Optimal Design for Mechanical Properties of Composite with Combination of Genetic Algorithm and Artificial Neural Network

摘要

Improved genetic algorithm, combined with artificial neural network, is present for the optimal design of 2.5D braided composite. Dispersal simulation data, including maximal stresses and elastics properties, are adopted by artificial neural network for the calculation of strength property. Based on calculation method of strength mentioned above and other calculation models for other mechanical properties, genetic algorithm is employed for the design of structure parameters of 2.5D braided composite, such as wrap fiber density, fill fiber density and interface strength. These structure optimal parameters are finally optimized for practical application.