Optimization of labyrinth seal leakage with independently varied tooth parameters using efficient global optimization

Abstract In this study, the Kriging model is employed alongside the Efficient Global Optimization (EGO) method to optimize the sealing performance of a labyrinth seal structure. Two optimization strategies are explored: one involving the consistent variation of design parameters and the other, independent variation. A comparative evaluation of the two approaches demonstrates their effectiveness, followed by an in-depth analysis of the leakage characteristics under varying operating conditions, both before and after optimization. Additionally, a sensitivity analysis is conducted for the independent optimization scheme to examine the influence of design variables on the objective function. The results indicate that both optimization schemes significantly improve the sealing performance, with the equivalent flow rate reduced by 14.5 % and 18.47 %, respectively, compared to the baseline design. The optimized models maintain a sealing advantage under variable operating conditions. Furthermore, the sensitivity analysis reveals key design parameters influencing leakage performance, providing valuable insights for the structural optimization of labyrinth seals.