The optimization of process parameters of three-dimensional needled composites based on ANN and GA

To build the relationship between needled process parameters and C/C-SiC composites stiffness, and get better process parameters to improve the stiffness performances of composites, a surrogate model for optimizing process parameters of three dimensional needled preforms of C/C-SiC composites was established based on back propagation (BP) neural network combing with improved genetic algorithm. The stiffness prediction was realized by BP network. The predicted value of the network is almost identical with the finite element calculation, so the prediction accuracy of the model is high. The genetic strategy and optimization strategy of genetic algorithm were improved. Two improved genetic algorithms were used to optimize the process parameters. The obtained process parameters can significantly improve the stiffness of the material.