Optimizing the operation safety and performance of an axial compressor using fluid-structure coupling and high-performance computing

Abstract

In this study, the impact of blade radial and axial deformations on the operation safety and performance of an axial compressor is analyzed using a partitioned fluid-structure coupling approach. High performance computing (HPC) clusters and the Message Passage Interface (MPI) parallelization method are utilized to optimize the mesh quality and computation time balance. The chosen high-resolution tip-clearance mesh is validated through a mesh convergence study on the fluid domain. Three turbulence models (k-ε, k-ω, and k-ω SST) are compared and the k-ε turbulence model is found to be the best option for agreement with experimental data. A multilevel factorial design of experiments (DOE) is conducted to investigate the influence of tip-clearance variation on the operation safety and performance of the compressor. A parametric study for several tip-clearance values and materials is performed using ANSYS Workbench, and the maximum deformation in the blade tip was predicted to be 0.7 mm, and the optimum design point is determined based on the weight and importance of the factors, which leads to an increase over 33% in the operation safety and a negligible loss in efficiency. The study also highlights the 42 times computational time savings obtained through the use of super-processors with high quality mesh in both solid and fluid domains in comparison with a personal computer. Future work could investigate the impact of other factors such as blade geometry, or operating conditions on the performance and safety of the compressor in a two-way strongly coupled transient manner.

Keywords

• NASA Compressor rotor 67
• Fluid-structure interaction
• HPC cluster
• MPI parallelization
• DOE
• Response optimization