Shape Optimization of Turbine Stage Using Adaptive Range Differential Evolution and Three-Dimensional Navier-Stokes Solver

Abstract

A new optimization method named as Adaptive Range Differential Evolution (ARDE) is proposed and developed for the turbine stage design. The mathematical tests are used to demonstrate the optimization performance of the present ARDE through compared with the Simple Genetic Algorithms (SGA) and the Differential Evolution (DE). Combined with the ARDE, surface modeling method and Navier-Stokes solver, a low aspect ratio transonic turbine stage is optimized, with 28 design variables in total, for the maximization of the isentropic efficiency. The optimization design of this case is performed on the cluster parallel Personal Computers. The optimal design turbine stage shows a better aerodynamic performance than that of the reference design while meeting the strength requirement. The robustness and reliability of the presented ARDE for the turbomachinery optimization design are also illustrated.