Oil and gas pipelines play a key role in the safe and efficient delivery of energy resources around the world. Crude oil by itself is not corrosive, but oil extracted from geological reservoirs is accompanied by varying amounts of water and acidic gases such as carbon dioxide (CO2), which can form a corrosive combination. Estimating the corrosion rate and depth in pipelines is essential for predicting their failure probability. In the present study, a Bayesian network has been developed for predicting the distribution of corrosion rate in oil pipelines given the point estimates generated using an empirical corrosion simulation model. For this purpose, the simulation model considers corrosion parameters such as pipe diameter, flow temperature, flow velocity, and CO2 partial pressure, among others. With the corrosion rate distribution predicted by the Bayesian network, corrosion depth–rate relationships have been employed to convert the corrosion rate distribution into failure probability distribution.

Major accidents in Western countries, receiving a lot of media attention in the 1970s, are starting point for research into internal and external domino effects in the chemical and petrochemical sectors and clusters. Initially, these reports are published by government institutions and government-related research centers. With the upcoming quantitative risk analyses in the 1970 and 1980s, the so-called colored books, published in the Netherlands, play a prominent role in quantifying these domino effects. Since the mid-1990s, the second European Seveso Directive encourages scientific research on domino effects, shown in substantial growth of academic publications on the topic. Research in Western countries is dominated by risk assessments, probabilities, and failure mechanisms are calculated for the complex phenomenon of domino effects and its consequences. Previous works are closely related to political, official, and private decision-making.

The present study aims to investigate the impacts of oil sands development in Canada on the economy, society and the environment as the three pillars of sustainability. Factors such as aquatic ecosystems, land disturbance and reclamation, air quality, public health, safety, aboriginal and local communities, gross domestic product, employment rate and job creation, government revenues and demographic changes have been considered. Based on a review of the available literature, this study shows that the oil sands industry has so far fallen short in keeping a balance among the three pillars of sustainability, with the negative impacts on society (e.g. changing the lifestyle of Aboriginal people) and the environment (e.g. land disturbance) overweighing the relatively positive economic impacts. This, along with the current pace of remedies (e.g. land reclamation), makes it hard to conclude that the oil sands industry is sustainable.

Recent catastrophic accidents in China and the USA urge and justify a thorough study on current & future research trends in the development of modeling methods and protection strategies for prevention and mitigation of large-scale escalating events or so-called domino effects in the process and chemical industries. This paper firstly provides an overview of what constitutes domino effects based on the definition and features, characterizing domino effect studies according to different research issues and approaches. The modeling approaches are grouped into three types while the protection strategies are divided into five categories, followed by detailed descriptions of representative modeling approaches and management strategies in chemical plants and clusters. The current research trends in this field are obtained based on the analysis of research work on domino effects caused by accidental events, natural events, and intentional attacks over a period of the past 30 years. A comparison analysis is conducted for the current modeling approaches and management strategies to pose their applications. Finally, this paper offers future research directions and identifies critical challenges in the field, aiming at improving the safety and security of chemical industrial areas so as to prevent and mitigate domino effects.

The Arctic is known for its abundant reserve of natural resources. Last decade has seen some exploration and production activities in this region. The assurance of safe operations in this region is a critical and challenging task because of the harsh environment, the remoteness of operation sites, the limited infrastructure, and resources available in response to emergent situations, the application of costly equipment and facilities, and the sensitive marine environment. For complex process systems operating in a harsh environment, the scope of conventional risk assessment is not enough because of the highly uncertain environment, and its impacts on equipment performance. Risk assessment needs to be extended to include both the pre-failure and the post-failure phases. Additionally, risk assessment approaches under normal operating, and environmental conditions may not be applicable in the Arctic regions with unique and uncertain characteristics of the harsh environment. Therefore, this study aims to develop a quantitative resilience assessment method for process units operating under Arctic extreme conditions. Dynamic Bayesian network (DBN) is applied to model the probabilistic relationships between causes and effects in a dynamic manner. The proposed method is applied to the resilience assessment of a separator (as part of an oil production system). The proposed approach will help reveal the critical operating parameters under extreme conditions for process units. It also helps identify potential design improvement to enhance process safety.

Success Likelihood Index Model (SLIM) is one of the widely-used deterministic techniques in human reliability assessment especially when data is insufficient. However, this method suffers from epistemic uncertainty as it extremely relies on expert judgment for determining the model parameters such as the rates and weights of the performance shaping factors (PSFs). Besides, given an operation consisting of several tasks, SLIM calculates the human error probability (HEP) by ignoring possible dependencies among the tasks.

