Pith. sign in

REVIEW 2 major objections 5 minor 50 references

Self-Replication of Turbulent Puffs: On the edge between chaotic saddles

T0 review · 2 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Puff splitting in pipe flow is mediated by a hidden edge state

desk verdict First direct DNS identification of a split edge state in pipe flow, with a credible mechanism and a fixable over-reliance on arbitrary classification thresholds. read the letter →

arxiv 2505.05075 v2 pith:WSZW2C6X submitted 2025-05-08 physics.flu-dyn nlin.CD

Anton Svirsky , Tobias Grafke , Anna Frishman This is my paper · ORCID
classification physics.flu-dynnlin.CD
keywords pipeflowturbulentpuffsself-replicationedgestatechaoticsaddlesslug-gap-splitmechanismsubcriticaltransitiondirectnumericalsimulation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper asks what exactly happens when a turbulent puff in pipe flow splits into two puffs, the process that sustains turbulence near the transition point. Using direct numerical simulations at $Re = 2200$, the authors identify a specific phase-space object — the split edge state — sitting on the boundary between the one-puff and two-puff states. They show that naturally occurring splitting events all pass near this state, and that the path follows a two-step 'slug-gap-split' mechanism: the puff first expands into a short slug with a uniform turbulent core, then a laminar gap nucleates and widens, completing the split. The result matters because it turns a blurry rare-event statistic into a concrete dynamical transition, and it suggests the same mechanism may govern turbulence spreading in other wall-bounded flows.

What carries the argument

The key object is the split edge state: a chaotic saddle on the phase-space boundary between the one-puff and two-puff states, found by iteratively bisecting two bounding states that evolve to one puff and two puffs respectively. The bisection uses a coarse-grained classification function $\Phi$ (Eq. 5) based on total turbulent length, number of turbulent patches, and laminar gap width, with per-Reynolds-number fitted means and standard deviations. The edge state's relevance is established by projecting instantaneous velocity fields onto the first two principal components of a dataset containing the split edge, one-puff, and two-puff states; in this $(p_1, p_2)$ plane, splitting trajectories form a clear tube passing through the edge state. The physical interpretation is delivered by correlating $p_2$ with streamwise turbulent kinetic energy (correlation 0.96) and $p_2 - p_1$ with laminar gap width (correlation 0.91).

What would settle it

Run the bisection algorithm at $Re = 2200$ with an alternative classification that treats three-patch states (currently lumped into $\Phi = -1$) as a separate transitional set, or with a different gap-width threshold $w_{th}$; if the resulting edge state no longer lies on the path of naturally occurring splitting trajectories, the claim that the split edge state mediates self-replication is refuted. Alternatively, a direct numerical simulation that captures a natural splitting event whose trajectory in the $(p_1, p_2)$ plane avoids the split edge state would falsify the claim.

Watch

Extended reading notes

Core claim

The central claim is that self-replication of turbulent puffs is not a featureless random escape from a chaotic saddle but a deterministic transition between two chaotic saddles, mediated by an edge state embedded in their phase-space boundary. The authors locate this 'split edge state' by a bisection edge-tracking algorithm, find that it resembles a short slug (an elongated puff with a uniform turbulent core), and demonstrate that all nine observed splitting trajectories pass near it in a PCA projection and in the full $L^2$ distance. The transition path decomposes into two straight segments in the reduced phase space, corresponding first to growth of streamwise turbulent kinetic energy and then to widening of a laminar gap. This directly confirms the previously proposed slug-gap-split mechanism, with the edge state as the tipping point between successful splitting and retraction.

Load-bearing premise

The entire edge-state picture rests on the classification function $\Phi$, whose thresholds (means, standard deviations, and gap-width cutoffs fitted separately at each Reynolds number) decide what counts as a one-puff or two-puff state; if that classifier mislabels transitional states, the bisection could converge to a boundary that is an artifact of the definition rather than the true one-puff/two-puff separator.

