KSEF R&D Excellence: Application of Genetic Algorithms to Flow Control Simulation

The next generation of aircraft will increasingly rely on flow control techniques to maximize performance. One potential system for active flow control is the use of a grid of micro-jets over the surface of an airfoil to control the shape of boundary layer. The challenge with such a system is developing the control system that will allow the jets to adjust rapidly and optimally for changing flow conditions. The proposed project will aim to develop a prototype system to train such a flow control approach. This project will combine the technologies of computational fluid dynamics (CFD), genetic algorithms (GA), and commodity cluster computing. Computational fluid dynamics provide the simulations. of the flow over the airfoil, determining the effects various jet configurations. Genetic algorithms allow the design space to be searched efficiently for the optimal configurations. Commodity clusters allow for parallel computations at relatively low costs. By aggressive optimization of the GA-CFD code and code-aware design of the cluster, we aim to considerably enhance our computational power and make larger problems more feasible. Current research is this area has been limited -the computational cost is high to do each simulation and genetic algorithms have rarely been used in conjunction with realistic CFD calculations. The proposed research will be on the leading edge in the area ofGA-CFD technology, a field that will likely have applications beyond the area of flow control.