Application of AD for the aero-structural optimization of wings

This workshop will focus on the coupled aerodynamic–structural optimization of wings, assisted by Algorithmic Differentiation (AD). Participants will learn how to set up and solve multidisciplinary optimization problems that account for both aerodynamic performance and structural constraints. A key aspect of the training will be the use of AD to efficiently compute the sensitivities required for gradient-based optimization techniques.

The workshop combines theoretical instruction with practical application. Teaching materials include the instructor’s lecture notes, key bibliographic references, and open-source software tools to support hands-on learning.

Overview

▪Introduction: new efficient wings, aerostructural coupling, and optimization

▪Optimization Foundations: definitions, classifications, numerical optimization methods (gradient-based optimization)

▪Sensitivity Analysis: Principles of gradient evaluation, including the adjoint method and an introduction to algorithmic differentiation.

▪Hands-On AD: practical session using an AD library on a simplified problem, comparing forward and reverse mode differentiation

▪MDO: introduction to MDO architectures and coupled gradient evaluation strategies

▪Coupled Gradients with AD: application of AD in evaluating gradients

▪High-Fidelity Aero-Structural Optimization: demonstration of a full aero-structural optimization workflow using high-fidelity models and AD-enhanced gradient evaluation