Editorial extensions

If this is right

  • The one-puff-to-two-puff transition in pipe flow has a well-defined tipping point, so split waiting times can be understood as escapes from a chaotic saddle through a specific edge state.
  • The slug-gap-split mechanism, confirmed here for pipe flow, becomes a candidate generic route for turbulence proliferation in other subcritical wall-bounded flows such as plane Couette and Taylor-Couette.
  • Because TKE$_z$ reaching the edge-state value is necessary but not sufficient for a split (only about 50% of such cases split), future observables targeting gap nucleation could predict individual splitting events.
  • The methodology of edge tracking between two chaotic saddles may transfer to other spatiotemporal chaotic systems, including climate and active-matter models, where transitions between coexisting chaotic saddles are relevant.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the split edge state is a genuine saddle on the boundary, rare-event algorithms could be tuned to sample splitting trajectories by targeting the edge state's stable manifold, dramatically shortening the waiting times needed in DNS.
  • The Reynolds-number dependence of the edge state (slug-like at 2200, semi-periodic at 2050 and 2100) hints that the transition may change character between 2100 and 2200; a systematic continuation of the edge state in $Re$ would reveal whether the slug-gap-split mechanism is the only route or one of several.
  • The strong correlations of $p_2$ and $p_2-p_1$ with TKE$_z$ and gap width suggest that a reduced two-variable model might capture split dynamics, which could guide control strategies to delay or promote turbulence spreading in pipes.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 5 minor

Summary. The manuscript studies the self-replication of turbulent puffs in pipe flow at Reynolds numbers near the onset of sustained turbulence. Using direct numerical simulations at Re = 2200, the authors apply an edge-tracking bisection algorithm in the full phase space to identify a 'split edge state' that lies on the boundary between initial conditions that evolve to a one-puff state and those that evolve to a two-puff state. They then analyze N = 9 naturally occurring splitting events, project them onto the first two principal components of a dataset built from one-puff, two-puff, and split-edge snapshots, and report that all splitting trajectories pass near the split edge state in both this PCA projection and a symmetry-reduced L2-distance check. The transition is interpreted as a two-step 'slug-gap-split' mechanism: the puff first expands into a slug-like structure with a homogeneous turbulent core (the split edge state), and then a laminar gap nucleates and widens to complete the split. The authors also report a qualitatively different, semi-periodic edge state at Re = 2050 and 2100, whose relation to the split mechanism is left open. The central claim is that the split edge state is the dynamical tipping point of self-replication, providing direct confirmation of a mechanism previously proposed in the Barkley model.

Significance. If the central claim holds, this would be a notable result: the first demonstration in a deterministic, high-dimensional fluid system that a transition between two chaotic saddles is mediated by an edge state, and a direct confirmation of the slug-gap-split mechanism for puff self-replication. The paper has real strengths: the DNS evidence is independent of the Barkley model, the PCA separation is cross-validated by an L2-distance check that does not rely on the low-dimensional projection, the robustness of the bisection algorithm is tested over the tolerance h1 and the stabilization time tau, and the simulations use the openly available openpipeflow code. The N = 9 splitting events are all consistent with the claimed pathway, and the correlation analysis linking PCA coordinates to turbulent kinetic energy and gap width is compelling. The main caveat is that the edge state is defined with respect to a coarse-grained classification function containing several tunable thresholds; the sensitivity of the central claim to those thresholds is not addressed, and the proximity statistics for splitting trajectories lack a null-model comparison.

