An Adjoint Method for the Optimal Boundary Control of Turbulent Flows Modeled with the Rans System

Abstract

In recent years, the optimal control in fluid dynamics has gained attention for the design and the optimization of engineering devices. One of the main challenges concerns the application of the optimal control theory to turbulent flows modeled by the Reynolds averaging Navier-Stokes equations. In this work we propose the implementation of an optimal boundary control problem for the ReynoldsAveraged Navier-Stokes system closed with a two-equations turbulence model. The optimal boundary velocity is sought in order to achieve several objectives such as the enhancement of turbulence or the matching of the velocity field over a well defined domain region. The boundary where the control acts can be the main inlet section or additional injection holes placed along the domain. By minimizing the augmented Lagrangian functional we obtain the optimality system comprising the state, the adjoint, and the control equations. Furthermore, we propose numerical strategies that allow to solve the optimality system in a robust way for such a large number of unknowns.

Full document