Surface Wettability Modeling and Predicting via Artificial Neural Networks

Abstract

Surface wettability, defined by the contact angle, describes the ability of a liquid to spread over, absorb or adhere to a solid surface. Surface wetting analysis is important in many applications, such as lubrication, heat transfer, painting and wherever liquids interact with solid surfaces. The behavior of liquids on surfaces depends mainly on the texture and chemical properties of the surface. Therefore, these studies show the possibility of modeling surface wettability by adjusting the parameters of the surface texturing process. The prediction of the contact angle describing the wettability of the surface was performed using artificial neural networks. In order to select the most effective prediction model, the activation functions of neurons, the number of hidden layers and the network training algorithms were changed. The neural network model presented in these studies is capable of predicting the contact angle with an efficiency defined by the coefficient of determination R² between real and predicted contact angles of over 0.9.

Keywords: artificial neural networks; contact angle; manufacturing; roughness; wettability.

Figure 1

The schematic view of the model for predicting the wettability of the surface.

Figure 2

Views of three examples of 3D images of the surface after EDM with discharge energies per single pulse: (a) 11,385 µJ, (b) 217,500 µJ, (c) 3,506,400 µJ.

Figure 3

Determination of coefficients R² between (a) the discharge energy in the EDM process and (b) the surface topographic characterization parameters and contact angles.

Figure 4

Schematic view of prediction tasks performed by artificial neural networks. 1—prediction of surface topography parameters based on EDM parameters, 2—prediction of surface wettability based on surface topography parameters, 3—prediction of surface wettability based on EDM parameters.

Figure 5

Results of a neural network for predicting surface topography parameters based on EDM parameters. The green background color indicates the parameters of the neural network with the best coefficient of determination R² between the experimental and predicted results.

Figure 6

Results of a neural network for predicting surface wettability based on surface topography parameters. The green background color indicates the parameters of the neural network with the best coeffi-cient of determination R² between the experimental and predicted results.

Figure 7

Results of a neural network for predicting surface wettability based on EDM parameters. The green background color indicates the parameters of the neural network with the best coeffi-cient of determination R² between the experimental and predicted results.