The present study is aimed at using Bayesian Network (BN) for improving the performance of SLIM in handling uncertainty arising from experts opinion and lack of data. To this end, SLIM is combined with BN to form the so-called BN-SLIM technique. We demonstrate how BN-SLIM can consider uncertainty associated with the rates of PSFs by using probability distributions. BN-SLIM is also able to provide a better estimation of human error probability by considering conditional dependencies resulting from common PSFs. The probability updating feature of BN-SLIM can be used to identify the PSFs contributing the most to human failure event. The outperformance of BN-SLIM over SLIM is demonstrated via an illustrative example.

The authors regret that an imprecise statement was made in this article, and wish to offer this Corrigendum as clarification. The authors would like to apologise for any inconvenience caused. Imprecise statement: Misuri et al. (2018, pp. 70) state in their work: “Differently from EN [evidential network], CN [credal network] can be used both to conduct forward analysis and to update probabilities”. This statement implies that EN cannot be used for belief updating and should be mapped into a corresponding CN for that purpose. Correction: The foregoing statement may be correct for conventional ENs that are based on Dempster's combination rule, but does not hold true for the EN developed by Simon and co-workers (2008, 2009) based on Bayesian network (BN) inference algorithms (herein, BN-based EN). Simon and Weber (2009) explicitly mention in their work that the developed EN can be used for belief updating: “The computation mechanism is based on the Bayes theorem, which is extended to the representation of uncertain information according to the framework of Dempster-Shafer theory. Specific evidence (Hard evidence) is modeled by a mass of 1 on one of the focal elements of the frame of discernment. Non-specific evidence (Soft evidence) corresponds to a mass distribution on the focal elements of the frame of discernment.” Proof: Fig. 1 displays the BN-based EN developed in Misuri et al. (2018) for security vulnerability assessment of a process plant where the belief masses have been updated given the evidence “Attack = Success”. Misuri et al. (2018) used the BN-based EN for predicting the probability of a successful attack (forward analysis), but to update the probabilities (backward analysis) they mapped the BN-based EN into an equivalent CN and calculated the updated probabilities given “Attack = Success”. They used two packages, JavaBayes and GL2U, to implement the CN, concluding that JavaBayes results in more consisting updated probabilities with regard to the evidence (the 2nd column in Table 1). In the present corrigendum, we used the BN-based EN (Fig. 1) for belief updating given “Attack = Success”. The updated beliefs were subsequently used to calculate updated probability intervals (the 3rd column in Table 1), showing a good agreement between the results of CN and BN-based EN. Conclusion: The EN developed by Simon and co-workers (2008, 2009) based on BN can be used for belief mass updating the same way BN can be used for probability updating, with no need for using CN.

Domino effects are high-impact phenomena that have caused catastrophic damage to several chemical and process plants around the world through secondary incidents caused by primary ones. With the increasing trend of cyberattacks targeting critical infrastructures, there is a concern that such cyberattacks may trigger domino effects, by manipulating industrial control systems in such a way that the physical consequences are likely to escalate. In this study, we have demonstrated that via network segmentation of industrial control systems, the plant robustness against cyberattack-related domino effects can be improved. To this end, a risk-based decision-making methodology is developed based on Bayesian network and graph theory to investigate and evaluate the robustness of segmentation alternatives. The application of the methodology to an illustrative case study shows the efficacy of the approach as a viable cyber risk mitigation measure in chemical and process plants.

The industrial clustering process in the chemical industry is becoming progressively more important due to economic, social and political issues. Industrial clustering means agglomeration of companies in the same geographical area in order to increase productivity and reduce costs. Nonetheless, clustering also has some important safety and security implications. The aim of this study is twofold: firstly, the development of an algorithm for the classification of chemical industrial clusters with regards to safety and security risks. Secondly, considering the importance of a multi-plant safety and security management system, highlighting the greater efficiency in the reduction of risk where adequate cooperation exists. The methodology is divided in three main steps, namely, “selection” (of the chemical parks to be processed), “assessment” of average hazard and vulnerability of installations within the cluster area, followed by an analysis of the relationships within companies in terms of strategic and operational cooperation, and “ranking”. The last step evaluates the strong influences of the above-mentioned parameters through the analytic network process (ANP) and leads to a final classification.

