Parameter optimization of the composite honeycomb tip in a turbine cascade

Abstract

The geometric parameters of the composite honeycomb tip have been extracted and optimized using Kriging model and genetic algorithm numerically. A total of 36 sets of input-response combination are included from full factor design. The Kriging model for the tip leakage mass flow rate is established and assessed by leave-one-out cross validation. The computation of the optimal composite honeycomb tip case has been completed to affirm the mathematical precision. Moreover, the tests of flat tip, normal honeycomb tip and optimal composite honeycomb tip were carried out in a low-speed wind tunnel. The results show that the combination of Kriging model and genetic algorithm is suitable for the present geometry optimization of the composite honeycomb tip. The optimal composite honeycomb tip reduces the tip leakage mass flow rate by up to 16.81%. In detail, the effects of these geometric parameters are discussed using the surrogate database. The pressure field on the casing is employed to analyze the blocking effect of these cavities on the tip leakage flow. The pressure distributions around the blade surface are also captured. Besides, the downstream secondary velocity streamlines and loss contours are compared. The simulated averaged loss is reduced by 5.49% in the optimal case.