Predicting Wear Rate and Friction Coefficient of Li2Si2O5 Dental Ceramic Using Optimized Artificial Neural Networks

Abstract

The tribological properties of dental materials, such as wear and friction, are crucial for ensuring their long-term reliability and performance. Traditional experimental approaches, while accurate, are often resource intensive and time consuming, prompting a need for efficient computational methods. This study explores the application of artificial neural networks (ANNs) to predict the tribological behavior of dental ceramic lithium disilicate (IPS e.max Cad). A genetic algorithm (GA) was used to optimize the ANN’s hyperparameters, improving its ability to model complex, nonlinear relationships between input variables, including normal load and velocity and output properties such as wear rate and friction coefficients. By integrating experimental data with an ANN, this study identifies key factors influencing tribological performance, reducing the dependency on extensive experimental testing. The results demonstrate that the optimized ANN model accurately predicts tribological behavior, offering a robust framework for material optimization. These findings emphasize the potential of combining ANNs and GAs to enhance the understanding and design of dental materials, accelerating innovation while addressing the challenges of traditional evaluation methods. This research underscores the transformative role of advanced computational approaches in tribology and material science.