Genetic and gradient-based algorithms for the multi-objective optimization of aircraft design with aeroelastic constraints

This work focuses on the implementation of a complex multidisciplinary computational environment for the design of aircraft. The MDA (MultiDisciplinary Analysis) is comprised of high-fidelity structural, aeroelastic and mission analyses. This allows designers to account for the aeroelastic constraints early in the design process. For this purpose, an aero-structural tool, called FEMWING, has been developed within the present research activity. It is able to handle the whole structural and aeroelastic assessment for the entire aircraft, comprised of wing, fuselage and tail. It can first create the structural and aerodynamic models, then perform the analyses through a Finite-Element commercial solver, and finally analyze the results. Furthermore, MDO problem is solved through a Multi-Objective Optimization (MOO). The optimization criteria include minimum structural weight, maximum lift-to-drag ratio and maximum mission range, taking into account also aeroelastic constraints. The 3-dimensional Pareto frontier is built by applying two different optimization methods. Specifically, a gradient-based algorithm (through Weighted Global Criterion method, WCG) and a stochastic algorithm (Multi Objective Genetic Algorithm, MOGA) are used for solving the multi-objective problem and their results are compared. The explicit multi-objective approach is proved to be more eective then the single-objective approach when the design space exploration is needed for the optimization of a new design.