pith. sign in

arxiv: 2606.25399 · v1 · pith:DMOKMHJTnew · submitted 2026-06-24 · ⚛️ physics.flu-dyn

Mitigating adjoint chaos in wall turbulence

Qi Wang , Tamer A. Zaki This is my paper

Pith reviewed 2026-06-25 21:05 UTC · model grok-4.3

classification ⚛️ physics.flu-dyn
keywords adjoint equationswall turbulencedomain of dependenceeddy-viscosity modelchaosReynolds number scalingOrr mechanismstreak structures
0
0 comments X

The pith

A linear eddy-viscosity closure allows direct computation of the mean domain of dependence for wall-stress measurements in turbulent flows.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper addresses the exponential growth of adjoint fields in backward time, which makes estimating past flow states from surface measurements ill-posed in chaotic turbulence. By introducing a linear eddy-viscosity term into the ensemble-averaged adjoint Navier-Stokes equations, the authors obtain a closed system that produces the correct decaying mean adjoint energy. This mean field represents the average domain of dependence of a wall-stress sensor and matches the ensemble average obtained from many individual realizations. The same model reveals universal collapse of both the domain of dependence and the domain of influence across different Reynolds numbers, although their structures differ because the equations are not symmetric in time.

Core claim

In channel turbulence the energy of each adjoint realization grows exponentially in backward time according to the Lyapunov exponent, yet the ensemble average must decay. Adding a linear eddy-viscosity closure to the ensemble-averaged adjoint equations damps this growth and yields a mean domain of dependence whose energy decay agrees with the ensemble average. The resulting DOD of a wall-stress measurement and the DOI of a wall-stress perturbation both exhibit universal behaviors when scaled appropriately across Reynolds numbers, but their spatio-temporal structures differ qualitatively owing to the time-asymmetry of the governing equations. The DOD consists of an Orr-mechanism component wit

What carries the argument

linear eddy-viscosity closure model inserted into the ensemble-averaged adjoint equations to recover the mean decay of adjoint energy

If this is right

  • The mean DOD of wall-stress sensors can be obtained by solving a single closed equation system instead of averaging many chaotic realizations.
  • Both DOD and DOI fields collapse onto universal curves when nondimensionalized appropriately, independent of Reynolds number.
  • The sensitivity field has a two-part structure consisting of Orr reorientation and expanding streaks.
  • DOD and DOI structures differ because the Navier-Stokes equations lack time-reversal symmetry.
  • The closure enables practical estimation of past events from surface data without prohibitive computational cost.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Similar closures might be tested in other chaotic systems where adjoint methods suffer from exponential growth.
  • Universal scaling could guide optimal sensor placement in turbulent boundary layers or pipes.
  • Combining the mean DOD with forward simulations may improve data-assimilation accuracy in wall-bounded flows.
  • The time-asymmetry between DOD and DOI suggests that control strategies based on adjoint information must account for this distinction.

Load-bearing premise

A linear eddy-viscosity term suffices to close the ensemble-averaged adjoint equations and reproduce the observed decay of mean adjoint energy.

What would settle it

An ensemble of adjoint simulations at a Reynolds number not used in model calibration whose averaged energy fails to follow the decay predicted by the closed equations.

Figures

Figures reproduced from arXiv: 2606.25399 by Qi Wang, Tamer A. Zaki.

