The static aeroelastic torsion divergence problem is the main obstacle to bring forward swept wing into massive applications. The aeroelastic tailoring technique-based radial basis function neural networks (RBFNNs) and genetic algorithm (GA) optimization in MATLAB considering the material orientation, thickness, and lay-up are elucidated in the present work. RBFNNs are used to build a surrogate model between the composite parameters and structure displacement, which is proved robust and accurate. Then an optimal structure is obtained by GA global search based on RBFNNs model with the weight constrain. The displacements of the forward swept wing caused by an approximate aerodynamic load are decreased 32.5% through finite element method (FEM) static structural analysis. The modal analysis illustrates that the first mode frequency increases by 33.0% and the second mode increases by 37.9%. A computational aeroelasticity approach is developed by in-house Hybrid Unstructured Reynolds-Averaged Navier-Stokes solver associating an open source FEM code – Calculix. The results of coupling calculations show effectiveness of aeroelastic tailoring optimization of composite forward swept wing without weight penalty. The results obtained demonstrate that for the forward swept wing, the most violent situation appears around Mach Number 1.0 where the aeroelastic tailoring optimization could decrease the torsion angle by nearly 70.0%. The torsion of forward swept wing will increase at subsonic and decrease at supersonic with the increase of velocity.
Multidisciplinary optimization of an NLF forward swept wing in combination with aeroelastic tailoring using CFRP
