Predicting Coating Degradation Using Machine Learning and Pattern Recognition in Accelerated Testing

RIP2025-00138: Coating degradation is a critical issue that affects the long-term performance and durability of materials in harsh environments. In this study, we employed pattern recognition techniques to predict the extent of degradation in three distinct coatings based on image analysis. The primary degradation mechanisms investigated included water uptake, substrate activation, expansion, and delamination, all of which contribute significantly to the failure of protective coatings. Accelerated testing was conducted in fog chambers under controlled conditions following the ASTM B117 methodology, which simulates corrosive environments through prolonged exposure to salt spray.

Images of the coatings were collected at various intervals to capture the progression of degradation over time. These images were processed and analyzed to identify key patterns and features associated with different degradation phases. By employing advanced image processing and pattern recognition algorithms, we successfully classified the coatings' degradation states and established predictive models for their performance.

This approach demonstrates the potential of combining non-destructive imaging techniques with computational methods to provide a deeper understanding of coating behavior under accelerated test conditions. The findings contribute to the development of more efficient predictive tools for assessing coating durability and guiding the design of advanced protective materials.