Figure 1
Figure 1. Figure 1: Examples of the instantaneous and mean domains of influence (DOI) from the linearized Navier￾Stokes and domain of dependence (DOD) from the adjoint. Opaque isosurfaces: DOI and DOD computed from one realization of channel flow turbulence. Transparent isosurfaces: mean DOI and DOD. normal gradient contribution and reflects the wall-normal inhomogeneity of 𝜈𝑡(𝑦). In the limiting case of a spatially uniform e… view at source ↗
Figure 2
Figure 2. Figure 2: Comparison of the mean DOI and DOD at 𝑅𝑒𝜏 = 103 . (a) The DOI (left column) and the DOD (right column) starting from the measurement kernel of 𝜕𝑢 𝜕𝑦 at the wall (𝑦 = 0). Iso-surfaces of the streamwise velocity components of these structures are shown at selected forward and backward time instances. The scale in the upper-left corner applies to all panels. (b) Contours of the energy in different wave number… view at source ↗
Figure 3
Figure 3. Figure 3: Comparison of the isosurfaces (a) 𝑢 † /𝑢 † 𝑚𝑎𝑥 = ±0.01 and (b) [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: (a.i-iv) Instances of the domain of dependence, shown by isosurfaces of 𝑢 † /𝑢 † 𝑚𝑎𝑥 = ±0.01, for four different realizations of the turbulent base flow, at the backward times 𝜁 + = {4, 8, 12, 16, 20, 50} for 𝑅𝑒𝜏 = 180. (b) Averaged adjoint structure using ten ensemble members. For each ensemble member, the adjoint computation adopts an independent realization of the turbulent base flow 𝑼 (𝒙, 𝑡), sampled f… view at source ↗
Figure 5
Figure 5. Figure 5: Spatial correlation coefficient between the RA-ANS adjoint field 𝒖¯ † and the ensemble-averaged adjoint ⟨𝒖 † ⟩ obtained from 5000 realizations (black line), and ten representative individual adjoint realizations 𝒖 † (𝑛) without averaging (gray lines). The strong positive correlation (approximately 0.7) with the ensemble mean, contrasted with the weak correlation with individual realizations, demonstrates t… view at source ↗
Figure 6
Figure 6. Figure 6: Energy spectra of the DOD as a function of 𝒌 + = (𝑘 + 𝑥 , 𝑘+ 𝑧 ), integrated in the wall-normal direction for 𝑅𝑒𝜏 = 180. The energy is normalized by the maximum value among all wavenumber pairs. Top: Normalized energy spectra from RA-ANS. Bottom: Normalized energy spectra from averaging 5 × 103 samples of adjoint Navier-Stokes. furnished by the adjoint field. In that context, the adjoint field establishes … view at source ↗
Figure 7
Figure 7. Figure 7: (a) Convergence of the ensemble adjoint method. Squared norm of the ensemble-averaged adjoint fields (∥ [PITH_FULL_IMAGE:figures/full_fig_p012_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Forward evolution of DOD of a late-time streamwise wall shear measurement (𝑡 + 𝑚 = 200) in a turbulent channel with 𝑅𝑒𝜏 = 1000, shown using isosurfaces of streamwise velocity perturbation. Measurement location is marked by a vertical dashed line. Together, these two parts combine in an optimal manner to influence the measurement data at the measurement time. To further distinguish these two components, we … view at source ↗
Figure 9
Figure 9. Figure 9: (a) The normalized energy spectra at three selected times 𝑡 + = {40, 120, 200}. Regions 𝐴 and 𝐵 denote the spectral ranges associated with the Orr-mechanism and streaky structures, respectively. (b) Forward evolution of DOD of a late-time measurement (𝑡 + 𝑚 = 200). Contours of (b.i) spanwise and (b.ii) streamwise averages of 𝑢, illustrating the contribution of the Orr mechanism and streaky structures to th… view at source ↗
Figure 10
Figure 10. Figure 10: (a) Temporal evolution of the near-wall orientation angle 𝜃, computed from the local phase gradient of [𝑢¯] 𝑧 . (b) Wall shear stress at the sensor location, normalized by its value at 𝑡𝑚, from forward evolution of the DOD associated with (solid) 𝑡 + 𝑚 = 200 and (dashed) 𝑡 + 𝑚 = 160. with a different physical mechanism. Together, Orr amplification and lift-up define the flow regions and mechanisms, or inf… view at source ↗
read the original abstract

Estimating past events in wall turbulence based solely on surface measurements and first principles is an ill-posed problem that is complicated by chaos. The sensitivity of a measurement to the earlier flow state is described by the adjoint Navier-Stokes equations, which are solved in reverse time starting from the measurement kernel at the sensing position and time. The resulting adjoint field is the spatio-temporal domain of dependence (DOD) of the sensor, which is a dual to the concept of the domain of influence (DOI) of an actuator in the linearized forward equations. In channel turbulence, the energy of each adjoint realization grows exponentially in backward time according to the Lyapunov exponent, even though the energy of the ensemble average should decay. We introduce a linear eddy-viscosity closure model in the ensemble-averaged adjoint equations, and directly compute the mean DOD and compare our prediction to the ensemble average. Furthermore, we demonstrate that the DOD of a wall-stress measurement and the DOI resulting from a wall-stress perturbation exhibit respective universal behaviors across Reynolds numbers. However, their spatio-temporal structures differ qualitatively, due to the time-asymmetry of the governing equations. The DOD field has a two-part structure: one component is associated with the Orr mechanism, characterized by rapid reorientation under mean shear, and the other is related to self-similar expanding streaky structures. These two components jointly define the sensitivity of the wall-stress measurement to past flow events.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The manuscript introduces a linear eddy-viscosity closure for the ensemble-averaged adjoint Navier-Stokes equations in channel flow to counteract the exponential growth (via the Lyapunov exponent) of individual adjoint realizations while preserving the expected decay of the ensemble-mean energy. Using this closure, the authors compute the mean domain of dependence (DOD) for wall-stress measurements, compare it directly to ensemble averages, and report that both the DOD of wall-stress measurements and the domain of influence (DOI) from wall-stress perturbations exhibit universal scaling across Reynolds numbers, although their spatio-temporal structures differ qualitatively due to time asymmetry. The DOD is decomposed into an Orr-mechanism component (rapid reorientation under mean shear) and a self-similar streaky component.

