Duct Shape Optimization Using Multi-Objective and Geometrically Constrained Adjoint Solver

In the recent years, adjoint optimization has gained popularity in the automotive industry with its growing applications. Since its inclusion in the mainstream commercial CFD solvers and its continuously added capabilities over the years, its productive usage became readily available to many engineers who were previously limited to interfacing the customized adjoint source code with CFD solvers. The purpose of this work is to demonstrate using an adjoint solver a method to optimize duct shape that meets multiple design objectives simultaneously. To overcome one of the biggest challenges in the duct design, i.e. the severe packaging constraints, the method here uses geometrically constrained adjoint to ensure that the optimum shape always fits into the user-defined packaging space. In this work, adjoint solver and surface sensitivity calculations are used to develop the optimization method. The development is carried out within commercial CFD solver settings, and a java macro is then utilized to automate the developed optimization cycles. The issues during the optimization process, mostly related to the morphing of mesh, are addressed and the solutions to avoid the issues are proposed. For test cases, an air-induction system duct and a rear floor duct are used for the shape optimization where it is aimed to reduce the pressure drop and the duct size simultaneously. The predictions show a significant reduction in both cases through a small shape change. After the optimization, the rear floor duct shapes from the optimized design and the baseline design are 3D-printed to perform the air-flow bench test. The test results showed good agreement against the CFD results, confirming the validity of the optimization method.