With the focus on the reduction of fossil fuel emissions, aircraft are continuously growing towards higher fuel efficiency. The traditional limits of aircraft performance can be surpassed through the use of composite materials which offer a reduction in aircraft weight. Due to the multidisciplinary nature of aircraft design, integration of different disciplinary analyses is required to arrive at a feasible design. The inclusion of composite design in the preliminary design process, however has become a challenge. This is due to the high computational cost associated with the composite aeroelastic tailoring tools used in the design process. A possible solution is available in the form of surrogate models which can reduce the computational costs. The current work focuses on the development of a methodology that allows the inclusion of surrogate-based model in the optimization for a computationally expensive aeroelastic tool (PROTEUS) developed at TU Delft. The resulting methodology can be expanded to a generic, computationally expensive tool in a multidisciplinary optimization setting. Wing design optimization is carried out based on surrogate modeling methodology and metal based design method. Comparison is made of the final optimized designs based on structural and performance parameters.

An analytical model for stiffness degradation of composite laminates with damage under static or cyclic loading is proposed. For static loading, two different modelling approaches have been created. Both account for shear-induced microscopic matrix damage and matrix cracking within each ply orientation of a laminate. In one approach the residual transverse tensile and shear moduli of a ply are determined by equating the energy density of the undamaged ply to that of the ply with damage, for a same applied load. Predictions excellent agreement with test results for cross-ply laminates, while stiffness degradation is for shear dominated layups. The other static approach applies energy equivalence only to the transverse tensile behavior of a ply. For shear, it predicts the residual shear modulus by accounting for the creation of permanent shear strains. Agreement with test results is excellent for cross-ply laminates, and excellent to good for shear dominated layups. The physical foundation behind this model is not as rigorous as the previous one, and as such needs more work. The proposed fatigue model assumes the matrix strength of a ply to be randomly distributed. Kassapoglou’s residual strength model for fatigue loading is used to express fatigue life as a function of matrix strength. The static model is then used to estimate stiffness degradation as a function the same matrix strength. Connecting the two, stiffness degradation as a function of fatigue cycles is obtained. Validation was performed on cross-ply laminates and agreement with results is excellent.