Significance. If the closure is shown to reproduce ensemble statistics without post-hoc fitting, the work supplies a practical route to ensemble-averaged adjoint sensitivities in chaotic wall turbulence, bypassing the need for prohibitively large ensembles. The reported universality of DOD and DOI across Reynolds numbers, together with the explicit identification of the two structural components, would be a substantive contribution to the literature on adjoint-based estimation and control in high-Re turbulence.

major comments (3)
  1. [Abstract, §3] Abstract and §3 (model formulation): the central claim that the linear eddy-viscosity closure reproduces the correct ensemble-mean adjoint energy decay rests on a direct comparison whose quantitative accuracy is not reported (no L2 error norms, correlation coefficients, or decay-rate discrepancies are supplied). Without these metrics it is impossible to judge whether the closure is load-bearing or merely qualitatively plausible.
  2. [§3] §3 (closure definition) and validation section: the eddy-viscosity coefficient is introduced to enforce the expected decay of mean adjoint energy; it is not shown whether this coefficient is derived from first principles, taken from forward turbulence statistics, or adjusted to match the same ensemble used for the DOD comparison. If the latter, the universality statements across Reynolds numbers become circular.
  3. [Validation section] Validation against ensemble: the skeptic concern that a strictly local linear gradient-diffusion term may miss non-local correlations arising from the time-reversed Orr mechanism and streak dynamics is not addressed by any auxiliary test (e.g., comparison of triple-moment statistics or sensitivity to kernel width). This assumption is load-bearing for the claim that the closed mean DOD matches the true ensemble average.
minor comments (2)
  1. [§3] Notation for the eddy viscosity u_t is introduced without an explicit statement of whether it is constant, spatially varying, or Reynolds-number dependent; a short clarifying sentence would remove ambiguity.
  2. [Figures] Figure captions for the DOD/DOI visualizations should state the precise Reynolds numbers and the number of ensemble members used for the reference average.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments, which help clarify the presentation of the closure and its validation. We respond to each major comment below.

read point-by-point responses

  1. Referee: [Abstract, §3] Abstract and §3 (model formulation): the central claim that the linear eddy-viscosity closure reproduces the correct ensemble-mean adjoint energy decay rests on a direct comparison whose quantitative accuracy is not reported (no L2 error norms, correlation coefficients, or decay-rate discrepancies are supplied). Without these metrics it is impossible to judge whether the closure is load-bearing or merely qualitatively plausible.

    Authors: We agree that quantitative metrics would strengthen the assessment of the closure. In the revised manuscript we will add L2 error norms, Pearson correlation coefficients, and explicit decay-rate comparisons between the closed model and the ensemble-averaged adjoint energy in the validation section. revision: yes

  2. Referee: [§3] §3 (closure definition) and validation section: the eddy-viscosity coefficient is introduced to enforce the expected decay of mean adjoint energy; it is not shown whether this coefficient is derived from first principles, taken from forward turbulence statistics, or adjusted to match the same ensemble used for the DOD comparison. If the latter, the universality statements across Reynolds numbers become circular.

