ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering

Forum papers are thought-provoking opinion pieces or essays founded in fact, sometimes containing speculation, on a civil engineering topic of general interest and relevance to the readership of the journal. The views expressed in this Forum article do not necessarily reflect the views of ASCE or the Editorial Board of the journal.

Failures in rail fasteners can lead to misalignments of the rails and even cause a train derailment. Current inspection and maintenance regimes for rail fasteners, however, do not adequately address the credible failure modes found in the field. In response to these improvement opportunities, a risk-based maintenance philosophy, driven by a risk management framework, is proposed for rail fasteners. The framework was primarily developed from international standards alongside high-level industry guidelines. Reliability tools were then incorporated, allowing practitioners to arrive at an appropriate combination of reliability tools based on the circumstances under which the assessment is to be conducted. Monte Carlo simulations were undertaken on the embedded anchors of rail fasteners to demonstrate how the resultant framework can be innovatively adopted in practice. The general findings highlight that accurate risk depiction is vital for track components (e.g., embedded anchors, the failure modes of which are dependent on time); thereby, the time frames in which risk for the component transits to different risk categories should be obtained. Note that the finding is unique to the example; thus, the proposed risk framework should be treated carefully before it is applied to other failure modes.

This paper compiled 159 centrifuge tests for foundations in clay with interbedded sand where punch-through failure was observed. The accuracy of initial- and failure-stress-dependent methods was quantified by the mean and coefficient of variation (COV) of a model factor defined as the ratio between the measured and calculated punch-through capacity. When the friction angle is reduced with consideration of the effect of sand dilation, both the initial- and failure-stress-dependent methods can accurately estimate the punch-through capacity of foundations in clay with sand. Numerical analyses were performed to verify the observation. The mean of the model factor is close to 1, and the COV is smaller than 0.3. For multilayer clays interbedded by carbonate sand, they could significantly overestimate the punch-through capacity. This is because carbonate sand has considerably higher compressibility, which is not incorporated into the initial- and failure-stress-dependent methods.

A fragility function quantifies the probability that a structural system exposed to a given hazard exceeds an undesirable limit state event conditioned on the occurrence of a hazard level. Multiple sources of uncertainty affect this function, including record-to-record variation, geometric and material properties, aging, modeling assumptions and errors, and even the analyzed dataset. This study presents a methodology for statistical model selection and uncertainty quantification of seismic fragility functions. The statistical models are created by implementing a hierarchical Bayesian framework with a sequential Monte Carlo technique. The most probable model is selected using Bayesian model selection. This model is validated through multiple metrics using predictive intervals and the Kolmogorov-Smirnov test. Then, the epistemic uncertainty is quantified as the variance of the area under the fragility functions. The methodology is implemented on a twenty-story steel benchmark model case study, demonstrating that the log-normal distribution yields superior performance relative to other models considered. Finally, further analysis of the case study demonstrates that the epistemic uncertainty is considerably reduced when using forty observations.

Risk identification in the case of technical projects is an issue that has been extensively studied in many previous research efforts; however, a thorough study of the respective literature clearly reveals a methodological and application gap between research and practice, which is further increased by a wide discrepancy of definitions and procedures applied. This paper proposes the consideration of risk sources in the risk identification process as a means to bridge the identified gap and reduce discrepancy. The paper presents the results of the processing and analysis of real projects’ risk data through unstructured interviews with the experts who provided the data, and examines their correlation with a comprehensive risk sources checklist drawn from previous research. The results of this cross-checking clearly highlighted the extensive misconceptions of the related terms and notions in practice (e.g., what is perceived as risk identification is actually risk source identification), as well as the partial implementation of procedures that reflects a similar partial approach to technical project risk analysis in practice (e.g., the risk sources actually identified are fewer than those that could be identified through a holistic approach to technical project risk analysis). These results draw attention to the need for more-focused risk source–related research, because the new paradigm shift in technical project risk analysis highlights the importance of risk sources.

Many civil infrastructure systems that deliver resources from source points to sinks, e.g., power distribution and gas pipeline networks, can be described as acyclic directed networks comprising nodes and links. Reliability assessment of these systems can be challenging, particularly for systems of increasing size and complexity and if the probabilities of rare events are of interest. This paper proposes a new analytical probability propagation method for reliability assessment of acyclic directed networks called the directed probability propagation method (dPrPm). Through a link-adding sequence to propagate a message consisting of the marginal and pairwise node reliabilities from source nodes to sink nodes, the method results in the upper and lower bounds of all sink node reliabilities. Reliability of a sink node is measured by the probability of reaching that node from a source node. Compared with previous methods, dPrPm addresses the case of multiple-sink networks, results in guaranteed reliability bounds, and analyzes acyclic directed networks as relevant for infrastructure systems. Proofs are provided guaranteeing the accuracy of dPrPm, and computation time is significantly reduced from typical exponential increases with system size to a polynomial increase. To assess performance, the proposed method was applied to three test applications: a directed grid network, a power distribution network, and a more complex gas pipeline network under seismic hazard. Results were compared with the exact solution and Monte Carlo simulations to evaluate accuracy and computational cost. Results showed that dPrPm performs equally well in terms of accuracy across network reliabilities and achieved order-of-magnitude increases in computational efficiency to obtain exact bounds on reliability assessments at all system sink nodes.