Domino Effect Analysis Using Bayesian Networks

A new methodology is introduced based on Bayesian network both to model domino effect propagation patterns and to estimate the domino effect probability at different levels. The flexible structure and the unique modeling techniques offered by Bayesian network make it possible to analyze domino effects through a probabilistic framework, considering synergistic effects, noisy probabilities, and common cause failures. Further, the uncertainties and the complex interactions among the domino effect components are captured using Bayesian network. The probabilities of events are updated in the light of new information, and the most probable path of the domino effect is determined on the basis of the new data gathered. This study shows how probability updating helps to update the domino effect model either qualitatively or quantitatively. The methodology is applied to a hypothetical example and also to an earlier‐studied case study. These examples accentuate the effectiveness of Bayesian network in modeling domino effects in processing facility.     

Exploring Risk Judgments in a Trade Dispute Using Bayesian Networks

Bayesian networks (BNs) are graphical modeling tools that are generally recommended for exploring what‐if scenarios, visualizing systems and problems, and for communication between stakeholders during decision making. In this article, we investigate their potential for exploring different perspectives in trade disputes. To do so, we draw on a specific case study that was arbitrated by the World Trade Organization (WTO): the Australia‐New Zealand apples dispute. The dispute centered on disagreement about judgments contained within Australia's 2006 import risk analysis (IRA). We built a range of BNs of increasing complexity that modeled various approaches to undertaking IRAs, from the basic qualitative and semi‐quantitative risk analyses routinely performed in government agencies, to the more complex quantitative simulation undertaken by Australia in the apples dispute. We found the BNs useful for exploring disagreements under uncertainty because they are probabilistic and transparently represent steps in the analysis. Different scenarios and evidence can easily be entered. Specifically, we explore the sensitivity of the risk output to different judgments (particularly volume of trade). Thus, we explore how BNs could usefully aid WTO dispute settlement. We conclude that BNs are preferable to basic qualitative and semi‐quantitative risk analyses because they offer an accessible interface and are mathematically sound. However, most current BN modeling tools are limited compared with complex simulations, as was used in the 2006 apples IRA. Although complex simulations may be more accurate, they are a black box for stakeholders. BNs have the potential to be a transparent aid to complex decision making, but they are currently computationally limited. Recent technological software developments are promising.     

Assessment of Catastrophic Risk Using Bayesian Network Constructed from Domain Knowledge and Spatial Data

Prediction of natural disasters and their consequences is difficult due to the uncertainties and complexity of multiple related factors. This article explores the use of domain knowledge and spatial data to construct a Bayesian network (BN) that facilitates the integration of multiple factors and quantification of uncertainties within a consistent system for assessment of catastrophic risk. A BN is chosen due to its advantages such as merging multiple source data and domain knowledge in a consistent system, learning from the data set, inference with missing data, and support of decision making. A key advantage of our methodology is the combination of domain knowledge and learning from the data to construct a robust network. To improve the assessment, we employ spatial data analysis and data mining to extend the training data set, select risk factors, and fine‐tune the network. Another major advantage of our methodology is the integration of an optimal discretizer, informative feature selector, learners, search strategies for local topologies, and Bayesian model averaging. These techniques all contribute to a robust prediction of risk probability of natural disasters. In the flood disaster's study, our methodology achieved a better probability of detection of high risk, a better precision, and a better ROC area compared with other methods, using both cross‐validation and prediction of catastrophic risk based on historic data. Our results suggest that BN is a good alternative for risk assessment and as a decision tool in the management of catastrophic risk.     

Improving Default Risk Prediction Using Bayesian Model Uncertainty Techniques

Credit risk is the potential exposure of a creditor to an obligor's failure or refusal to repay the debt in principal or interest. The potential of exposure is measured in terms of probability of default. Many models have been developed to estimate credit risk, with rating agencies dating back to the 19th century. They provide their assessment of probability of default and transition probabilities of various firms in their annual reports. Regulatory capital requirements for credit risk outlined by the Basel Committee on Banking Supervision have made it essential for banks and financial institutions to develop sophisticated models in an attempt to measure credit risk with higher accuracy. The Bayesian framework proposed in this article uses the techniques developed in physical sciences and engineering for dealing with model uncertainty and expert accuracy to obtain improved estimates of credit risk and associated uncertainties. The approach uses estimates from one or more rating agencies and incorporates their historical accuracy (past performance data) in estimating future default risk and transition probabilities. Several examples demonstrate that the proposed methodology can assess default probability with accuracy exceeding the estimations of all the individual models. Moreover, the methodology accounts for potentially significant departures from “nominal predictions” due to “upsetting events” such as the 2008 global banking crisis.     

