Simplified Eigenvalue Analysis for Turbomachinery Aerodynamics with Cyclic Symmetry

TL;DR

This paper introduces a computationally efficient method for eigenvalue analysis of cyclic symmetric turbomachinery flows by exploiting symmetry to reduce problem size, enabling faster analysis without losing accuracy.

Contribution

A novel approach that leverages cyclic symmetry to simplify eigenvalue computations in turbomachinery aerodynamics, reducing computational resources required.

Findings

Significant reduction in CPU time and memory usage.

Accurate eigenvalue spectrum obtained from a single sector.

Validated on a 22-sector annular compressor cascade.

Abstract

Eigenvalue analysis is widely used for linear instability analysis in both external and internal aerodynamics. It typically involves finding the steady state, linearizing around it to obtain the Jacobian, and then solving for its eigenvalues and eigenvectors. When the flow is modelled with Reynolds-averaged Navier--Stokes equations with a large boundary-layer-resolving mesh, the resulting eigenvalue problem can be of very high dimensions, and is thus computationally very challenging. To reduce the computational cost, a simplified approach is proposed to compute the eigenvalues and eigenvectors, by exploiting the cyclic symmetric nature of annular fluid domain for typical compressors. It is shown that via a rotational transformation, the Jacobian can be reduced to a block circulant matrix, whose eigenvalues and eigenvectors then can be computed using only one sector of the entire domain. This simplified approach significantly lowers the memory overhead and the CPU time of the eigenvalue analysis without compromising on the accuracy. The proposed method is applied to the eigenvalue analysis of an annular compressor cascade with 22 repeated sectors and it is shown the spectrum of the whole annulus can be obtained by using the information of 1 sector only, demonstrating the effectiveness of the proposed method.