Extracting dominant dynamics about unsteady base flows
Alec J. Linot, Barbara Lopez-Doriga, Yonghong Zhong, Kunihiko Taira · 2025
Read Extracting dominant dynamics about unsteady base flows by Alec J. Linot, Barbara Lopez-Doriga, Yonghong Zhong, Kunihiko Taira (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-24075.txt.
Introduction
Alec Linot and co-authors review how to pull out the leading dynamics of flows that never settle, in a 2025 paper. Many methods exist for steady or repeating base flows, but unsteady, non repeating flows got far less attention. Such flows appear when a pipe's flow rate jumps, when a wing stalls, or when a vortex forms.
For these cases, knowing the base flow's own motion is not enough; the features that grow around it matter too. The review covers data driven decompositions, operator based methods, and causality analysis, among other tools. It closes with an outlook that flags where new work is most needed. The paper serves as a map of a young field.
Catalog record
| CREATOR | Alec J. Linot, Barbara Lopez-Doriga, Yonghong Zhong, Kunihiko Taira |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 80 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-24075
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Extracting dominant dynamics about unsteady base flows Alec J. Linot, Barbara Lopez-Doriga, Yonghong Zhong, Kunihiko Taira Abstract A wide range of techniques exist for extracting the dominant flow dynamics and features about steady, or periodic base flows. However, there have been limited efforts in extracting the dominant dynamics about unsteady, aperiodic base flow. These flows appear in many applications such as when there is a sudden change in the flow rate through a pipe, when an airfoil experiences stall, or when a vortex forms. For these unsteady flows, it is valuable to know not only the dynamics of the base flow but also the features that form around this base flow. Here, we discuss the current state of research on extracting important flow structures and their dynamics in such cases with time-varying base flows. In particular, we consider data-driven decompositions, operator-based methods, causality analysis, and some other approaches. We also offer an outlook and call attention to key areas that require future efforts. June 2025 1 Introduction Unsteady flows exhibit rich dynamics and spatial patterns. Modern simulations and experiments can capture these unsteady flow fields with remarkable spatial and temporal resolutions, but the underlying dynamics may still be difficult to interpret. Instead of attempting to track every flow structure and its temporal evolution, it is beneficial to extract the dominant low-dimensional dynamics that play an important role in the overall behavior of the system. In fact, our research community has been successful in extracting such dominant flow structures, for which modal analysis has served as a foundation. Data-driven and operator-based modal analysis techniques [ 1 , 2 , 3 , 4 , 5 ] have been widely used for a variety of research efforts [ 6 ] . Particularly noteworthy are the proper orthogonal decomposition [ 7 , 8 ] , dynamic mode decomposition [ 9 ] , stability analysis [ 10 , 11 , 12 , 13 ] , and resolvent analysis [ 14 , 15 , 16 ] . With these techniques, the research community has made great strides in analyzing fluid flows with steady base states. However, the examination of complex flow dynamics with respect to time-varying base states, as shown in figure 1 , remains a challenge. The difficulty stems from the lack of theories and analysis techniques available for time-varying base states, as well as the guidelines on how to choose the time-varying base states themselves. Floquet analysis is one method for analyzing perturbation dynamics about time-periodic base flows [ 17 , 18 , 19 ] , but this method does not readily extend to aperiodic time-varying base flows. Figure 1 : Perturbation dynamics about (left) steady and (right) unsteady base flows. Base flows are shown in black and the perturbations on top of the base flow are shown in blue. In fact, many fluid-based systems possess unsteady base flows, whose acceleration and deceleration significantly influence the overall flow dynamics. For example, notable acceleration and deceleration occur in both internal flows, such as pipe flow [ 20 ] , and external flows around moving bodies, such as maneuvering airfoils [ 21 ] . Even when there does not exist a clear transient force driving the system, transient periods of acceleration or deceleration are observed in many examples, including in turbulent channel flows [ 22 , 23 ] , mixing in a tank [ 24 , 25 ] , dynamic stall [ 26 ] , and unpredictable weather patterns [ 27 , 28 ] . Here, we call attention to the shortcomings of the standard