major comments (2)
  1. [End matter, 'State classification' and Eq. (5)] The split edge state is obtained by bisection between sets S1 and S2 whose definitions depend on the classification function Phi of Eq. (5), which contains several tunable constants: the threshold wth = 20D for the laminar gap width, the 2*sigma_i statistical tolerances on the turbulent length, the turbulence intensity threshold qth, and the stabilization time tau. The reported robustness checks vary h1 and tau, but not wth or the 2*sigma_i tolerances. Because the two-puff set is defined only for wgap > wth, the phase-space boundary being bisected includes a level set set by an arbitrary constant; changing wth changes the target set S2 and can therefore shift the boundary and the edge state found by the algorithm. The claim that the split edge state is the dynamically selected tipping point is load-bearing for the paper, so the authors should either vary wth (e.g., 10D and 30D) and the statistical tolerances and show that the edge state, its core length, and its unstable direction are unchanged, or give a dynamical argument for why these classification constants do not affect the boundary.
  2. [Relevance for splitting events, Fig. 3, and the L2-distance check] The statement that all N = 9 splitting trajectories pass near the split edge state needs a control against the classification threshold itself. Every natural splitting event must at some time cross from a state with n = 1 to a state with n = 2 and wgap > wth, so proximity to the classifier-defined boundary is expected by construction; what needs to be shown is that proximity to the particular state found by bisection is significantly stronger than proximity to generic states on that boundary or to typical non-splitting fluctuations within the one-puff set. The paper should compare the minimal distances des(t) for splitting trajectories with the distribution of the same quantity for non-splitting trajectories that approach the one-puff saddle (or for randomized one-puff initial conditions). Without such a null model, the N = 9 events support the consistency of the pathway but not yet the claim that the split edge state is a unique or unusually relevant gateway for splitting.
minor comments (5)
  1. [Fig. 2 caption and main text] The statement that the upstream and downstream fronts of the split edge are 'exactly identical' to those of the puff state is too strong for time-averaged profiles with visible standard deviations; 'indistinguishable within statistical uncertainty' would be more accurate.
  2. [End matter, 'Definition of the phase space boundary'] There is a typo in the sentence 'a precise dentition requires considering a finite time horizon'; 'dentition' should be 'definition'.
  3. [Principal Component Analysis, end matter] The PCA validation would be more informative if the authors reported the fraction of total variance captured by the first two principal components, since the reduced phase space is central to Fig. 3(a).
  4. [Main text, 'Split edge state'] The phrase 'approximatelyλ = 0.48± 0.04' has a missing space; it should read 'approximately λ = 0.48 ± 0.04'.
  5. [State classification, end matter] The manuscript states that the chosen qth is consistent with common values such as [47], but it does not discuss the sensitivity of the classification to qth; a brief comment on whether the edge state persists for, say, qth in a factor-of-two range would address a natural concern.

Circularity Check

1 steps flagged · score 2.0 of 10

Minor self-citation for the splitting mechanism; no load-bearing circularity in the DNS edge-state derivation.

  1. other [Main text, 'Puff-splitting mechanism' section, reference [16]]
    "The two-step process described above aligns with the previously proposed “slug-gap-split” mechanism [16]"

    Reference [16] is authored by two of the present authors (Frishman and Grafke), so the label 'previously proposed' is a self-citation. The paper uses this citation to name the mechanism, but the DNS edge state and splitting trajectories are computed independently in this work, so the self-citation is not load-bearing for the empirical findings. The circularity is limited to the interpretive framing, not the data derivation.

full rationale

The paper's central derivation—identifying the split edge state and showing natural splits pass near it—is based on DNS bisection in the full phase space, not on any fitted prediction. The classification function Φ (Eq. 5) is used only to label one-puff and two-puff sets; the edge state is found as an attracting set on the dynamically defined boundary, and its structure (homogeneous ~8D core) is not encoded in Φ. The proximity of splitting trajectories is verified by direct L2 distances in the full phase space, independent of the PCA projection, and the PCA 'tube' is a data-driven reduction, not a fitted prediction. The only self-citation is the 'slug-gap-split' mechanism from reference [16], by two of the same authors, which provides the interpretive label but is confirmed by independent DNS evidence. The arbitrary threshold wth=20 in Φ is a robustness concern rather than a circularity, because the edge state is not defined as a level set of wgap; it is a dynamical saddle on the boundary. Overall, the paper is self-contained against external DNS benchmarks, with only a minor non-load-bearing self-citation, so no significant circularity is present.

