Abstract
It is difficult to clarify the detailed propagation mechanisms of back-side metal loss of bottom floors because this task is complicated by several factors, including the soil and the environment. Therefore, authors assumed the propagation of back-side metal loss of the bottom floors to be a stochastic process and applied the directed percolation model to the phenomena. The back-side metal loss generated on the bottom floors was simulated using the directed percolation model with various percolation probabilities. It was then verified that similar risk curves for the dataset measured with the continuous thickness measurement device were created with a simulated dataset. Using the simulated dataset, the dependency of number of data for calculating the Corrosion Risk Parameter (CRP) obtained from the statistical analysis using the measured dataset was discussed. Compared with the CRP calculated from the simulated dataset and the maximum depth of the clusters, the more than 1, 000 data were needed for the calculation of the appropriate CRP.