Static Aeroelastic Scaling

Abstract

Classical aerodynamic wind tunnel tests are performed with stiff or rigid wind tunnel models. Over the years wing structures have become more flexible, and with the rise of non-conventional designs, such as strut-braced wings featuring high aspect ratios, even more flexible designs have risen. To capture the aerodynamic performance of these concepts under varying load conditions, resulting in different flight shapes, a static aeroelastically scaled wind tunnel model is required. The thesis research objective is to structurally design a static-aeroelastically scaled wind tunnel test model using optimization routines to find the optimal scaled aeroelastic characteristics representing the reference wing in wind tunnel test conditions while incorporating the safety constraints. Use cases included in the research are both an unconventional high aspect ratio strut-braced wing as well as a highly flexible cantilever wing.

To obtain aeroelastic and aerodynamic similarity, the aerodynamic similarity parameters should be the same and the elastic properties of the structure should adhere to static-aeroelastic scaling laws, which were found to solely depend on the length and pressure ratios. For finding a structure that has those elastic properties, design optimization is needed. After length scaling for the wind tunnel, little room for structural elements is left, especially in the slender strut-braced wing concept. For this reason, a composite shell with solid foam is used, as composites allow control over the directional stiffness. However, the discrete nature of composites, consisting of various layers with each a variable ply direction, complicates the optimization. To overcome this complication, lamination parameters are used to make the design space continuous, allowing the use of gradient-based solvers. An optimization framework was developed to optimize the lamination parameter space including a finite element solver in combination with scaled aeroelastic loads. The design variables considered in the optimization are the lamination parameters and the thicknesses of the laminate with the root mean square of the difference in deformation as objective. After optimizing in the lamination parameter space, the stacking sequence is retrieved using an open-source tool, OptiBless.

The results showed that the design space of the unconventional strut-braced wing configuration is highly non-convex which results in finding local minimum solutions. For a cantilever wing, the optimization is much more robust, as the optimization yields comparable results for multiple starting points which can be considered as the global optimum. The optimization framework developed was able to design a static aeroelastically scaled wind tunnel model, consistently resulting in a representative match in terms of deformation with the full-scale design for the cantilever wing. For the strut-braced wing, a representative match can also be found but these results were inconsistent. The optimization framework proved to be successful to optimize composite designs for cantilever wings. For non-conventional designs, this method is less powerful as using a great number of starting values is required, diminishing the computational benefit of using a gradient-based solver, and other strategies could be considered that may yield better results.