao Zuo, Jiahui Wang, Shuangjiao Fan
Summary: This study summarized 19 frequently occurring basic scenarios in emergency processes for fire accidents in oil-gas storage and transportation through statistical analysis of actual accident cases, and calculated the state probabilities of scenario nodes using fuzzy numbers and dynamic Bayesian network. By establishing an accident consequence prediction model, it provided decision makers with a scientific basis for effective emergency measures.
JOURNAL OF LOSS PREVENTION IN THE PROCESS INDUSTRIES
(2021)
Article
Ergonomics
Joshua L. Hunte, Martin Neil, Norman E. Fenton
Summary: Product risk assessment is the process of determining whether a product is safe for consumer use. The primary method used by regulators in the UK and EU is RAPEX, but it has limitations. To address these limitations, a Bayesian Network (BN) model is developed to provide an improved systematic approach. The BN approach can assess products with or without data, providing deeper insights. It complements traditional methods and provides quantifiable, auditable assessments when data is lacking.
JOURNAL OF SAFETY RESEARCH
(2022)
Article
Engineering, Civil
Donatella de Silva, M. Andreini, A. Bilotta, G. De Rosa, S. La Mendola, E. Nigro, O. Rios
Summary: This paper presents a performance-based approach for the design and verification of fire damage levels in RC tunnel linings. The main parameters controlling each damage state are determined through back analyses of real fires. The proposed approach is applied to a real concrete tunnel, and the potential fire damage level is evaluated by analyzing the temperatures and stresses of the structural lining.
FIRE SAFETY JOURNAL
(2022)
Article
Public, Environmental & Occupational Health
Amir Hossein Khoshakhlagh, Saeid Yazdanirad, Masoud Motalebi Kashani, Elham Khatooni, Yaser Hatamnegad, Sohag Kabir
Summary: The study showed that high job stress could decrease the occurrence of high safety climate and increase the likelihood of accidents, while a low safety climate condition could also increase the accident occurrence. Additionally, the simultaneous presence of high job stress and low safety climate could further raise the risk of accidents.
Article
Computer Science, Interdisciplinary Applications
Haidong Yang, Luying Chen, Biyu Liu, Athanasios Migdalas
Summary: This study focuses on the suppliers' optimal emergency procurement/production decision (EPD) with information updating in the event of sudden accidents (SAs). The research proposes a real-time updated emergency decision-making model (EDM) based on Stackelberg game and Bayesian inference, and quantitatively analyzes the value of information updating and critical factors affecting the suppliers' optimal EPD. Results show that information is crucial for effective decision-making, and the proposed model helps suppliers determine when and how many products to procure/produce to maximize profits. The cost parameters significantly impact the suppliers' EPD with information updating.
COMPUTERS & INDUSTRIAL ENGINEERING
(2021)
Article
Engineering, Chemical
Chao Chen, Hang Chen, Li Mo, Shenbin Xiao, Changjun Li, Ming Yang, Genserik Reniers
Summary: Fire-induced domino effect is a major threat to hazardous material storage tanks. Previous research neglected the impact of wind load on the thermal buckling behavior of tanks exposed to fire. This paper conducts a numerical simulation to analyze the synergistic effects of fire and wind loads on the thermal post-buckling behavior of storage tanks, considering wind parameters and heat radiation parameters.
JOURNAL OF LOSS PREVENTION IN THE PROCESS INDUSTRIES
(2024)
Article
Engineering, Chemical
Zhang Chu, Liu Lili, Li Wei, Li Beibei, Liu Mingxing
Summary: This study predicts the fatigue life of blasting discs using a combined simulation method and measures the deformation of conventional slotted blasting discs subjected to repeated loading using a high-precision laser scanning method. A relationship model is established between deformation, loading pressure, cycle number, and change in blasting pressure. The results show that the deformation varies linearly with the cycle number and follows a power function relationship with the change in blasting pressure. The deformation can be used to accurately predict the fatigue life of blasting discs.
JOURNAL OF LOSS PREVENTION IN THE PROCESS INDUSTRIES
(2024)
Article
Engineering, Chemical
Shin-Mei Ouyang, Jia-Chi Ye, Wei-Chun Chen, Yu-Jen Chen, Chin-Feng Chen
Summary: This study employed the Taguchi method and analysis of variance to investigate the factors that influence the deterioration of pyrotechnics when stored under different humidity and temperature conditions. Various experimental methods were used to study the microstructure, heat flow, and safety of the pyrotechnics. The results showed that humidity had a significant impact on the pyrotechnics, and controlling humidity could reduce the hazards associated with storing pyrotechnics.
JOURNAL OF LOSS PREVENTION IN THE PROCESS INDUSTRIES
(2024)
Article
Engineering, Chemical
Bei Li, Huiqiang Liu, Heng Gao, Chi-Min Shu, Mingshu Bi
Summary: The moisture content sprayed into the dust layer in industrial sites is a main factor affecting reignition. The study found that moisture has a certain influence on the thermophysical properties of the dust layer, playing an important role in heat transfer and heat absorption processes. The results showed that as the moisture content increased, the thermal conductivity and heat sink density of the sample increased, indicating the need to have a heat sink density higher than the heat increase value to suppress ignition.
JOURNAL OF LOSS PREVENTION IN THE PROCESS INDUSTRIES
(2024)
Article
Engineering, Chemical
Yue Jing, Yong Pan, Fan Yang, Dan Wei, Wenhe Wang
Summary: A comprehensive risk assessment method for the esterification process was proposed in this study. It established a comprehensive evaluation index system and utilized improved Dempster-Shafer evidence theory and cloud model for assessment. The method was validated through case studies and showed consistency with established indexes.
JOURNAL OF LOSS PREVENTION IN THE PROCESS INDUSTRIES
(2024)