Application of Machine Learning for Predicting the Incubation Period of Water Droplet Erosion in Metals

Water droplet erosion (WDE) is a critical phenomenon that leads to material degradation in many engineering applications, particularly in power generation and aerospace industry. Accurate prediction of the incubation period is essential for optimizing material selection and maintenance strategies. Traditional empirical models, while helpful, often lack predictive accuracy due to their reliance on numerous parameters with limited physical interpretation. In this study, a machine learning (ML) approach was developed to predict the incubation period for different materials. A range of ML models—including linear regression (LR), decision tree regressor (DT), random forest regressor (RF), gradient boosting regressor (GBR), and artificial neural networks (ANN)—was employed to capture the complex relationships between material properties and erosion conditions. Despite hyperparameter optimization using techniques such as grid search, no substantial improvement in model predictions was observed. Data transformation methods—logarithmic, Yeo-Johnson, and Box-Cox transformations—were applied to enhance model performance. A dataset derived from experimental measurements on five different alloys was used to train and validate the ML models. The results indicate that the ML models significantly outperform conventional empirical approaches. Notably, the LR model with Box-Cox transformation achieved an R² (coefficient of determination) of over 90% and low mean absolute error (MAE), while the ANN model with Yeo-Johnson transformation attained an R² above 85% with a correspondingly low MAE. Additionally, feature impact and importance analyses provided insights into the key factors influencing the duration of the incubation period, further validating the robustness of the models. This study offers a robust tool for predicting the incubation period of WDE, with broad applicability in engineering design and material selection across various industries.