Application of artificial neural network technique in modifying carbon-carbon composites

A matrix-modification process has great importance for carbon/carbon (C/C) composites. It is the main method to protect C/C composites from oxidation. As is well known, the matrix modification effects are influenced by many complicated factors, so a mathematical model cannot be exactly formulated. In this paper an artificial neural network (ANN) model is developed to predict the burning rate of matrix-modified C/C composites by the use of the Levenberg-Marquardt algorithm. The relationship between the modifying additives and burning rate is analyzed on the basis of the model. Results show that the relative error between the expected value and the predicted output of the network is less than 0.5%. Employing the ANN model, an optimized combination of these additives is obtained. The burning rate of the additive-optimized C/C composite decreases by 49.3%, which indicates that the ANN model is effective and feasible and could be used to reveal the relationships between the additive contents and the burning rate of C/C composites.