techniques established for steady base flows when applied to time-varying base flows. We also present existing methods that can extract dominant, low-dimensional flow dynamics and spatial flow structures with respect to time-varying flows. Analysis techniques for unsteady base flows are far less prevalent than the methods for steady base flows listed above. We hope that calling attention to these approaches will expand their use and motivate the development of new methods. Also presented in this paper are the challenges associated with studying unsteady flows, including the determination of the unsteady base flow, examination of the time-varying flow evolution operator, and the greater computational requirements compared to analysis for flows with steady base states. In what follows, we discuss our perspective on recent approaches for analyzing unsteady flows. First, we outline some fundamental challenges with applying analysis techniques to flows with unsteady base states and with finding time-varying base flows in section 2 . Then, we present data-driven modal approaches in section 3.1 , operator-based approaches (nonmodal stability analysis, optimally time-dependent modes analysis, and input-output analysis) in sections 3.2 - 3.4 , causality analysis in section 3.5 , and other approaches in section 3.6 . Finally, we provide concluding remarks in section 4 . We hope that this work stimulates future discussions and the development of computationally efficient techniques for the analysis of statistically evolving flows in time. 2 Preliminaries In this paper, we aim to understand the evolution of perturbations (fluctuations) around some unsteady base flow (state). These perturbations may be analyzed using both linear and nonlinear approaches. If the perturbation is sufficiently small, linear methods are a valid approach. However, once the magnitudes of these perturbations become large, nonlinear effects need to be taken into account. In this section, we present the equations of motion for the perturbation about an unsteady base flow to highlight both the linear and nonlinear effects driving the perturbation dynamics. We will also describe why simple extensions of traditional linear methods are insufficient for understanding unsteady flows. 2.1 Governing equations Let us consider an n -dimensional state vector \mathbf{q}(t)\in\mathbb{R}^{n} of some dynamical system \dfrac{d\mathbf{q}(t)}{dt}=\mathbf{N}(\mathbf{q}(t))+\mathbf{f}(t), (1) where \mathbf{N}:\mathbb{R}^{n}\rightarrow\mathbb{R}^{n} is a nonlinear function and \mathbf{f}(t)\in\mathbb{R}^{n} is a time-varying forcing term. For the systems of interest in this paper, ( 1 ) comes from discretizing a partial differential equation for a field variable (or vector of field variables) q(\mathbf{x},t) onto a finite grid, or a finite set of basis functions, to give the time-dependent vector \mathbf{q}(t) . Although, we write the state vector as a real-valued vector for simplicity, there should be a similar formulation if the state vector is complex, for example, from a spatial Fourier transform. Next, we express the state variable and forcing term to be comprised of their base variables [\bar{\mathbf{q}}(t),\bar{\mathbf{f}}(t)] and perturbations [\mathbf{q}^{\prime}(t),\mathbf{f}^{\prime}(t)] , such that \mathbf{q}(t)=\bar{\mathbf{q}}(t)+\mathbf{q}^{\prime}(t) (2) and \mathbf{f}(t)=\bar{\mathbf{f}}(t)+\mathbf{f}^{\prime}(t). (3) By inserting these expressions into ( 1 ) and performing a Taylor series expansion of \mathbf{N} about \bar{\mathbf{q}}(t) , we find that \dfrac{d\mathbf{q}^{\prime}(t)}{dt}=\mathbf{L}[\bar{\mathbf{q}}(t)]\mathbf{q}^{\prime}(t)+\check{\mathbf{f}}(t), (4) where \mathbf{L}[\bar{\mathbf{q}}(t)]=\nabla\mathbf{N}|_{\bar{\mathbf{q}}(t)} and \check{\mathbf{f}}(t)=-\dfrac{d\bar{\mathbf{q}}(t)}{dt}+\mathbf{N}[\bar{\mathbf{q}}(t)]+\bar{\mathbf{f}}(t)+\mathbf{f}^{\prime}(t)+\mathcal{O}(||\mathbf{q}^{\prime}(t)||^{2}) (5) collects all of the remaining terms. This formulation frames the evolution of perturbations in a time-varying, forced linear system. While the above equation may suggest a simple application of linear dynamical systems theory around some chosen base flow to understand the perturbation dynamics, care is needed in adopting this formulation. Different treatments of the forcing term will lead to different interpretations of the results. These different treatments correspond to choices of the base states [\bar{\mathbf{q}}(t),\bar{\mathbf{f}}(t)] and the characteristics of the perturbations [\mathbf{q}^{\prime}(t),\mathbf{f}^{\prime}(t)] . If the base states are the solution to the Navier–Stokes equations (NSE), the first three terms of ( 5 ) sum to zero. In this case, \check{\mathbf{f}}(t) reduces to \mathbf{f}^{\prime}(t)+\mathcal{O}(||\mathbf{q}^{\prime}(t)||^{2})