Polymeric composites or reinforced plastics are materials made with a polymer matrix and a reinforcement. These materials have many advantages when compared to conventional engineering materials. Among the manufacturing methods for composite parts from continuous fiber reinforcement is the filament winding, often used to fabricate closed-surface structures such as tubes and tanks. In this work artificial neural networks, a computational tool inspired in the human brain, were applied in the filament winding process to predict the thermal behavior of composite tubes during the curing step. Information about the thermal behavior of composite parts may help in the selection of the appropriate cure cycle, which is one of the challenges in obtaining quality parts at low cost. Networks were trained with data obtained with the Lee-Springer model. The methodology was validated with experimental results from the literature.
Polymeric composites; thermal behavior; neural networks; modeling
