Shape Optimization of a Labyrinth Seal Applying the Simulated Annealing Method

Abstract

In the present work, an environment for the shape optimization of a labyrinth seal is described. A program for a parameterized, automated grid generation is coupled with a commercial Computational Fluid Dynamics (CFD) flow solver and an optimization algorithm. Standard optimization strategies, like gradient-based methods, mostly are trapped to local optima. Therefore, the simulated annealing method is applied. It allows the finding of global minima or maxima of arbitrary functions.
The presented optimization method is used to minimize the leakage through a three-finned, stepped labyrinth seal. For the optimization, the step position and the step height are chosen to be variable. The characteristics of the flow fields of selected seal configurations are compared and discussed against the background of the leakage behavior of the seal.