Artificial neural network modeling of mechanical behaviors of zea fiber-polyester composites

The tensile, flexural, and impact behaviors of zea fiber-reinforced polyester composites were evaluated. The different proportion of fiber length, fiber content in weight percentage, and molding pressure was used as process parameters in the fabrication of zea-polyester composites. The parameters and their levels were selected based on full factorial design as per the design of experiments. The composites were fabricated and tested as per ASTM standards. An artificial neural network algorithm was developed to predict the mechanical behaviors over the specified range of conditions, and the average absolute percentage of error of less than 4% was observed for tensile, flexural, and impact models.