A Robust Blade Design Method based on Non-Intrusive Polynomial Chaos Considering Profile Error

摘要: To weaken the influence of profile error on compressor aerodynamic performance, especially pressure ratio and efficiency, a robust design method considering is built to improve robustness performance blade. The characteristics are random small-scaled, which means that evaluate blade time-intensive high-precision work. For this reason, non-intrusive polynomial chaos (NIPC) Kriging surrogate model introduced in efficiency uncertainty quantification (UQ) ensure accuracy. satisfies Gaussian distribution, NIPC carried out do since it has advantages prediction accuracy get statistical behavior error. In integrand points NIPC, several models established based Latin hypercube sampling (LHS) + Kriging, further reduces costs optimization by replacing calling computational fluid dynamic (CFD) repeatedly. results show can significantly shorter time (40 times faster) without losing accuracy, meaningful engineering application reduce manufacturing cost premise ensuring performance. Mechanism analysis improvement samples current work help find key parameter dominating under disturbance error, robustness.