Efficient Global Optimization of Helicopter Rotor Blades for Vibration Reduction in Forward Flight

The effectiveness of surrogate based optimization of helicopter rotor blades for vibration reduction using kriging global approximations is investigated. The search for the optimal design is conducted using two methods: (1) the “one-shot ” approach in which the surrogate is optimized directly and (2) a method based on the expected improvement function which seeks to find optimal designs while reducing the uncertainty in the surrogate. The results indicate that the more global search based on the expected improvement function can lead to more optimal rotor blade designs with less function evaluations than the “one-shot” approach, and thus is more efficient. In addition, the results indicate that the application of parallel computation to the expected improvement function based method aids in finding more optimal blade designs even if the kriging approximations are not accurate everywhere in the design space. Nomenclature c Blade chord CW Helicopter weight coefficient Cd0 Blade profile drag coefficient Cdf Flat plate drag coefficient D Vector of design variables EIF (x) Expected improvement function f(x) Assumed polynomials which account for the ‘global ’ behavior in krigin