Droplet Spreading Prediction of the Textured Surfaces: A Machine Learning Approach

Surface wettability plays a crucial role in a wide range of applications, including waterproof coatings, medical diagnostics, microfluidics, and the textile industry. It is typically characterized by contact angle measurements and droplet spreading behavior on surfaces. However, traditional methods for assessing surface wettability often come with limitations, such as the need for complex experimental setups and high costs. To address these challenges, machine learning (ML) offers a promising alternative for making accurate and rapid predictions of surface wettability properties. In this study, we developed and evaluated several regression models, including Support Vector Regression (SVR), K-Nearest Neighbors (KNN), Extreme Gradient Boosting (XGB), Random Forest Regression (RFR), and Artificial Neural Networks (ANN). Our analysis focused on predicting droplet spreading dynamics on a variety of surfaces, using a dataset of 2,500 samples generated through Lattice Boltzmann Method (LBM) simulations. The performance of these models was rigorously assessed using key metrics such as Root Mean Squared Error (RMSE), Mean Squared Error (MSE), Mean Absolute Percentage Error (MAPE), and the coefficient of determination ($\mathrm{R}^{2}$). Among the models, Random Forest Regression stood out, delivering exceptional performance with an $\mathrm{R}^{2}$ score of 0.999 during training and 0.991 during testing. The model’s superior performance, reflected in its low RMSE and MAE values, underscores the potential of ML to surpass traditional methods, providing a faster, more data-driven approach to predicting and analyzing droplet spreading behavior on surfaces.