Bayesian Inference Method for Failure Probability of Wall Thinning Pipe in Corrosion Rate Fluctuation Model

To rationalize the inspection interval for the wall-thinning piping element, the linear-Bayes method was proposed in the previous paper. To derive the simple formula, the linear-Bayes method ignores the corrosion rate change against time. However, this change may be caused by the one of the operational environment. Therefore, without the sufficient monitoring of the environment, the linear-Bayes method may underestimate the failure probability. In this paper, the linear-Bayes method is extended for the wall-thinning model with the corrosion rate fluctuation, which imitates the unexpected change of the corrosion rate. The extension is carried out through following two approaches: the “correction-term” and the “error-term” approaches. The correction-term approach can evaluate the change of corrosion rate, however, it requires sufficient number of inspections. The error-term approach evaluates the failure probability conservatively.