Prediction of corrosion behaviour of Alloy 22 using neural network as a data mining tool

Objective of this work was to develop an algorithm to predict behaviour of corrosion resistant metal alloys using a supervised neural network method as a data mining tool. We studied corrosion data on a nickel-based alloy, Alloy 22 which is of great industrial interest. This is an extension of a previously reported study on metallic glasses, carbon steel, and grade-2 titanium. The data mining results allow us to categorize and prioritize certain parameters (i.e. pH, temperature, time of exposure, electrolyte composition, metal composition, etc.) and help us understand the synergetic effects of the parameters and variables on corrosion behaviour.

Research highlights
► A supervised neural network (NN) model showed excellent agreement with experimental data on localized corrosion of Alloy 22.
► Developed NN models using AC impedance data on Alloy 22 predicted that, under applied potentials >550 mVSCE, pit growth is significant. This observation is consistent with experiments.
► Temperature effect on corrosion rate of Alloy 22 was shown to follow Arrhenius rule.