Design Optimization of Complex Flowfields Using Evolutionary Algorithms and Hybrid Unstructured CFD

Abstract

In this paper we present design optimization studies of complex fluid dynamic problems involving viscous, turbulent, multi-phase flow phenomena utilizing evolutionary algorithms. The shape optimization process is carried out by utilization of high fidelity CFD based comprehensive framework. The framework comprises of a genetic algorithm based design optimization procedure coupled to the hybrid unstructured CRUNCH CFD code and a grid generator. The genetic algorithm based optimization procedure is very robust, and searches the complex design landscape in an efficient and parallel manner. Furthermore, it can easily handle complexities in constraints and objectives and is disinclined to get trapped in local extrema regions. The fitness evaluations are carried out through a RANS based hybrid unstructured solver. The utilization of hybrid unstructured methodology provides flexibility in incorporating large changes in design and mesh regeneration is carried out in an automated manner through a scripting process within the grid generator GRIDGEN. The design optimization procedure is discussed in the context of two representative problems involving multi-element airfoils and cavitating airfoils. The design optimization studies revealed interesting insights on the influence of the stabilizer trailing edge shape on gap flow and lift and torque characteristics of the flap. Design optimization studies in multiphase regimes also correlate the length of cavitation zones with airfoil shapes and its influence on lift.