# A Robust Approach for Stability Analysis of Complex Flows Using Navier-Stokes Solvers

## TL;DR

This paper introduces a robust, Jacobian-free method for extracting global stability modes in complex 3D flows, reducing computational costs and handling diverse flow conditions effectively.

## Contribution

It proposes a new approach combining high-order schemes and dynamic mode decomposition to efficiently obtain global modes without iterative procedures or orthonormalization.

## Key findings

- Method accurately captures physically meaningful modes.
- Significant computational savings over Arnoldi-based methods.
- Applicable to complex, compressible, curvilinear flow domains.

## Abstract

Direct methods to obtain global stability modes are restricted by the daunting sizes and complexity of Jacobians encountered in general three-dimensional flows. Jacobian-free iterative approaches such as Arnoldi methods have greatly alleviated the required computational burden. However, operations such as orthonormalization and shift-and-invert transformation of matrices with appropriate shift guesses can stll introduce computational and parameter-dependent costs that inhibit their routine application to general three-dimensional flowfields. The present work addresses these limitations by proposing and implementing a robust, generalizable approach to extract the principal global modes, suited for curvilinear coordinates as well as the effects of compressibility. Accurate linearized perturbation snapshots are obtained using high-order schemes by leveraging the same non-linear Navier-Stokes code as used to obtain the basic state by appropriately constraining the equations using a body-force. It is shown that with random impulse forcing, dynamic mode decomposition (DMD) of the subspace formed by these products yields the desired physically meaningful modes when appropriately scaled. The leading eigenmodes are thus obtained without spurious modes or the need for an iterative procedure. Further, since orthonormalization is not required, large subspaces can be processed to capture converged low frequency or stationary modes. The validity and versatility of the method are demonstrated with numerous examples encompassing essential elements expected in realistic flows, such as compressibility effects and complicated domains requiring general curvilinear meshes. Favorable comparisons with Arnoldi-based method, complemented with substantial savings in computational resources show the potential of the current approach for relatively complex flows.