Uncertainty Quantification in Resolvent Analysis of Experimental Wall-Bounded Turbulent Flows
Salvador R. Gomez, Tomek M. Jaroslawski · 2025
Read Uncertainty Quantification in Resolvent Analysis of Experimental Wall-Bounded Turbulent Flows by Salvador R. Gomez, Tomek M. Jaroslawski (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-2511-00258.txt.
Introduction
Salvador Rey Gomez and Tomek Jaroslawski ask how much trust to place in a popular tool for studying turbulence near walls. The tool, resolvent analysis, breaks the flow equations into modes that show which patterns grow strongest. It often runs on measured average flows, but sensors cannot resolve the thin layer right at the wall, leaving the near-wall speed uncertain.
The authors show that the effect of that uncertainty can be measured at little extra computing cost. Tests use artificial disturbances to average profiles and real camera-based flow measurements with different near-wall fits. In one boundary layer case, poor near-wall data flipped the conclusion. Careful uncertainty checks, they argue, should accompany the method.
Catalog record
| CREATOR | Salvador R. Gomez, Tomek M. Jaroslawski |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 60 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-00258
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Uncertainty Quantification in Resolvent Analysis of Experimental Wall-Bounded Turbulent Flows Salvador R. Gomez, Tomek M. Jaroslawski Abstract Experimental mean flows are commonly used to study wall-bounded turbulence. However, these measurements are often unable to resolve the near-wall region and thus introduce ambiguity in the velocity closest to the wall. This poses a source of uncertainty in equation-based approaches that rely on these mean flow measurements such as resolvent analysis. Resolvent analysis provides a scale-dependent decomposition of the linearized Navier Stokes equations that identifies optimal gains, response modes and forcing modes that has been used to great effect in turbulent wall-bounded flows. Its potential in the development of predictive tools for a variety of wall-bounded flows is high but the limitations of the input data must be addressed. Here, we quantify the sensitivity of resolvent analysis to common sources of experimental uncertainty and show that this sensitivity can be quantified with minimal additional computational cost. This approach is applied to both local and biglobal resolvent analysis by using artificial disturbances to mean profiles in the former and particle image velocimetry measurements with differing near-wall fits in the latter. We also highlight an example where poor near-wall resolution can lead to erroneous conclusions compared to the full-resolution data in an adverse pressure gradient turbulent boundary layer. keywords Turbulent boundary layers, uncertainty quantification, resolvent analysis, mean flow reconstruction † † equal-contributors: These authors contributed equally to this work. 1 Introduction Wall-bounded turbulent flows exhibit strong inhomogeneity and a wide range of interacting scales, making them particularly difficult to measure and characterize. Resolvent analysis provides a linear framework to study the amplification of disturbances by the mean flow, but its accuracy depends critically on the fidelity of that mean profile. In practice, experimental measurements often lack sufficient resolution close to the wall. Similarly, wall-modeled large-eddy simulations (WMLES) rely on models to approximate near-wall dynamics, introducing significant uncertainty in that region. As a result, mean flow profiles may contain inaccuracies—especially near the wall—raising an important question: how sensitive is resolvent analysis to such errors? This study addresses that question through a systematic sensitivity analysis. The mean flow field of a zero-pressure-gradient (ZPG) turbulent boundary layer (TBL) on a hydraulically smooth surface exhibits self-similar behavior across different wall-normal regions, each with its own characteristic length and velocity scales. These scales give rise to the classical law of the wall and the law of the wake, which are valid in the near-wall and outer regions of the flow, respectively, with the log layer classically characterized as an overlap layer where both scaling laws apply ( Coles 1956 ) . These scaling laws can be used to model the near-wall flow field from outer region measurements ( Van Driest 1956 ) . The near-wall scales are determined by viscosity and wall shear via the