Real-time risk analysis of road tanker containing flammable liquid based on fuzzy Bayesian network
Introduction
According to statistics, there were approximately 2636 hazardous chemical accidents in China from 2010 to 2015, and 54 % of them were related to road transportation. Many hazardous chemicals are flammable liquids transported with road tankers. In the case of a loss of containment (LOC), the released liquid may trigger fires and explosions and poses considerable threats to the surrounding people and vehicles, i.e., the Yanhou tunnel accident in China, 2014; the Bologna accident in Italy, 2018. Road tankers are “moving hazards”. Both the leakage probability and the consequences dynamically change due to the ever-varying environment and other conditions. Thus, a real-time risk analysis of road tankers is necessary. Recently, GPS tracking of road tankers has been compulsory in some cities in China, and together with other information (i.e., weather, traffic), real-time risk analysis of road tankers is becoming feasible.
In the past two decades, a growing number of studies have been conducted for the safety transportation of hazardous materials by road, mainly focusing on the following four aspects:(i) accident statistical analysis (Oggero et al., 2006; Yang et al., 2010; Zhao et al., 2012; Shen et al., 2013; Ambituuni et al., 2015), including the accident features and the classification, the occurrence frequency, and the consequences; (ii) framework design for safety regulation and emergency response (Fabiano et al., 2002, 2005; Ambituuni et al., 2015); (iii) selection and optimization of transport routes (Tena-Chollet et al., 2013; Mahmoudabadi, 2015); and (iv) evaluating the probabilities (Ronza et al., 2007), the risk magnitudes (Verter and Kara, 2001; Gheorghe, 2006; Tomasoni et al., 2010; Chakrabarti and Parikh, 2012; Landucci et al., 2017), or minimizing the risk through quantitative assessment (Guo and Verma, 2010; Mahmoudabadi et al., 2018). These works have made considerable efforts in qualitative and quantitative analysis, but few have depicted the dynamic change in risk.
Real-time risk is a kind of dynamic risk, which is often regarded as a composite measure of the time-varying likelihood and severity of adverse effects. The methods of dynamic risk analysis are summarized in Ref. (Khan et al., 2015; Villa et al., 2016), and are applied in a variety of fields (Paltrinieri et al., 2014a, 2014b, 2015; Khakzad et al., 2012, 2013; Koromila et al., 2015; Xin et al., 2017). In general, there are two widely accepted dynamic risk analysis methods. One is based on the Bayesian approach, which updates the prior failure probabilities in the form of a likelihood function by inputting new observations or evidence and applying Bayes' theorem (Cai et al., 2013, 2019b; Abimbola et al., 2015; Zarei et al., 2017; Gyftakis et al., 2018). The other is a non-Bayesian method that provides new data through real-time monitoring, physical reliability models or inspection of process equipment (Khakzad et al., 2012; Paltrinieri et al., 2014b; Abimbola et al., 2014). The Bayesian network method is an easy-to-update model that has recently become increasingly popular for risk analysis of process industries (Meel and Seider, 2008; Nordgard and Sand, 2010; Khakzad et al., 2013; Wu et al., 2016, 2017, 2019; Khan et al., 2018; Leonia et al., 2019; Cai et al., 2019a). It is highly recommended for real-time risk analysis, for it can provide a more reliable and practical forecast (Islam et al., 2017). Since the conditions in a road transportation system change over time and some of the condition parameters can be monitored with GPS data or other real-time information, the combination of Bayesian network and the real-time condition parameters, can well address the complex and uncertain situations in dynamic risk analysis.
The present study is aimed at showing the application of the Bayesian Network in real-time risk analysis of road tankers containing flammable liquids. Furthermore, the case of road tanker is applied in this research. Section 2 describes the methods used in this paper, and Section 3 shows the detailed process of dynamic hazard identification. Section 4 shows the construction of the Bayesian network. Section 5 discusses the main results of the study, and Section 6 draws the main conclusion.
Section snippets
Methodology
This section provides an overview of the proposed method, as shown in Fig. 1. First, the hazards in a flammable liquid transportation system are identified based on the Bow-tie model, and the Bayesian network structure is further determined accordingly. Second, a probabilistic estimation model, which combines expert judgment and fuzzy set theory, is established to determine the prior probabilities and the conditional probability tables of the nodes in the Bayesian network. Finally, the
Flammable liquid road transportation system
Flammable liquid, as an important chemical raw material, is usually transported by road tankers. Fig. 3 briefly describes the transportation system. It is influenced by both the internal and external factors. The internal factors can be the driver, the materials (flammable liquids), the vehicle, the tank and other apparatus. The external factors consist of weather, traffic, route location and other environmental elements. It should be noted that both the internal and external factors change
Risk analysis based on the Bayesian network
BT can show a comprehensive risk evolution of road tankers; however, it has limitations in risk assessment. First, BT is static and cannot update with new information (Khakzad et al., 2011). Hence, it is not useful in dynamic risk assessment, especially for road transportation systems with various ever-changing parameters. In addition, the variables in BT are binary, which means there are only two states: occurrence or nonoccurrence (Martins et al., 2014). Another limitation is that the
Case study: the Yanhou tunnel accident
To demonstrate the applicability of the proposed BN model, the Yanhou tunnel accident is used as a case study (Ingason and Li, 2017). The accident causes, reported by the Ministry of Emergency Management of the People's Republic of China, are used as the evidence in the BN model.