Assumptions & free parameters 6 free parameters · 5 assumptions · 1 invented entities

The central claim rests on the dynamical-systems interpretation of puffs as chaotic saddles, a coarse-grained classification function with per-Reynolds-number fitted thresholds, and the assumption that the bisection algorithm converges to the relevant boundary. The split edge state is a new dynamical object but is directly observed in simulation rather than postulated as a hidden mechanism.

free parameters (6)
  • turbulent length classification statistics (l1_bar, sigma1, l2_bar, sigma2) per Reynolds number = Re=2200: (10.75, 2.3, 21.5, 3.1); Re=2100: (9.4, 2.1, 18.8, 2.5); Re=2050: (9.05, 2.0, 18.1, 2.4)
    Measured from 1000D/Ubar DNS runs for one- and two-puff states; used in Eq. (5) to define the sets S1 and S2 for edge tracking.
  • qth (turbulence classification threshold) = 5e-4
    Threshold on q(z,t) to classify a cross-section as turbulent; chosen to match commonly used values in the literature.
  • wth (minimal gap width for two-puff classification) = 20D
    Threshold on laminar gap width used in Eq. (5); set to 20D for all Reynolds numbers considered.
  • tau (stabilization time in classification) = 50D/Ubar
    Duration for which Phi must be constant to classify a state; robustness checked over 25 to 100.
  • PCA representative two-puff gap = 25 +/- 5D
    Only two-puff states with this gap were used in the PCA dataset; different spacings correspond to different velocity fields even after symmetry reduction.
  • bisection tolerance h1 = 2e-6
    Distance threshold for convergence of edge tracking; robustness checked over 1e-7 to 1e-3.
assumptions (5)
  • standard math Navier-Stokes with no-slip boundary conditions describes the pipe flow under study
    The equations of motion are stated in Eq. (3) and integrated with openpipeflow; this is the standard continuum model for incompressible flow.
  • domain assumption Puffs correspond to chaotic saddles and the laminar state to a stable fixed point, with an edge of chaos boundary
    This is the standard dynamical-systems picture for subcritical shear flows, supported by many cited works; the paper extends it to the boundary between one- and two-puff saddles.
  • ad hoc to paper The bisection algorithm converges to an attracting set on the phase-space boundary
    The existence of a single attracting edge state on the boundary is assumed and then tested; the algorithm is initialized with two nearby states that go to S1 and S2 and iteratively refined.
  • domain assumption The Barkley model is a faithful minimal model for puff dynamics, so agreement with it indicates universality
    Used in the conclusion and in comparisons of the edge state structure; the model is phenomenological but has been validated in prior work.
  • ad hoc to paper The first two PCA components provide a faithful projection of the transition dynamics
    The reduced phase space uses only two components; the authors partially validate this with an L2-distance check, but the projection itself is built from data that includes the split edge state.
invented entities (1)
  • Split edge state independent evidence
    purpose: An invariant set on the boundary between the one-puff and two-puff chaotic saddles, proposed to be the tipping point for puff splitting.
    Directly observed in DNS at Re=2200 through edge tracking; its structure and role are reproducible with the described algorithm, though it is introduced as a new dynamical object.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Self-Replication of Turbulent Puffs: On the edge between chaotic saddles." pith.science (2026). https://pith.science/paper/WSZW2C6X

@misc{pith2026250505075,
  author       = {Pith},
  title        = {Pith review of: Self-Replication of Turbulent Puffs: On the edge between chaotic saddles},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WSZW2C6X}},
  note         = {Machine review of arXiv:2505.05075}
}
read the original abstract

Pipe flow is a canonical example where turbulence first appears intermittently in space and time, taking the form of localized structures termed puffs. Turbulence spreads via puff self-replication, which must out-compete puff decays to sustain it. Here we study the self-replication process, a transition from one to two puffs, using direct numerical simulations. We identify an edge state on the phase space boundary between the two states, demonstrate that it mediates the transition, and show that self-replication follows a previously proposed mechanism, with the edge state as its tipping point.