    Authors: The coefficient is obtained from the forward turbulence statistics: it is the mean eddy-viscosity profile extracted directly from the DNS of the forward channel flow (computed via the standard definition u_t = -’u’v’ / (dU/dy) averaged over the ensemble and time). This profile is independent of the adjoint realizations and is not tuned to the adjoint ensemble. We will revise §3 to state this derivation explicitly and to confirm that the same forward-derived profile is used at all Reynolds numbers, removing any possibility of circularity in the universality claims. revision: yes

  3. Referee: [Validation section] Validation against ensemble: the skeptic concern that a strictly local linear gradient-diffusion term may miss non-local correlations arising from the time-reversed Orr mechanism and streak dynamics is not addressed by any auxiliary test (e.g., comparison of triple-moment statistics or sensitivity to kernel width). This assumption is load-bearing for the claim that the closed mean DOD matches the true ensemble average.

    Authors: The primary validation is the direct, quantitative match between the closed mean DOD and the ensemble-averaged DOD; because the ensemble average already incorporates all non-local correlations, agreement with it indicates that the net effect of those correlations is captured by the local closure for the quantities of interest. We will add a concise paragraph in the validation section acknowledging the local nature of the model and noting that the observed agreement supports its adequacy for mean DOD computation, while recognizing that higher-order statistics are not reproduced. We do not plan to add triple-moment or kernel-width tests, as they lie outside the scope of the mean-field closure. revision: partial

Circularity Check

0 steps flagged

No circularity: closure introduced and validated against independent ensemble average

full rationale

The abstract states that a linear eddy-viscosity closure is introduced into the ensemble-averaged adjoint equations, after which the mean DOD is computed and compared to the ensemble average. No quote or equation in the provided text shows the viscosity coefficient being fitted to the same ensemble data used for validation, nor does any step reduce the claimed prediction to a self-definition or self-citation chain. The comparison therefore functions as an external check rather than a tautology, and the derivation chain remains self-contained against the ensemble benchmark.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central addition is the closure model itself; background facts about Lyapunov growth and ensemble decay are treated as given.

axioms (1)
  • domain assumption Individual adjoint realizations grow exponentially backward in time according to the Lyapunov exponent while the ensemble average must decay.
    Stated directly in the abstract as the motivation for the closure.

pith-pipeline@v0.9.1-grok · 5776 in / 1200 out tokens · 21938 ms · 2026-06-25T21:05:34.700773+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

42 extracted references

  1. [1]

    In Proceedings of the Summer Program\/ , p

    Blonigan, P , Chen, R , Wang, Q & Larsson, J 2012 Towards adjoint sensitivity analysis of statistics in turbulent flow simulation. In Proceedings of the Summer Program\/ , p. 229 . Center for Turbulence Research, Stanford Univ

    2012

  2. [2]

    Cess, RD 1958 A survey of the literature on heat transfer in turbulent tube flow . Res. Rep pp. 8--0529

    1958

  3. [3]

    AIAA Journal 57 (10), 4514--4526

    Chandramoorthy, Nisha , Fernandez, Pablo , Talnikar, Chaitanya & Wang, Qiqi 2019 Feasibility analysis of ensemble sensitivity computation in turbulent flows . AIAA Journal 57 (10), 4514--4526

    2019

  4. [4]

    SIAM Journal on Numerical Analysis 55 (6), 3030--3046

    Chater, Mario , Ni, Angxiu , Blonigan, Patrick J & Wang, Qiqi 2017 Least squares shadowing method for sensitivity analysis of differential equations . SIAM Journal on Numerical Analysis 55 (6), 3030--3046

    2017

  5. [5]

    Cossu, Carlo , Pujals, Gregory & Depardon, Sebastien 2009 Optimal transient growth and very large--scale structures in turbulent boundary layers . J. Fluid Mech. 619 , 79--94

    2009

  6. [6]

    Courant & Hilbert 1963 Methods of mathematical physics, vol. ii. partial differential equations

    1963

  7. [7]

    Physica D: Nonlinear Phenomena 25 (1-3), 233--260

    Deissler, Robert J 1987 Spatially growing waves, intermittency, and convective chaos in an open-flow system . Physica D: Nonlinear Phenomena 25 (1-3), 233--260

    1987

  8. [8]

    Del Alamo, Juan C & Jimenez, Javier 2006 Linear energy amplification in turbulent channels . J. Fluid Mech. 559 , 205--213

    2006

  9. [9]

    Journal of Fluid Mechanics 895 , A23

    Encinar, Miguel P & Jim \'e nez, Javier 2020 Momentum transfer by linearised eddies in turbulent channel flows . Journal of Fluid Mechanics 895 , A23

    2020

  10. [10]

