In this paper, the Levenberg-Marquardt algorithm was applied to predict the dynamic-mechanical properties of carbon-fiber reinforced epoxy composite. The Standard Test Method for Measuring Vibration-Damping Properties of Materials (ASTM E-756) was used for a cantilever beam that gives experimental curves of magnitude versus time. The composite is used in aeronautics and its configuration is [0º/45º/90º/0º]S. A multilayered neural network perceptron (MLP) was used, and the results showed that the application of the Levenberg-Marquardt learning algorithm leads to a high predictive quality to epoxy composites, i.e. nearly 64% of standard error of prediction was found to be >0.9. The initial tests considered a simple architecture 2-[30-30]2-1 resulting in low predictive quality. However, increasing the number of neurons in the hidden layers resulted in an enhancement in an optimized architecture composed by 2-[100-100]2-1 neurons.
Artificial neural networks (ANN); composites; carbon fiber; dynamic-mechanical properties