Sensitivity Analysis, Monte Carlo Risk Analysis, and Bayesian Uncertainty Assessment

Standard statistical methods understate the uncertainty one should attach to effect estimates obtained from observational data. Among the methods used to address this problem are sensitivity analysis, Monte Carlo risk analysis (MCRA), and Bayesian uncertainty assessment. Estimates from MCRAs have been presented as if they were valid frequentist or Bayesian results, but examples show that they need not be either in actual applications. It is concluded that both sensitivity analyses and MCRA should begin with the same type of prior specification effort as Bayesian analysis.     

Analysis of Regional Scale Risk of Whirling Disease in Populations of Colorado and Rio Grande Cutthroat Trout Using a Bayesian Belief Network Model

Introduction and spread of the parasite Myxobolus cerebralis, the causative agent of whirling disease, has contributed to the collapse of wild trout populations throughout the intermountain west. Of concern is the risk the disease may have on conservation and recovery of native cutthroat trout. We employed a Bayesian belief network to assess probability of whirling disease in Colorado River and Rio Grande cutthroat trout (Oncorhynchus clarkii pleuriticus and Oncorhynchus clarkii virginalis, respectively) within their current ranges in the southwest United States. Available habitat (as defined by gradient and elevation) for intermediate oligochaete worm host, Tubifex tubifex, exerted the greatest influence on the likelihood of infection, yet prevalence of stream barriers also affected the risk outcome. Management areas that had the highest likelihood of infected Colorado River cutthroat trout were in the eastern portion of their range, although the probability of infection was highest for populations in the southern, San Juan subbasin. Rio Grande cutthroat trout had a relatively low likelihood of infection, with populations in the southernmost Pecos management area predicted to be at greatest risk. The Bayesian risk assessment model predicted the likelihood of whirling disease infection from its principal transmission vector, fish movement, and suggested that barriers may be effective in reducing risk of exposure to native trout populations. Data gaps, especially with regard to location of spawning, highlighted the importance in developing monitoring plans that support future risk assessments and adaptive management for subspecies of cutthroat trout.     

Using Bayesian Networks to Model Expected and Unexpected Operational Losses

This report describes the use of Bayesian networks (BNs) to model statistical loss distributions in financial operational risk scenarios. Its focus is on modeling "long" tail, or unexpected, loss events using mixtures of appropriate loss frequency and severity distributions where these mixtures are conditioned on causal variables that model the capability or effectiveness of the underlying controls process. The use of causal modeling is discussed from the perspective of exploiting local expertise about process reliability and formally connecting this knowledge to actual or hypothetical statistical phenomena resulting from the process. This brings the benefit of supplementing sparse data with expert judgment and transforming qualitative knowledge about the process into quantitative predictions. We conclude that BNs can help combine qualitative data from experts and quantitative data from historical loss databases in a principled way and as such they go some way in meeting the requirements of the draft Basel II Accord (Basel, 2004) for an advanced measurement approach (AMA).     

Application of Bayesian Networks in Quantitative Risk Assessment of Subsea Blowout Preventer Operations

This article proposes a methodology for the application of Bayesian networks in conducting quantitative risk assessment of operations in offshore oil and gas industry. The method involves translating a flow chart of operations into the Bayesian network directly. The proposed methodology consists of five steps. First, the flow chart is translated into a Bayesian network. Second, the influencing factors of the network nodes are classified. Third, the Bayesian network for each factor is established. Fourth, the entire Bayesian network model is established. Lastly, the Bayesian network model is analyzed. Subsequently, five categories of influencing factors, namely, human, hardware, software, mechanical, and hydraulic, are modeled and then added to the main Bayesian network. The methodology is demonstrated through the evaluation of a case study that shows the probability of failure on demand in closing subsea ram blowout preventer operations. The results show that mechanical and hydraulic factors have the most important effects on operation safety. Software and hardware factors have almost no influence, whereas human factors are in between. The results of the sensitivity analysis agree with the findings of the quantitative analysis. The three‐axiom‐based analysis partially validates the correctness and rationality of the proposed Bayesian network model.     

