Abstract
A new optimization method for the design of axial turbine stages is described, which can search for the global optimum solution and realize the robust design. The feature of the new method include creating orthogonal arrays, changing the level values of design factors, and evaluating standard deviation of the objective function for every search step. As a result of applying the method to the optimum design of turbine stages, it is shown that the method can search the global optimum solution and the tolerances of design factors greatly affect the robustness of turbine efficiency.