Conclusions
The road transportation of flammable liquid is full of uncertainties. Both the probability for road tanker LOC and the consequences afterward are strongly dependent on the varying passed locations and the environment. This paper provides a method for evaluating the real-time risk level for road tankers containing flammable liquids based on a Bayesian network. The Bayesian network was built after a revision of the hazard identification result from a bow-tie diagram. Domain experts’ knowledge was
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This work was supported by the National Key R&D Program of China (Grant No. 2018YFC0809300) and by the National Natural Science Foundation of China (Grant No. 51806247).
References (57)
- et al.
Dynamic safety risk analysis of offshore drilling
J. Loss Prev. Process Ind.
(2014) - et al.
Risk assessment of petroleum product transportation by road: a framework for regulatory improvement
Saf. Sci.
(2015) - et al.
Safety and risk analysis of managed pressure drilling operation using Bayesian network
Saf. Sci.
(2015) - et al.
Applying HAZAN methodology to hazmat transportation risk assessment
Process Saf. Environ. Prot.
(2012) - et al.
Analyzing system safety and risks under uncertainty using a bow-tie diagram: an innovative approach
Process Saf. Environ. Prot.
(2013) - et al.
A framework for risk assessment and decision-making strategies in dangerous good transportation
J. Hazard. Mater.
(2002) - et al.
Dangerous good transportation by road: from risk analysis to emergency planning
J. Loss Prev. Process Ind.
(2005) - et al.
Choosing vehicle capacity to minimize risk for transporting flammable materials
J. Loss Prev. Process Ind.
(2010) - et al.
Experimental investigation on transverse ceiling flame length and temperature distribution of sidewall confined tunnel fire
Fire Saf. J.
(2017) - et al.
Real time risk analysis of kick detection: testing and validation
Reliab. Eng. Syst. Saf.
(2017)
Spilled liquid fires in tunnels
Fire Saf. J.
Safety analysis in process facilities: comparison of fault tree and Bayesian network approaches
Reliab. Eng. Syst. Saf.
Dynamic risk analysis using bow-tie approach
Reliab. Eng. Syst. Saf.
Dynamic safety analysis of process systems by mapping bow-tie into Bayesian network
Process Saf. Environ. Prot.
Methods and models in process safety and risk management: past, present and future
Process Saf. Environ. Prot.
An operational risk analysis tool to analyze marine transportation in Arctic waters
Reliab. Eng. Syst. Saf.
Quantitative risk analysis on leakage failure of submarine oil and gas pipelines using Bayesian network
Process Saf. Environ. Prot.
Risk assessment of mine ignition sources using fuzzy Bayesian network
Process Saf. Environ. Prot.
HazMat transportation risk assessment: a revisitation in the perspective of the Viareggio LPG accident
J. Loss Prev. Process Ind.
Developing a chaotic pattern of dynamic risk definition for solving hazardous material routing-locating problem
J. Loss Prev. Process Ind.
Real-time risk analysis of safety systems
Comput. Chem. Eng.
Application of Bayesian networks for risk analysis of MV air insulated switch operation
Reliab Eng Syst Saf.
A survey of accidents occurring during the transport of hazardous substances by road and rail
J. Hazard. Mater. A
Dynamic approach to risk management: application to the Hoeganaes metal dust accidents
Process Saf. Environ. Prot.
A fuzzy reliability assessment of basic events of fault trees through qualitative data processing
Fuzzy Sets Syst.
Using transportation accident databases to investigate ignition and explosion probabilities of flammable spills
J. Hazard. Mater.
Development of a spatial risk assessment tool for the transportation of hydrocarbons: methodology and implementation in a geographical information system
Environ. Model. Softw.
Towards dynamic risk analysis: a review of the risk assessment approach and its limitations in the chemical process industry
Saf. Sci.
Cited by (54)
Integrating FBN and FDS for quantitative risk assessment of cable fire in utility tunnel
2024, Journal of Loss Prevention in the Process IndustriesProbabilistic risk analysis and characterization of hazardous material road transportation routes under explosion accidents
2024, Journal of Loss Prevention in the Process IndustriesRisk analysis for hazardous chemical vehicle-bridge transportation system: A dynamic Bayesian network model incorporating vehicle dynamics
2024, Reliability Engineering and System SafetyDynamic analysis of pilot transfer accidents
2023, Ocean EngineeringA dynamic community gas risk-prediction method based on temporal knowledge graphs
2023, Process Safety and Environmental ProtectionA novel dynamic risk assessment method for the petrochemical industry using bow-tie analysis and Bayesian network analysis method based on the methodological framework of ARAMIS project
2023, Reliability Engineering and System Safety