A Methodology for Risk Analysis Based on Hybrid Bayesian Networks: Application to the Regasification System of Liquefied Natural Gas Onboard a Floating Storage and Regasification Unit

This article presents an iterative six‐step risk analysis methodology based on hybrid Bayesian networks (BNs). In typical risk analysis, systems are usually modeled as discrete and Boolean variables with constant failure rates via fault trees. Nevertheless, in many cases, it is not possible to perform an efficient analysis using only discrete and Boolean variables. The approach put forward by the proposed methodology makes use of BNs and incorporates recent developments that facilitate the use of continuous variables whose values may have any probability distributions. Thus, this approach makes the methodology particularly useful in cases where the available data for quantification of hazardous events probabilities are scarce or nonexistent, there is dependence among events, or when nonbinary events are involved. The methodology is applied to the risk analysis of a regasification system of liquefied natural gas (LNG) on board an FSRU (floating, storage, and regasification unit). LNG is becoming an important energy source option and the world's capacity to produce LNG is surging. Large reserves of natural gas exist worldwide, particularly in areas where the resources exceed the demand. Thus, this natural gas is liquefied for shipping and the storage and regasification process usually occurs at onshore plants. However, a new option for LNG storage and regasification has been proposed: the FSRU. As very few FSRUs have been put into operation, relevant failure data on FSRU systems are scarce. The results show the usefulness of the proposed methodology for cases where the risk analysis must be performed under considerable uncertainty.     

Lessons learned from Toulouse and Buncefield disasters: from risk analysis failures to the identification of atypical scenarios through a better knowledge management

The recent occurrence of severe major accidents has brought to light flaws and limitations of hazard identification (HAZID) processes performed for safety reports, as in the accidents at Toulouse (France) and Buncefield (UK), where the accident scenarios that occurred were not captured by HAZID techniques. This study focuses on this type of atypical accident scenario deviating from normal expectations. The main purpose is to analyze the examples of atypical accidents mentioned and to attempt to identify them through the application of a well-known methodology such as the bow-tie analysis. To these aims, the concept of atypical event is accurately defined. Early warnings, causes, consequences, and occurrence mechanisms of the specific events are widely studied and general failures of risk assessment, management, and governance isolated. These activities contribute to outline a set of targeted recommendations, addressing transversal common deficiencies and also demonstrating how a better management of knowledge from the study of past events can support future risk assessment processes in the identification of atypical accident scenarios. Thus, a new methodology is not suggested; rather, a specific approach coordinating a more effective use of experience and available information is described, to suggest that lessons to be learned from past accidents can be effectively translated into actions of prevention.     

Aggregate Human Health Risk Assessment from Dust of Daily Life in the Urban Environment of Beijing

Because of the high emissions of polycyclic aromatic hydrocarbons (PAHs) into the environment by the increasing number of vehicles in Beijing and the absorption of these PAHs onto particulates, the performance of a preliminary health risk assessment of the aggregate exposure to PAHs of urban citizens in daily life is very important. Urban dust can be used to examine the aggregation of atmospheric particulates from local pollution sources over a long time period and the direct exposure of the urban human population. The environment's correlative with clothing, dining, residing, and traveling in urban daily life was assessed using exposure‐receptor‐oriented analysis. The multipathway exposure model was used to simulate the lifetime exposure of a female citizen to PAHs in dust. All of the PAH concentrations in dust for each behavior and its correlative environment in Beijing were acceptable because all of the carcinogenic risks of PAHs in the dust were approximately 1.0 × 10^–6. The dominant induced carcinogenic risks in the dust were Benzo(a)pyrene and Dibenzo(a,h)anthracene. The main carcinogenic risk routes for humans were dermal contact and oral intake, which contributed on average 99.78% of the risk. Indoor risk is especially important, as the decoration and height within the building were important impact factors for carcinogenic risk induced by indoor PAHs. For people living in an urban area, a healthy lifestyle includes less decoration per room, living on a low floor, wearing a respirator, and reducing exposed skin area when traveling.     

