離散定性検査結果からのリスク評価のためのベイズ法を用いた破損確率評価手法の検討

抄録

In this study, evaluation method of the probability of failure of a structure from qualitative inspection data, such as visual inspection, is proposed. In recent years, the aging of social infrastructure structures and the resulting increase in maintenance costs have become a problem. RBM is a method of optimiz-ing the timing and content of maintenance by defining the risk as the product of the probability of failure of the structure and the consequence of the failure. For the Risk Based Maintenance, optimization is done by homogenizing risk, which requires the calculation of risk, i.e., the probability of failure. Corrosion is a problem in outdoor structures such as bridges and roadway ancillary facili-ties. However, unlike structures in factories, quantitative thickness control is not used in many target structures, and only qualitative inspections such as visu-al inspections are used to rank the thickness. However, in such inspections, un-like in the case of control by thickness measurement, a method for evaluat-ing the probability of failure from inspection results has not been established. In this paper, we propose a method to evaluate the quantitative wall thin-ning rate from discrete qualitative inspection data using Bayesian estimation, and examine the effectiveness of the proposed method by comparing it with the evaluation of quantitative inspection data. The method is based on the probabilistic estimation of the inspection results obtained from any combi-nation of the year when the wall thinning started and the wall thinning rate. In this paper, the effectiveness of the proposed method, the evaluation accuracy, and the influence of the qualitative method are clarified by examining the evaluation of the qualitative inspection classified into four ranks, the number of rank divi-sions in the qualitative inspection, and the influence of the rank division method on the evaluation accuracy.