Chemical industrial areas comprising various hazardous installations may be attacked by adversaries, triggering possible intentional domino effects. Compared with accidental domino effects, intentional domino effects may be more difficult to prevent since intelligent and strategic adversaries can adapt their tactics according to protection measures. However, how and to what extent domino effects affect security management is ignored in previous studies. This study proposes a methodology to prevent and mitigate intentional domino effects taking into consideration economic issues in the decision-making process on safety and security resources. The methodology is divided into five parts: threat analysis, vulnerability analysis of installations with respect to intentional attacks, vulnerability analysis of installations subject to possible domino effects caused by the attacks, cost-benefit analysis, and optimization. Net present value of benefits (NPVB) is employed and quantified in the cost-benefit analysis to determine whether a protection strategy (a combination of safety and security measures) is profitable, or not. Besides, an optimization algorithm called “PROTOPT” based on “maximin” strategy is developed to achieve the most profitable protection strategy. An illustrated case study shows that domino effects can not be ignored in security management since they may have a profound impact on adversaries’ strategies.

Quantitative risk assessment (QRA) has played an effective role in improving safety of process systems during the last decades. However, QRA conventional techniques such as fault tree and bow-tie diagram suffer from drawbacks as being static and ineffective in handling uncertainty, which hamper their application to risk analysis of process systems. Bayesian network (BN) has well proven as a flexible and robust technique in accident modeling and risk assessment of engineering systems. Despite its merits, conventional applications of BN have been criticized for the utilization of crisp probabilities in assessing uncertainty. The present study is aimed at alleviating this drawback by developing a Fuzzy Bayesian Network (FBN) methodology to deal more effectively with uncertainty. Using expert elicitation and fuzzy theory to determine probabilities, FBN employs the same reasoning and inference algorithms of conventional BN for predictive analysis and probability updating. A comparison between the results of FBN and BN, especially in critically analysis of root events, shows the outperformance of FBN in providing more detailed, transparent and realistic results.

Because of modern societies' dependence on industrial control systems, adequate response to system failures is essential. In order to take appropriate measures, it is crucial for operators to be able to distinguish between intentional attacks and accidental technical failures. However, adequate decision support for this matter is lacking. In this paper, we use Bayesian Networks (BNs) to distinguish between intentional attacks and accidental technical failures, based on contributory factors and observations (or test results). To facilitate knowledge elicitation, we use extended fishbone diagrams for discussions with experts, and then translate those into the BN formalism. We demonstrate the methodology using an example in a case study from the water management domain.

Risk analysis in process systems is very important to design effective strategies for preventing and mitigating potential major accidents. Although conventional techniques as Bow-tie (BT) have widely been used in risk assessment of process systems, they fall short in effectively modelling epistemic uncertainty which is prevailing in risk assessment of process systems. The present study is aimed at alleviating this shortcoming by incorporating fuzzy set theory into BT, developing a so-called fuzzy extended Bow tie (FEBT) model. FEBT, compared with previous fuzzy BT methods, uses the intuitionistic fuzzy numbers and thus provides a more accurate cause – consequence model of accident scenarios. A natural gas transmission network is used to demonstrate the application of FEBT.

Chemical industrial parks, being critical infrastructures, are susceptible to domino effects triggered by intentional attacks. Previous research on security risk management has mainly focused on using security measures to prevent intentional attacks, neglecting the effects of safety barriers. Safety barriers are able to reduce the potential consequences and decrease the attractiveness of chemical industrial parks to terrorists who aim to maximize the damage. From a systematic perspective, the potential consequence of intentional attacks is defined as the expected loss which is the sum-product of damage probability and consequence of installations. A consequence-based method including a Dynamic Vulnerability Assessment Graph (DVAG) model is proposed to integrate safety and security resources for reducing the risk of intentional attacks. The DVAG model is developed based on dynamic graphs, considering the effects of security measures, safety barriers, and emergency response. This method can assess the consequences and damage probabilities of possible intentional attacks so as to mitigate the risk via evaluation and allocation of security measures and safety barriers with fast computation speed.

Optimal selection of safety measures (SMs) is a challenging task for safety managers due to its importance, complexity, and incapability of traditional approaches in considering all the aspects of SMs optimal selection. Sophisticated mathematical models can be used to overcome the limitations of traditional approaches. However, setting the objective functions while considering their priorities as well as possible synergistic effects of the SMs on the hazards are still among the main concerns in the development and application of mathematical models.
The present study is aimed at developing a methodology to optimize the SMs selection while addressing the aforementioned challenges and considering both the budget and the risks. To do so, first the Pareto set of the solutions is obtained by NSGA-II technique - a multi-objective genetic algorithm technique - where a lexicographic model is used to select the optimal solution from the Pareto set based on the priority of the objective functions. A pessimistic strategy is used to account for the synergistic effects and the overlaps between the selected SMs.
Two mathematical models are developed to represent different policies in optimal SMs selection in a gas wellhead and surface facility. The results show a notable difference between the two policies, indicating the importance of setting proper objective functions in multi-objective optimization problems. The results also show that the methodology is able to effectively satisfy different safety management policies and constraints with no need for much extra information except the cost and impact of SMs on the hazards’ risk