friction velocity, u_{\tau}=\sqrt{\tau_{w}/\rho} , where \tau_{w} is the wall shear stress and \rho the density, and the viscous length scale, \ell_{\nu}=\nu/u_{\tau} , where \nu is the kinematic viscosity. Scales normalized with u_{\tau} and \ell_{\nu} are here denoted with a + superscript. The outer region can be described by the outer length-scale, \delta . A measure of scale separation in a ZPG TBL is the friction Reynolds number, Re_{\tau}=\delta/\ell_{\nu} . The importance of the mean flow field is that it encodes valuable information about the dynamics of the turbulent fluctuations through turbulent production and linear amplification. The linearized Navier Stokes equations (LNSE) create a linear operator, \mathbf{L} , that can provide many useful predictions for turbulent fluctuations. Linear stability analysis is used to determine the growth of initial disturbances ( Schmid and Henningson 2002 ) , though these results are sensitive to perturbations in \mathbf{L} when \mathbf{L} is non-normal ( Trefethen et al. 1993 ; Reddy et al. 1993 ) , i.e. \mathbf{L}\mathbf{L}^{\dagger}\neq\mathbf{L}^{\dagger}\mathbf{L} . Transient growth analysis relies on the non-normality of \mathbf{L} ( Trefethen et al. 1993 ) and has identified length-scales and flow structures similar to those observed in turbulent flows from the optimal linear amplification ( Del Alamo and Jimenez 2006 ; Cossu et al. 2009 ; Hwang and Cossu 2010 ) . While these approaches can augment \mathbf{L} with an eddy viscosity to model the nonlinearities ( Reynolds and Hussain 1972 ) , they ultimately ignore the presence of the nonlinear terms in the fluctuating equations. Resolvent analysis remedies this by modeling the nonlinearities as a forcing input and creating the resolvent operator as a linear transfer function between the fluctuations and said forcing ( McKeon and Sharma 2010 ) . The singular value decomposition (SVD) of the resolvent operator identifies an orthonormal forcing basis and an orthonormal response basis each ranked by their singular values, the linear gain of the transfer function. With deep ties to the psuedospectrum of \mathbf{L} , resolvent analysis takes advantage of non-normal mechanisms present to find the most amplified motions ( Symon et al. 2018 ) . Resolvent analysis has identified similarities in the response modes to those observed in the near-wall turbulent flows ( Moarref et al. 2013 ; Abreu et al. 2020 ) , increased amplification of the large-scales similar to the observed increased energization in APG TBLs ( Gomez and McKeon 2025 ) , provides a useful basis for flow reconstruction ( Moarref et al. 2014 ) , and can agree with data-driven modal decompositions ( Towne et al. 2018 ) . Extensions to high Reynolds number wall-bounded flows, especially those with non-canonical effects like pressure gradients or roughness, are limited by the availability of well-resolved data and often rely on experimental measurements. Experimental mean flow fields have been used in resolvent analysis to great success ( Beneddine et al. 2017 ; He et al. 2019 ; Lesshafft et al. 2019 ; Symon et al. 2019 ; Preskett et al. 2024 ) , and we highlight the observations of Chavarin et al. 2020 that their resolvent gains were sensitive to the mean flow estimation. However, experimental measurements in high Reynolds number flows often have a great deal of uncertainty near the wall where the law-of-the-wall is often used to supplement the lack of measurements in this region. The sensitivity of linear stability analysis to experimental measurements and fitting of the mean profile has been studied by Boutilier and Yarusevych 2013 and can be rigorously quantified with psuedospectral methods ( Reddy et al. 1993 ; Trefethen 1999 ) . The sensitivity of resolvent gains to parameters has been used to great effect to study parameterization of the mean flow and the LNSE ( de Pando et al. 2014 ; Fosas de Pando and Schmid 2017 ; Skene and Schmid 2019 ) and to search large amplification regions in wavenumber space ( Gomez et al. 2022 ) . Here, we use this approach to quantify the uncertainty in the resolvent analysis gains to experimental mean flow error and mean flow reconstruction. Furthermore, we extend the approaches in de Pando et al. 2014 ; Fosas de Pando and Schmid 2017 ; Skene and Schmid 2019 ; Gomez et al. 2022 to introduce a formulation that predicts the sensitivity in the resolvent modes as well. In particular, we focus on uncertainty in the near-wall measurements since the near-wall mean shear is a source of amplification responsible