Figures

Figures reproduced from arXiv: 2505.05075 by the authors.

Figure 1
Figure 1. Illustration of one- and two-puff configurations as [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Comparison of flow profiles for the split edge and turbulent puff at Re = 2200. (a) Time-averaged turbulence level [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. (a) Splitting trajectories in a reduced phase-space, projected onto the two largest PCA components, [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 49 canonical work pages

  1. [1]

    Self-Replication of Turbulent Puffs: On the edge between chaotic saddles

    with viscosity ν, flowing in a pipe with circular cross-section of diameter D and length L with periodic boundary conditions in the stream-wise direction. The Reynolds number is defined as Re = UD/ν , where U is the cross-section mean flow velocity which is kept con- stant, enforcing constant mass flux. The equations are non-dimensionalized with velocity ...

    work page Pith review arXiv 2025

  2. [2]

    An experimental investigation of the circumstances which determine whether the motion of water shall be direct or sinuous, and of the law of resis- tance in parallel channels,

    Osborne Reynolds, “An experimental investigation of the circumstances which determine whether the motion of water shall be direct or sinuous, and of the law of resis- tance in parallel channels,” Philosophical Transactions of the Royal Society of London 17, 935–982 (1883)

    work page

  3. [3]

    77 (Springer Netherlands, 2005)

    Tom Mullin and Rich Kerswell, eds., IUTAM Symposium on Laminar-Turbulent Transition and Finite Amplitude Solutions, Fluid Mechanics and its Applications, Vol. 77 (Springer Netherlands, 2005)

    work page 2005

  4. [4]

    P. G. Drazin, Introduction to Hydrodynamic Stability , Cambridge Texts in Applied Mathematics (Cambridge University Press, 2002)

    work page 2002

  5. [5]

    The onset of turbulence in pipe flow,

    Kerstin Avila, David Moxey, Alberto De Lozar, Marc Avila, Dwight Barkley, and Bj¨ orn Hof, “The onset of turbulence in pipe flow,” Science 333, 192–196 (2011)

    work page 2011

  6. [6]

    On transition in a pipe. part 1. the origin of puffs and slugs and the flow in a turbulent slug,

    I. J. Wygnanski and F. H. Champagne, “On transition in a pipe. part 1. the origin of puffs and slugs and the flow in a turbulent slug,” Journal of Fluid Mechanics 59, 281–335 (1973)

    work page 1973

  7. [7]

    On tran- sition in a pipe. part 2. the equilibrium puff,

    I. Wygnanski, M. Sokolov, and D. Friedman, “On tran- sition in a pipe. part 2. the equilibrium puff,” Journal of Fluid Mechanics 69, 283–304 (1975)

    work page 1975

  8. [8]

    Laminar-to-turbulent transition of pipe flows through puffs and slugs,

    Mina Nishi, B¨ ulent ¨Unsal, Franz Durst, and Gautam Biswas, “Laminar-to-turbulent transition of pipe flows through puffs and slugs,” Journal of Fluid Mechanics 614, 425–446 (2008)

    work page 2008

Show all 50 references
  1. [9]

    The flow structure of a puff,

    Casimir W. H. van Doorne and Jerry Westerweel, “The flow structure of a puff,” Philosophical Transactions: Mathematical, Physical and Engineering Sciences 367, 489–507 (2009), publisher: Royal Society

    2009

  2. [10]

    Scaling of the turbu- lence transition threshold in a pipe,

    B. Hof, A. Juel, and T. Mullin, “Scaling of the turbu- lence transition threshold in a pipe,” Phys. Rev. Lett. 91, 244502 (2003)

    2003

  3. [11]

