Nonlinear beam optimization tool for integration into a Gradient-based Aero-structural Optimization Framework

The study conducted consists in the development of an open-source tool, able to perform linear and nonlinear structural weight optimization of a frame made of beam elements, whose sections are rectangular wingboxes reinforced with stiffeners.The software is made up of two modules : a core written in C++,where math calculations are done , and an external wrapping written in Python to promote the easy management of inputs, outputs and where the optimization is performed.The whole project relies on an existing homemade structural nonlinear beam solver , named PyBeam .The enlarging of this solver,initially able to solve just nonlinear structural analysis,was deemed necessary to obtain as outputs, values of objective function(weight), constraint and their gradients , giving in input the Design Variables (geometric parameters of the section).PyBeam incorporates an AD-based Adjoint solver for gradient computation, which relies on the AD library CoDiPack. Regarding the constraint, it has been computed Kreisselmeier-Steinhauser function to have just one aggregated constraint. Optimization is performed with openMDAO, an open-source Multidisciplinary design optimization (MDO) framework.Finally the optimization is validated, demonstrating the adherence to Karush–Kuhn–Tucker conditions by the use of Matlab symbolic tool.The main goal of this study is to provide an integrated infrastructure for structural design and optimization of wings.For the future development, this solver should be integrated in a bigger Aerostructural optimization framework.