Reliability updating for slope stability of dikes : approach with fragility curves (background report)

Slope stability assessments of dikes, just like most geotechnical problems, are typically dominated by the large uncertainties in soil properties, often resulting in rather large estimated probabilities of (slope) failure compared to the actual failure rates observed in the field. Observations of past performance such as survival of significant loading can be incorporated to improve such reliability estimates. The present report describes how survival of observed load conditions can be taken into account quantitatively. In particular, a simplified approach with fragility curves is introduced in order to enable practitioners to adjust slope failure probabilities to field observations. Besides a scientific description of the theory, the report contains considerations for applying the approach to dike stability problems. Moreover, it describes how the results can be used in the Dutch safety assessment framework for primary flood defenses. In essence, the updated probability of failure obtained from the analysis can be directly compared to the target probability of failure for a specific dike section as defined in the statutory safety assessment of Dutch primary flood defenses (WBI-2017). Four fictitious benchmark examples illustrate the application of the method to cases with a varying degree of complexity, each highlighting different aspects of the method. All examples confirm that the results obtained with the simplified approach with fragility curves reasonably match the results by conventional reliability methods such as Monte Carlo simulation.