Unsupervised classification of organic coating performance using electrochemical impedance data

Abstract In this study, the use of unsupervised machine learning for grouping organic coatings performance during corrosion testing is evaluated. Coatings with various pigments and pigment volume concentrations (PVC) were applied to AA2024-T3 and cold-rolled mild steel. The coated specimens were immersed in selected environments and periodically monitored with electrochemical impedance spectroscopy. Post-testing, the specimens were classified into three performance groups based on their appearance. The KMeans algorithm was employed to group the systems using three datasets based on the electrochemical. Comparing KMeans results with performance-based groups yielded an accuracy of 73%, highlighting both the potential and limitations of this approach. Graphical abstract