    Critical behavior in the relaminarization of localized turbulence in pipe flow,

    Ashley P. Willis and Rich R. Kerswell, “Critical behavior in the relaminarization of localized turbulence in pipe flow,” Phys. Rev. Lett. 98, 014501 (2007)

    2007

  4. [12]

    Theoretical perspective on the route to tur- bulence in a pipe,

    D. Barkley, “Theoretical perspective on the route to tur- bulence in a pipe,” J. Fluid Mech. 803, P1 (2016)

    2016

  5. [13]

    Finite lifetime of turbulence in shear flows,

    Bj¨ orn Hof, Jerry Westerweel, Tobias M. Schneider, and Bruno Eckhardt, “Finite lifetime of turbulence in shear flows,” Nature 443, 59–62 (2006), publisher: Nature Publishing Group

    2006

  6. [14]

    Transition to turbulence in wall- bounded flows: Where do we stand?

    Paul Manneville, “Transition to turbulence in wall- bounded flows: Where do we stand?” Mechanical En- gineering Reviews 3, 15–00684–15–00684 (2016)

    2016

  7. [15]

    Directed percolation and the transition to turbulence,

    Bj¨ orn Hof, “Directed percolation and the transition to turbulence,” Nat Rev Phys 5, 62–72 (2023)

    2023

  8. [16]

    Onset of meso-scale turbulence in active nematics,

    Amin Doostmohammadi, Tyler N Shendruk, Kristian Thijssen, and Julia M Yeomans, “Onset of meso-scale turbulence in active nematics,” Nature communications 8, 15326 (2017)

    2017

  9. [17]

    Mechanism for turbulence proliferation in subcritical flows,

    Anna Frishman and Tobias Grafke, “Mechanism for turbulence proliferation in subcritical flows,” Proceed- ings of the Royal Society A: Mathematical, Physi- cal and Engineering Sciences 478, 20220218 (2022), https://royalsocietypublishing.org/doi/pdf/10.1098/rspa.2022.0218

    2022

  10. [18]

    The openpipeflow navier–stokes solver,

    Ashley P. Willis, “The openpipeflow navier–stokes solver,” SoftwareX 6, 124–127 (2017)

    2017

  11. [19]

    A mathematical example displaying fea- tures of turbulence,

    Eberhard Hopf, “A mathematical example displaying fea- tures of turbulence,” Communications on Pure and Ap- plied Mathematics 1, 303–322 (1948)

    1948

  12. [20]

    The dynamics of bursting process in wall turbulence,

    Tomoaki Itano and Sadayoshi Toh, “The dynamics of bursting process in wall turbulence,” Journal of the Physical Society of Japan 70, 703–716 (2001), https://doi.org/10.1143/JPSJ.70.703

    2001 doi

  13. [21]

    Turbulence transition and the edge of chaos in pipe flow,

    Tobias M. Schneider, Bruno Eckhardt, and James A. Yorke, “Turbulence transition and the edge of chaos in pipe flow,” Phys. Rev. Lett.99, 034502 (2007), publisher: American Physical Society

    2007

  14. [22]

    Turbulence transition in pipe flow,

    Bruno Eckhardt, Tobias M. Schneider, Bjorn Hof, and Jerry Westerweel, “Turbulence transition in pipe flow,” Annu. Rev. Fluid Mech. 39, 447–468 (2007)

    2007

  15. [23]

    Transi- tion to turbulence in pipe flow,

    Marc Avila, Dwight Barkley, and Bj¨ orn Hof, “Transi- tion to turbulence in pipe flow,” Annual Review of Fluid Mechanics 55, 575–602 (2023)

    2023

  16. [24]

    Transition in pipe flow: the saddle structure on the boundary of turbulence,

    Y. Duguet, A. P. Willis, and R. R. Kerswell, “Transition in pipe flow: the saddle structure on the boundary of turbulence,” J. Fluid Mech. 613, 255–274 (2008)

    2008

  17. [25]

