Prediction of salt spray test results of micro arc oxidation coatings on AA2024 alloys by combination of accelerated electrochemical test and artificial neural network

Abstract

Micro arc oxidation (MAO) coating was prepared on AA2024 as a substitute of sulfuric acid anodizing (SAA) with the aim to improve anticorrosion behavior. Evaluation of corrosion resistance was first performed by means of salt spray test (SST) as a conventional but long way in time method, usually more than 1000 h for aeronautical needs. In parallel, an original approach combining electrochemical characterization and Artificial Intelligence was evaluated in order to predict corrosion occurrence in SST. The electrochemical setup consists in an improvement of ISO17463 accelerated cyclic electrochemical method (ACET) with an additional linear polarization step. The whole generates up to 50 parameters that are treated by artificial neural network (ANN). After training and refinement, the average accuracy of predicted corrosion resistance and corroded surface after 200 h of SST are 99 and 90% respectively. This promising combination of ACET and ANN enables the evaluation of corrosion behavior of coating in less than 24 h.