Design optimization of 3D nozzle shapes using CFD and parallel simulated annealing

Abstract

Aerodynamic shape design of 3D nozzles using global optimization methods and CFD solver usually require enormous computational resources for searching the global optimum design. The focus of this work is on the application of a parallel simulated annealing algorithm (PSA) implemented on multi-process or computer system to optimize the aerodynamic shape design objective functions which are evaluated by numerically solving the 3D Euler and Navier-Stokes equations using a TVD approach and a LU implicit scheme. For illustrative purposes the thrust of a 3D high speed internal flow system is taken as an objective function and the shape of the internal flow system is represented using bicubic splines. A series of tests are conducted to identify the efficiency and speedups using the PSA. The results show that PSA can be implemented as an efficient global optimizer for aerodynamic shape design and can serve as a robust tool resulting in considerable reductions in wall-clock time on multiple processors.