    Sensitive depen- dence on initial conditions in transition to turbulence in pipe flow,

    Holger Faisst and Bruno Eckhardt, “Sensitive depen- dence on initial conditions in transition to turbulence in pipe flow,” Journal of Fluid Mechanics 504, 343–352 (2004)

    2004

  18. [26]

    Relative periodic orbits form the backbone of turbulent pipe flow,

    N. B. Budanur, K. Y. Short, M. Farazmand, A. P. Willis, and P. Cvitanovi´ c, “Relative periodic orbits form the backbone of turbulent pipe flow,” Journal of Fluid Me- chanics 833, 274–301 (2017)

    2017

  19. [27]

    Transition to turbulence in a shear flow,

    Bruno Eckhardt and Alois Mersmann, “Transition to turbulence in a shear flow,” Phys. Rev. E 60, 509–517 (1999)

    1999

  20. [28]

    Edge of chaos in a parallel shear flow,

    Joseph D. Skufca, James A. Yorke, and Bruno Eckhardt, “Edge of chaos in a parallel shear flow,” Phys. Rev. Lett. 96, 174101 (2006)

    2006

  21. [29]

    Relative periodic orbits in transitional pipe flow,

    Yohann Duguet, Chris C. T. Pringle, and Rich R. Ker- swell, “Relative periodic orbits in transitional pipe flow,” Physics of Fluids 20, 114102 (2008)

    2008

  22. [30]

    Transition in local- ized pipe flow turbulence,

    Fernando Mellibovsky, Alvaro Meseguer, Tobias M. Schneider, and Bruno Eckhardt, “Transition in local- ized pipe flow turbulence,” Phys. Rev. Lett. 103, 054502 (2009)

    2009

  23. [31]

    Slug gen- esis in cylindrical pipe flow,

    Y. Duguet, A. P. Willis, and R. R. Kerswell, “Slug gen- esis in cylindrical pipe flow,” Journal of Fluid Mechanics 663, 180–208 (2010)

    2010

  24. [32]

    Streamwise-localized solutions at the onset of turbu- lence in pipe flow,

    M. Avila, F. Mellibovsky, N. Roland, and B. Hof, “Streamwise-localized solutions at the onset of turbu- lence in pipe flow,” Phys. Rev. Lett. 110, 224502 (2013), publisher: American Physical Society

    2013

  25. [33]

    Heteroclinic path to spatially localized chaos in pipe flow,

    N. B. Budanur and B. Hof, “Heteroclinic path to spatially localized chaos in pipe flow,” Journal of Fluid Mechanics 827, R1 (2017)

    2017

  26. [34]

    (Cambridge University Press, 2002)

    Edward Ott, Chaos in Dynamical Systems , 2nd ed. (Cambridge University Press, 2002)

    2002

  27. [35]

    Edge state in pipe flow experiments,

    A. De Lozar, F. Mellibovsky, M. Avila, and B. Hof, “Edge state in pipe flow experiments,” Phys. Rev. Lett. 108, 214502 (2012)

    2012

  28. [36]

    The Proper Orthogonal Decomposition in the Analysis of Turbulent Flows,

    G. Berkooz, P. Holmes, and J. L. Lumley, “The Proper Orthogonal Decomposition in the Analysis of Turbulent Flows,” Annual Review of Fluid Mechanics 25, 539–575 (1993)

    1993

  29. [37]

    Principal component analysis: a review and recent developments,

    Ian T. Jolliffe and Jorge Cadima, “Principal component analysis: a review and recent developments,” Philosoph- ical Transactions of the Royal Society A 374, 20150202 (2016)

    2016

  30. [38]

    See Supplemental Material,

  31. [39]

    Splitting of a turbulent puff in pipe flow,

    Masaki Shimizu, Paul Manneville, Yohann Duguet, and 6 Genta Kawahara, “Splitting of a turbulent puff in pipe flow,” Fluid Dynamics Research 46, 061403 (2014), pub- lisher: IOP Publishing

