A life prediction model for coatings based on the statistical analysis of hot salt corrosion performance

Abstract

This paper examines the use of a probabilistic approach to corrosion analysis to develop a life prediction model for coating systems subject to a pitting mode of attack. The analysis is used to predict the rate of growth of type II hot salt corrosion pits, formed under marine service conditions. Examination of engine components has shown that particulate impaction can lead to oxide scale fracture at temperatures close to and below 700°C. These sites of local scale disruption can act as initiation points for the type II hot salt corrosion reaction. The growth of the observed pitting morphologies has been modelled using extreme value statistics and this forms the basis of the life model. Hot salt corrosion tests using a dynamic burner rig are used to evaluate the time dependent growth of type II corrosion for pack aluminised coatings on IN738 at 700°C. The resulting corrosion profiles are multimodal and hence cannot be described by a single statistical model. Areas of maximum attack follow extreme value statistics and hence the time dependent variation in the extreme value model parameters can be assessed. From these studies a coating life prediction model is developed. The predicted corrosion rates using this model agrees well with published service data.