The use of anticorrosive coatings has been a powerful method to be applied on the surface of metallic materials to mitigate the corrosive process. In this study, the focus is composite coatings that are commonly used on the internal surface of storage tanks in petrochemical industries. The development of non-destructive methods for inspection of faults in this field is desired due to unhealthy access and mainly because undercoating corrosion is difficult to detect by visual inspection. Pulsed thermography (PT) was employed to detect undercoating corrosion and adhesion loss of anticorrosive composite coatings defects. Additionally, a computational simulation model was developed to complement the PT tests. According to the experimental results, PT was able to detect all types of defects evaluated. The results obtained by computational simulation were compared with experimental ones. Good correlation (similarity) was verified, regarding both the defect detection and thermal behavior, validating the developed model. Additionally, by reconstructing the thermal behavior according to the defect parameters evaluated in the study, it was estimated the limit of the remaining thickness of the defect for which it would be possible to obtain its detection using the pulsed modality.
Detection of Defects in Additively Manufactured Metallic Materials with Machine Learning of Pulsed Thermography Images