    2014

  32. [40]

    Dynamics and prolif- eration of turbulent stripes in plane-Poiseuille and plane- Couette flows,

    E. Marensi, G. Yalnız, and B. Hof, “Dynamics and prolif- eration of turbulent stripes in plane-Poiseuille and plane- Couette flows,” Journal of Fluid Mechanics 974, A21 (2023)

    2023

  33. [41]

    Saddle avoidance of noise-induced transitions in multiscale sys- tems,

    Reyk B¨ orner, Ryan Deeley, Raphael R¨ omer, Tobias Grafke, Valerio Lucarini, and Ulrike Feudel, “Saddle avoidance of noise-induced transitions in multiscale sys- tems,” Phys. Rev. Res. 6, L042053 (2024)

    2024

  34. [42]

    Extreme events in transitional turbulence,

    S. Gom´ e, L. S. Tuckerman, and D. Barkley, “Extreme events in transitional turbulence,” Phil. Trans. R. Soc. A. 380, 20210036 (2022)

    2022

  35. [43]

    Turbulent spots in channel flow: An experimental study,

    Gr´ egoire Lemoult, Konrad Gumowski, Jean-Luc Aider, and Jos´ e Eduardo Wesfreid, “Turbulent spots in channel flow: An experimental study,” The European Physical Journal E 37, 25 (2014)

    2014

  36. [44]

    Phase transition to turbulence in spatially extended shear flows,

    Lukasz Klotz, Gr´ egoire Lemoult, Kerstin Avila, and Bj¨ orn Hof, “Phase transition to turbulence in spatially extended shear flows,” Phys. Rev. Lett. 128, 014502 (2022)

    2022

  37. [45]

    Rare Event Algorithm Links Transitions in Turbulent Flows with Activated Nucleations,

    Freddy Bouchet, Joran Rolland, and Eric Simonnet, “Rare Event Algorithm Links Transitions in Turbulent Flows with Activated Nucleations,” Physical Review Let- ters 122, 074502 (2019), publisher: American Physical Society

    2019

  38. [46]

    Boundary crisis and long transients of the Atlantic overturning circulation mediated by an edge state,

    Reyk B¨ orner, Oliver Mehling, Jost von Hardenberg, and Valerio Lucarini, “Boundary crisis and long transients of the Atlantic overturning circulation mediated by an edge state,” (2025), arXiv:2504.20002 [nlin]

    2025 arXiv

  39. [47]

    D. J. Acheson and D. J. Acheson, Elementary Fluid Dy- namics, Oxford Applied Mathematics and Computing Science Series (Oxford University Press, 1990)

    1990

  40. [48]

    Speed and structure of turbulent fronts in pipe flow,

    Baofang Song, Dwight Barkley, Bj¨ orn Hof, and Marc Avila, “Speed and structure of turbulent fronts in pipe flow,” Journal of Fluid Mechanics813, 1045–1059 (2017), arXiv:1603.04077 [physics]

    2017 arXiv

  41. [49]

    Revealing the state space of turbulent pipe flow by symmetry reduc- tion,

    A. P. Willis, P. Cvitanovi´ c, and M. Avila, “Revealing the state space of turbulent pipe flow by symmetry reduc- tion,” Journal of Fluid Mechanics 721, 514–540 (2013). END MATTER Direct numerical simulations of pipe flow Considering the flow in a circular pipe as described in...

    2013

  42. [50]

    In the case that Φ(t) =−1 is ever encountered the procedure is terminated (which never occurred in our investigation)

    is constant for a time of at least τ = 50D/ ¯U. In the case that Φ(t) =−1 is ever encountered the procedure is terminated (which never occurred in our investigation). The parameters used: For Re = 2200: ¯l1 = 10.75, σ1 = 2.3, ¯l2 = 21.5, σ2 = 3.1. For Re = 2100: ¯l1 = 9.4, σ1 ...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.