Risk Analysis for U.S. Offshore Wind Farms: The Need for an Integrated Approach

Wind power is becoming an increasingly important part of the global energy portfolio, and there is growing interest in developing offshore wind farms in the United States to better utilize this resource. Wind farms have certain environmental benefits, notably near‐zero emissions of greenhouse gases, particulates, and other contaminants of concern. However, there are significant challenges ahead in achieving large‐scale integration of wind power in the United States, particularly offshore wind. Environmental impacts from wind farms are a concern, and these are subject to a number of on‐going studies focused on risks to the environment. However, once a wind farm is built, the farm itself will face a number of risks from a variety of hazards, and managing these risks is critical to the ultimate achievement of long‐term reductions in pollutant emissions from clean energy sources such as wind. No integrated framework currently exists for assessing risks to offshore wind farms in the United States, which poses a challenge for wind farm risk management. In this “Perspective”, we provide an overview of the risks faced by an offshore wind farm, argue that an integrated framework is needed, and give a preliminary starting point for such a framework to illustrate what it might look like. This is not a final framework; substantial work remains. Our intention here is to highlight the research need in this area in the hope of spurring additional research about the risks to wind farms to complement the substantial amount of on‐going research on the risks from wind farms.     

Bayesian Data Analysis of Severe Fatal Accident Risk in the Oil Chain

We analyze the risk of severe fatal accidents causing five or more fatalities and for nine different activities covering the entire oil chain. Included are exploration and extraction, transport by different modes, refining and final end use in power plants, heating or gas stations. The risks are quantified separately for OECD and non‐OECD countries and trends are calculated. Risk is analyzed by employing a Bayesian hierarchical model yielding analytical functions for both frequency (Poisson) and severity distributions (Generalized Pareto) as well as frequency trends. This approach addresses a key problem in risk estimation—namely the scarcity of data resulting in high uncertainties in particular for the risk of extreme events, where the risk is extrapolated beyond the historically most severe accidents. Bayesian data analysis allows the pooling of information from different data sets covering, for example, the different stages of the energy chains or different modes of transportation. In addition, it also inherently delivers a measure of uncertainty. This approach provides a framework, which comprehensively covers risk throughout the oil chain, allowing the allocation of risk in sustainability assessments. It also permits the progressive addition of new data to refine the risk estimates. Frequency, severity, and trends show substantial differences between the activities, emphasizing the need for detailed risk analysis.     

Risk Management of Domino Effects Considering Dynamic Consequence Analysis

Domino effects are low‐probability high‐consequence accidents causing severe damage to humans, process plants, and the environment. Because domino effects affect large areas and are difficult to control, preventive safety measures have been given priority over mitigative measures. As a result, safety distances and safety inventories have been used as preventive safety measures to reduce the escalation probability of domino effects. However, these safety measures are usually designed considering static accident scenarios. In this study, we show that compared to a static worst‐case accident analysis, a dynamic consequence analysis provides a more rational approach for risk assessment and management of domino effects. This study also presents the application of Bayesian networks and conflict analysis to risk‐based allocation of chemical inventories to minimize the consequences and thus to reduce the escalation probability. It emphasizes the risk management of chemical inventories as an inherent safety measure, particularly in existing process plants where the applicability of other safety measures such as safety distances is limited.     

Quantitative Risk Assessment from Farm to Fork and Beyond: A Global Bayesian Approach Concerning Food-Borne Diseases

A novel approach to the quantitative assessment of food-borne risks is proposed. The basic idea is to use Bayesian techniques in two distinct steps: first by constructing a stochastic core model via a Bayesian network based on expert knowledge, and second, using the data available to improve this knowledge. Unlike the Monte Carlo simulation approach as commonly used in quantitative assessment of food-borne risks where data sets are used independently in each module, our consistent procedure incorporates information conveyed by data throughout the chain. It allows back-calculation in the food chain model, together with the use of data obtained downstream in the food chain. Moreover, the expert knowledge is introduced more simply and consistently than with classical statistical methods. Other advantages of this approach include the clear framework of an iterative learning process, considerable flexibility enabling the use of heterogeneous data, and a justified method to explore the effects of variability and uncertainty. As an illustration, we present an estimation of the probability of contracting a campylobacteriosis as a result of broiler contamination, from the standpoint of quantitative risk assessment. Although the model thus constructed is oversimplified, it clarifies the principles and properties of the method proposed, which demonstrates its ability to deal with quite complex situations and provides a useful basis for further discussions with different experts in the food chain.     

