Machine Learning-Based Corrosion Rate Prediction for Zinc and Steel Using Real-Time Corrosion Sensors in Accelerated Corrosion Test Conditions

Electrical Resistance (ER) corrosion sensors are widely used for real-time monitoring of corrosion in laboratory and field conditions. Combining data acquired by ER sensors with machine learning (ML) offers new perspectives for the development of models aiming to predict atmospheric corrosion. A predictive model was therefore developed using key environmental parameters from more than ten accelerated corrosion tests, trained using machine learning techniques, and validated through a specific corrosion test. The results demonstrate the potential of this method that can then provide deeper insight into reaction dynamics and the impact of climatic conditions on corrosion rates within cyclic corrosion tests.