| A limitation of the current reliability model has been that the failure probability was based on assumed deterministic and probabilistic models, which may not necessarily represent the state of Nature. The probability obtained in the current model is therefore at best a "calculated probability". Observed performances can provide valuable information to calibrate this calculated probability for the model error that may have been accrued. This thesis presents a Bayesian procedure in which observed performance records, effect of system deterioration, and other unaccounted for factors can be incorporated systematically to convert the calculated reliability index of a given system to an estimate of the real probability of failure. The probability will correct for the under or over conservatism of the true probability by the index. A method is also proposed to assess the uncertainty associated with the calculated probability, resulting from various factors including the assumed probabilistic models and the availability of performance records for updating. The method was applied to the stability performance of soil cut slopes in Hong Kong. Through this updating process, the reliability estimate of a given slope (of similar type and under similar environmental condition) and its associated confidence range can be realistically assessed to provide inputs for a more defensible decision. |
