Conference Paper

Combining Static/Dynamic Fault Trees and Event Trees Using Bayesian Networks

DOI: 10.1007/978-3-540-75101-4_10 Conference: Computer Safety, Reliability, and Security, 26th International Conference, SAFECOMP 2007, Nuremberg, Germany, September 18-21, 2007.
Source: DBLP


In this study, an alternative approach for combining Fault Trees (FT) and Event Trees (ET) using capabilities of Bayesian
networks (BN) for dependency analysis is proposed. We focused on treating implicit and explicit weak s-dependencies that may
exist among different static/dynamic FTs related to an ET. In case of combining implicit s-dependent static FTs and ET that
combinatorial approaches fail to get the exact result, the proposed approach is accurate and more efficient than using Markov
Chain (MC) based approaches. In case of combining implicit weak s-dependent dynamic FTs and ET where the effect of implicit
s-dependencies have to be manually inserted into the MC, the proposed approach is more efficient for getting an acceptable

21 Reads
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper discusses how non-marked Petri nets can be used to model and analyze event trees where the pivotal (branching) events are dependent and modeled by fault trees. The dependencies may, for example, be caused by shared utilities, shared components, or general common cause failures that are modeled by beta-factor models. These dependencies are cumbersome to take into account when using standard event-/fault tree modeling techniques, and may lead to significant errors in the calculated end-state probabilities of the event tree if they are not properly analyzed. A new approach is proposed in this paper, where the whole event tree is modeled by a non-marked Petri net and where P-invariants, representing the structural properties of the Petri net, are used to obtain the frequency of each end-state of the event tree with dependencies. The new approach is applied to a real example of an event tree analysis of the Strahov highway tunnel in Prague, Czech Republic, including two types of dependencies (shared Programmable Logic Controllers and Common Cause Failures).
    Reliability Engineering [?] System Safety 08/2012; 104:45–57. DOI:10.1016/j.ress.2012.03.013 · 2.41 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: In this paper, we review briefly the different works published in the field of Dynamic Bayesian Network (DBN) reliability analyses and estimation, and we propose to use DBNs as a tool of knowledge extraction for constructing DBN models modeling the reliability of systems. This is doing, by exploiting the data of (tests or experiences feedback) taken from the history of the latter’s. The built model is used for estimating the system reliability via the inference mechanism of DBNs. The proposed approach has been validated using known system examples taken from the literature.
    Ref. No: 978-3-319-00560-7, Year: 04/2013
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper describes how dynamic simulations of a manufacturing process can be used to construct informed prior distributions for the failure probabilities of alarm and safety interlock systems. Bayesian analysis is used starting with prior distributions and enhancing them with likelihood distributions, constructed from real-time alarm data, to form posterior distributions, which are used to estimate failure probabilities. The use of alarm data to build likelihood distributions has previously been investigated[2,11]. Rare-event historical data are typically sparse and have high-variance likelihood distributions. When high-variance likelihood distributions are combined with typical high-variance prior distributions, the resulting posterior distributions naturally have high variances preventing reliable failure predictions. In contrast with prior distributions obtained by maximizing entropy[19] and those that are based on expert knowledge[2], this paper introduces a new repeated-simulation method to construct informed prior distributions having smaller variances, which in turn leads to posterior distributions with lower variances and a more reliable prediction of the failure probabilities of alarm and safety interlock systems. The application of the proposed method is demonstrated for offline dynamic risk analysis of a steam-methane reformer (SMR) process.
    Industrial & Engineering Chemistry Research 02/2015; 54(16):150216151052007. DOI:10.1021/ie5038769 · 2.59 Impact Factor