Risk Analysis for Unintentional Slide Deployment During Airline Operations

We present a risk analysis undertaken to mitigate problems in relation to the unintended deployment of slides under normal operations within a commercial airline. This type of incident entails relevant costs for the airline industry. After assessing the likelihood and severity of its consequences, we conclude that such risks need to be managed. We then evaluate the effectiveness of various countermeasures, describing and justifying the chosen ones. We also discuss several issues faced when implementing and communicating the proposed measures, thus fully illustrating the risk analysis process.     

Eliciting and Combining Decision Criteria Using a Limited Palette of Utility Functions and Uncertainty Distributions: Illustrated by Application to Pest Risk Analysis

Utility functions in the form of tables or matrices have often been used to combine discretely rated decision‐making criteria. Matrix elements are usually specified individually, so no one rule or principle can be easily stated for the utility function as a whole. A series of five matrices are presented that aggregate criteria two at a time using simple rules that express a varying degree of constraint of the lower rating over the higher. A further nine possible matrices were obtained by using a different rule either side of the main axis of the matrix to describe situations where the criteria have a differential influence on the outcome. Uncertainties in the criteria are represented by three alternative frequency distributions from which the assessors select the most appropriate. The output of the utility function is a distribution of rating frequencies that is dependent on the distributions of the input criteria. In pest risk analysis (PRA), seven of these utility functions were required to mimic the logic by which assessors for the European and Mediterranean Plant Protection Organization arrive at an overall rating of pest risk. The framework enables the development of PRAs that are consistent and easy to understand, criticize, compare, and change. When tested in workshops, PRA practitioners thought that the approach accorded with both the logic and the level of resolution that they used in the risk assessments.     

基于贝叶斯网络的操作风险预警机制研究

鉴于商业银行的操作风险管理具有样本量小、结构复杂等特点,较难运用传统方法构建风险预警系统,本文采用贝叶斯网络.研究了商业银行操作风险预警系统的建模过程并根据我国商业银行操作风险的外部数据给出了算例.通过构建由关键风险指标和关键风险诱因组成的商业银行操作风险拓扑结构,本文分析了各类风险指标对操作风险的作用形式,在对各类节点赋值的基础上,通过贝叶斯推理,建立起商业银行操作风险的预警系统.以便在出现可能导致巨额损失时,商业银行能够及时采取措施化解操作风险.     

A Bayesian Method to Mine Spatial Data Sets to Evaluate the Vulnerability of Human Beings to Catastrophic Risk

Vulnerability of human beings exposed to a catastrophic disaster is affected by multiple factors that include hazard intensity, environment, and individual characteristics. The traditional approach to vulnerability assessment, based on the aggregate‐area method and unsupervised learning, cannot incorporate spatial information; thus, vulnerability can be only roughly assessed. In this article, we propose Bayesian network (BN) and spatial analysis techniques to mine spatial data sets to evaluate the vulnerability of human beings. In our approach, spatial analysis is leveraged to preprocess the data; for example, kernel density analysis (KDA) and accumulative road cost surface modeling (ARCSM) are employed to quantify the influence of geofeatures on vulnerability and relate such influence to spatial distance. The knowledge‐ and data‐based BN provides a consistent platform to integrate a variety of factors, including those extracted by KDA and ARCSM to model vulnerability uncertainty. We also consider the model's uncertainty and use the Bayesian model average and Occam's Window to average the multiple models obtained by our approach to robust prediction of the risk and vulnerability. We compare our approach with other probabilistic models in the case study of seismic risk and conclude that our approach is a good means to mining spatial data sets for evaluating vulnerability.     

Bayesian Inference for Reliability of Systems and Networks Using the Survival Signature

The concept of survival signature has recently been introduced as an alternative to the signature for reliability quantification of systems. While these two concepts are closely related for systems consisting of a single type of component, the survival signature is also suitable for systems with multiple types of component, which is not the case for the signature. This also enables the use of the survival signature for reliability of networks. In this article, we present the use of the survival signature for reliability quantification of systems and networks from a Bayesian perspective. We assume that data are available on tested components that are exchangeable with those in the actual system or network of interest. These data consist of failure times and possibly right‐censoring times. We present both a nonparametric and parametric approach.