Project C2 - M. Hinze and Thomas Rung: Hydrodynamic shape-optimization for turbulent industrial flows

Scientific Background and Motivation

Shape design is one of the most important drivers for the use of simulation approaches in fluids-engineering industry. Continuous improvements of computer performance and algorithm efficiency in conjunction with reductions of hardware expenses allow simulation approaches to rival experimental techniques for the analysis of turbulent flows in the framework of industrial design with many ten Mio. degrees of freedom. In comparison to experiments, Computational Fluid Dynamics (CFD) offers a competitive advantage if it is able to guide the analyst to a better design. Examples refer to the drag of an aircraft or ship and propulsive efficiency of aeroengines or propellers. However, when many design parameters are involved, the computational effort for the optimization of an existing shape quickly becomes prohibitive expensive. In order to deploy mathematical optimization strategies in an industrial framework, the associated time-to-solution has to be confined to a few days for the cycle.

Aims and Objectives

The focal point of the project is on adjoint shape-optimization techniques under the aegis of industrial CFD. The adjoint approach is most suitable for optimization of complex shapes, since the related effort decouples from the amount of shape parameters. Establishing efficient computational shape-optimization processes requires advanced strategies from various disciplines and cross-institutional collaboration. The resulting large-scale, non-linear optimization problem has to be discretized and approximated appropriately and is associated to a massive computational effort. Moreover, the shape modifications are constraint to practical realizability and must adhere to the existing design process.

PhD Student: Anne Gerdes, geb. Lincke. Thesis defended February 22, 2018

Publications

1. A. Lincke, A. Stück, and T. Rung: Derivation of an adjoint high-Reynolds number model for thermally driven flows. Proc. Appl. Math. Mech., 12: 681-682, 2012. doi: 10.1002/pamm.201210329.

2. A. Lincke, T. Rung: Adjoint-Based Sensitivity Analysis for Buoyancy-Driven Incompressible Navier-Stokes Equations with Heat Transfer, in B.H.V. Topping, (Editor), "Proceedings of the Eighth International Conference on Engineering Computational Technology", Civil-Comp Press, Stirlingshire, UK, Paper 100, 2012. doi:10.4203/ccp.100.100

3. A. Gerdes, M. Hinze, and T.Rung: An efficient line search technique and its application to adjoint topology optimisation. Proc. Appl. Math. Mech., 14: 719-720, 2014.

4. A. Gerdes, M. Hinze and T. Rung: Enhancement of adjoint topology optimization by approximated L1-regularization for promotion of sparse distributed control. Proc. Appl. Math. Mech., 16: 637–638, 2016,