Security management of critical infrastructures is a complex task as a great variety of technical and socio-political information is needed to realistically predict the risk of intentional malevolent acts. In the present study, a methodology based on Limited Memory Influence Diagram (LIMID) has been developed for the protection of critical infrastructures via cost-effective allocation of security measures. LIMID is an extension of Bayesian network (BN) intended for decision-making, allowing for efficient modelling of complex systems while accounting for interdependencies and interaction of system components. The probability updating feature of BN has been used to investigate the effect of vulnerabilities on adversaries’ preferences when planning attacks. Moreover, the proposed methodology has been shown to be able to identify an optimal defensive strategy given an attack through maximizing defenders’ expected utility. Despite being demonstrated via a chemical facility, the methodology can easily be tailored to a wide variety of critical infrastructures.

Success Likelihood Index Model (SLIM) is one of the widely-used methods in human reliability assessment especially when data is insufficient. However, this method suffers from uncertainty as it heavily relies on expert judgment for determining the model parameters such as the rates and weights of the performance shaping factors. The present study is aimed at using Bayesian Network (BN) for improving the performance of SLIM in handling the uncertainty arising from experts opinion and lack of data. To this end, SLIM is combined with BN to form the so-called BN-SLIM technique. We applied both SLIM and BN-SLIM models to a hypothetical example and compared the results. It is shown that BN-SLIM is able to provide a better estimation of human error probability by considering dependencies. The probability updating feature of BN-SLIM in particular makes it possible to use new information to update the prior beliefs about the rates of the performance shaping factors, thus updating the resultant human error probabilities.

Complex networks play a vital role in reliability analysis of real-world applications, demanding for precise and accurate analysis methods for optimal allocations of cost and reliability. Since the configuration of a system may change with every feasible solution of cost allocation optimization equation, finding the best arrangement of the system can become very challenging. This paper presents a novel methodology by combining Genetic Algorithm (GA) and Monte Carlo (MC) simulation approaches to simultaneously optimize cost allocation and system configuration in complex network. GA is used to generate configuration-cost pairs while MC is used to evaluate the reliability of the system for each pair. The application of the developed methodology is demonstrated for power grids as an example of critical complex networks. The results show that the proposed methodology can be readily used in practice.

Hazard and Operability analysis (HAZOP) is a popular technique for hazard identification and risk ranking in hazardous facilities. Conventional HAZOP, however, has some drawbacks: (i) it considers a limited number of risk factors, i.e., only the frequency and the severity of hazards; (ii) it assumes equal weights for the risk factors, thus ranking low-probability high-consequence hazards equally important as high-probability low-consequence hazards; and (iii) it uses crisp and precise data which is rarely available or highly uncertain, especially in the case of complex oil and gas facilities.
The present study is an attempt to alleviate the foregoing drawbacks of conventional HAZOP via a Fuzzy Multi-Attribute HAZOP technique (FMA-HAZOP). To do this, Analytic Hierarchy Process (AHP) and the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) are used, both in a fuzzy environment, to determine the weight of risk factors and to prioritize the hazards. The application of the FMA-HAZOP on a gas wellhead facility shows that FMA-HAZOP presents a more transparent and more detailed information about the rank of hazards compared to conventional HAZOP.

System safety and reliability assessment relies on historical data and experts opinion for estimating the required failure probabilities. When data comes from different sources, be it different databases or subject domain experts, the estimation of accurate probabilities would be very challenging, if not impossible, and subject to high epistemic uncertainty. In such cases, the use of imprecise probabilities to reflect the incomplete knowledge of analysts and their epistemic uncertainty is inevitable.
Evidence theory is an effective tool for manipulating imprecise probabilities. However, challenges in the assignment of prior belief masses and the lack of effective inference algorithms for combining and updating the belief masses have impeded the widespread application of evidence theory.
To address the foregoing issues, in the present study, (i) an innovative heuristic approach is developed to determine the prior belief masses based on the prior imprecise probabilities, and (ii) it is demonstrated how Bayesian network can be used for both propagating and updating the belief masses. In a nutshell, the developed methodology converts the prior imprecise probabilities into prior belief masses, propagates and updates the belief masses using Bayesian network, and back-transforms the predicted/updated belief masses to posterior imprecise probabilities.