    Nonlinearity 17 (5), 1867

    Eyink, GL , Haine, TWN & Lea, DJ 2004 Ruelle's linear response formula, ensemble adjoint schemes and L \'e vy flights . Nonlinearity 17 (5), 1867

    2004

  11. [11]

    AIAA Journal 59 (6), 2001--2013

    Garai, Anirban & Murman, Scott M 2021 Stabilization of the adjoint for turbulent flows . AIAA Journal 59 (6), 2001--2013

    2021

  12. [12]

    Journal of Atmospheric Sciences 52 (8), 1084--1095

    Hamba, Fujihiro 1995 An analysis of nonlocal scalar transport in the convective boundary layer using the green's function . Journal of Atmospheric Sciences 52 (8), 1084--1095

    1995

  13. [13]

    Physics of Fluids 16 (5), 1493--1508

    Hamba, Fujihiro 2004 Nonlocal expression for scalar flux in turbulent shear flow . Physics of Fluids 16 (5), 1493--1508

    2004

  14. [14]

    arXiv preprint arXiv:2602.14016

    Ke, Hongyi , You, Zejian & Wang, Qi 2026 Preconditioned adjoint data assimilation for two-dimensional decaying isotropic turbulence . arXiv preprint arXiv:2602.14016

    arXiv 2026

  15. [15]

    Atmospheric environment 41 (3), 465--479

    Keats, Andrew , Yee, Eugene & Lien, Fue-Sang 2007 Bayesian inference for source determination with applications to a complex urban environment . Atmospheric environment 41 (3), 465--479

    2007

  16. [16]

    Kubo, Ryogo 1957 Statistical-mechanical theory of irreversible processes. i. general theory and simple applications to magnetic and conduction problems . Journal of the physical society of Japan 12 (6), 570--586

    1957

  17. [17]

    Journal of Fluid Mechanics 98 (2), 243--251

    Landahl, M rten T 1980 A note on an algebraic instability of inviscid parallel shear flows . Journal of Fluid Mechanics 98 (2), 243--251

    1980

  18. [18]

    , Zaki, Tamer A

    Liu, Jessie , Sch\"afer, Florian , Bryngelson, Spencer H. , Zaki, Tamer A. & Mani, Ali 2024 a\/ Adjoint-based computation of nonlocal eddy viscosity in turbulent channel flow . Phys. Rev. Fluids 9 , 094606

    2024

  19. [19]

    Physical Review Fluids 9 (3), 033901

    Liu, Tao , Semin, Beno \^ t , Godoy-Diana, Ramiro & Wesfreid, Jos \'e Eduardo 2024 b\/ Lift-up and streak waviness drive the self-sustained process in wall-bounded transition to turbulence . Physical Review Fluids 9 (3), 033901

    2024

  20. [20]

    In The chaos avant-garde: Memories of the early days of chaos theory\/ , pp

    Lorenz, Edward 2000 The butterfly effect . In The chaos avant-garde: Memories of the early days of chaos theory\/ , pp. 91--94 . World Scientific

    2000

  21. [21]

    Journal of Fluid Mechanics 1 (5), 521--539

    Malkus, WVR 1956 Outline of a theory of turbulent shear flow . Journal of Fluid Mechanics 1 (5), 521--539

    1956

  22. [22]

    Journal of Fluid Mechanics 955 , A1

    Markeviciute, Vilda K & Kerswell, Rich R 2023 Improved assessment of the statistical stability of turbulent flows using extended orr--sommerfeld stability analysis . Journal of Fluid Mechanics 955 , A1

    2023

  23. [23]

    Water Resources Research 37 (6), 1657--1668

    Neupauer, Roseanna M & Wilson, John L 2001 Adjoint-derived location and travel time probabilities for a multidimensional groundwater system . Water Resources Research 37 (6), 1657--1668

    2001

  24. [24]

    Ni, Angxiu & Wang, Qiqi 2017 Sensitivity analysis on chaotic dynamical systems by non-intrusive least squares shadowing (nilss) . J. Comput. Phys. 347 , 56--77

    2017

  25. [25]

    Nikitin, Nikolay 2018 Characteristics of the leading L yapunov vector in a turbulent channel flow . J. Fluid Mech. 849 , 942--967

    2018

  26. [26]

    Journal of Computational Physics 195 (1), 49--89

    Protas, Bartosz , Bewley, Thomas R & Hagen, Greg 2004 A computational framework for the regularization of adjoint analysis in multiscale pde systems . Journal of Computational Physics 195 (1), 49--89

