ABSTRACT
The paper presents a procedure for numerically modelling the surface of corroded specimens. A 3D laser scan is utilised to derive surface morphology and corrosion pit depth distribution. This corrosion depth is analyzed with a probability model to assess the morphology difference and the characteristic of the corrosion pit depth at a particular volumetric change. Furthermore, the goodness fit test statistic is carried out to observe the propensity of corrosion depth to a specific distribution, i.e. the Gaussian and non-Gaussian. The corroded specimens that conform to Gaussian distribution are numerically modelled with ANSYS APDL to generate a Gaussian surface. Furthermore, the non-Gaussian surface model is simplified as a single corrosion pit with various geometrical shapes. The mesh convergence is carried out to provide accurate stress distribution . The procedure is adjustable and applicable to the other surface morphology.
Acknowledgment
The author would like to acknowledge the German academic exchange service (DAAD) for generous financial support.
Disclosure statement
No potential conflict of interest was reported by the author(s).