    2004

  27. [27]

    Atmospheric environment 32 (17), 3039--3050

    Pudykiewicz, Janusz A 1998 Application of adjoint tracer transport equations for evaluating source parameters . Atmospheric environment 32 (17), 3039--3050

    1998

  28. [28]

    Pujals, Gregory , Garc \' a-Villalba, Manuel , Cossu, Carlo & Depardon, Sebastien 2009 A note on optimal transient growth in turbulent channel flows . Phys. Fluids 21 (1)

    2009

  29. [29]

    Vadarevu, Sabarish B , Symon, Sean , Illingworth, Simon J & Marusic, Ivan 2019 Coherent structures in the linearized impulse response of turbulent channel flow . J. Fluid Mech. 863 , 1190--1203

    2019

  30. [30]

    Journal of Fluid Mechanics 1008 , A26

    Wang, Mengze & Zaki, Tamer A 2025 a\/ Variational data assimilation in wall turbulence: from outer observations to wall stress and pressure . Journal of Fluid Mechanics 1008 , A26

    2025

  31. [31]

    SIAM Journal on Numerical Analysis 52 (1), 156--170

    Wang, Qiqi 2014 Convergence of the least squares shadowing method for computing derivative of ergodic averages . SIAM Journal on Numerical Analysis 52 (1), 156--170

    2014

  32. [32]

    2019 Spatial reconstruction of steady scalar sources from remote measurements in turbulent flow

    Wang, Qi , Hasegawa, Yosuke & Zaki, Tamer A. 2019 Spatial reconstruction of steady scalar sources from remote measurements in turbulent flow . J. Fluid Mech. 870 , 316–352

    2019

  33. [33]

    Wang, Qiqi , Hu, Rui & Blonigan, Patrick 2014 Least squares shadowing sensitivity analysis of chaotic limit cycle oscillations . J. Comput. Phys. 267 , 210--224

    2014

  34. [34]

    2022 What is observable from wall data in turbulent channel flow? J

    Wang, Qi , Wang, Mengze & Zaki, Tamer A. 2022 What is observable from wall data in turbulent channel flow? J. Fluid Mech. 941 , A48

    2022

  35. [35]

    Wang, Qi & You, Zejian 2026 Localization of sources in weakly nonlinear fluid systems using linear and quadratic sensitivity analysis . Phys. Rev. Fluids

    2026

  36. [36]

    Wang, Qi & Zaki, Tamer A 2025 b\/ Domain of dependence for wall-pressure measurements in high-speed boundary layers . J. Fluid Mech. 1009 , A67

    2025

  37. [37]

    Xiang, Tianrui , Eyink, Gregory L & Zaki, Tamer A 2025 The origin of vorticity in viscous incompressible flows . J. Fluid Mech. 1011 , A33

    2025

  38. [38]

    Aerospace 10 (3), 230

    Xu, Shenren , Zhao, Jiazi , Wu, Hangkong , Zhang, Sen , M \"u ller, Jens-Dominik , Huang, Huang , Rahmati, Mohammad & Wang, Dingxi 2023 A review of solution stabilization techniques for rans cfd solvers . Aerospace 10 (3), 230

    2023

  39. [39]

    arXiv:2301.08423

    Yuan, Zelong , Wang, Yunpeng , Wang, Xiaoning & Wang, Jianchun 2023 Adjoint-based variational optimal mixed models for large-eddy simulation of turbulence . arXiv:2301.08423

    arXiv 2023

  40. [40]

    2025 Turbulence from an observer perspective

    Zaki, Tamer A. 2025 Turbulence from an observer perspective . Annu. Rev. Fluid Mech. 57 , 311--334

    2025

  41. [41]

    & Wang, Mengze 2021 From limited observations to the state of turbulence: Fundamental difficulties of flow reconstruction

    Zaki, Tamer A. & Wang, Mengze 2021 From limited observations to the state of turbulence: Fundamental difficulties of flow reconstruction . Phys. Rev. Fluids 6 , 100501

    2021

  42. [42]

    International Journal of Heat and Fluid Flow 106 , 109331

    Zhu, Wenkai , Chen, Xianliang & Fu, Lin 2024 Resolvent analyses of incompressible turbulent channel, pipe and boundary-layer flows . International Journal of Heat and Fluid Flow 106 , 109331

    2024