Scaling of the minimal energy for turbulence transition in pipe flow
summary
The gist
Predicting the transition of turbulence in pipe flow remains a fundamental problem in fluid dynamics.
In short
This work investigates how turbulence starts in pipe flow by finding 'nonlinear optimal perturbations' (NLOPs). It shows that these perturbations grow by combining linear instability mechanisms (Orr and lift-up) with nonlinear interactions. The minimum energy needed to cause transition scales as O(Re^-3), providing a way to predict the onset of turbulence.
Key concepts
- Nonlinear Optimal Perturbations (NLOPs)
- These are specific, localized perturbations that grow fastest in time. They achieve this by combining two mechanisms: the Orr mechanism and the lift-up mechanism, linked by a nonlinear oblique phase. These structures are crucial because they efficiently trigger transition from laminar flow.
- Orr Mechanism
- This is an initial growth phase where the energy gain of a perturbation is independent of its initial size (E0) but increases linearly with the Reynolds number (Re). It sets up the initial instability that allows for further nonlinear development and energy extraction.
- Lift-up Mechanism
- This mechanism involves strong nonlinear interactions between various oblique modes, leading to the growth of streaks and cross-sectional vortices. This phase is highly dependent on the initial energy (E0) and Reynolds number, significantly amplifying the perturbation's energy before transition occurs.
Terminology used across episodes
This episode discusses
The paper
Scaling of the minimal energy for turbulence transition in pipe flow · Read on arXiv
Center of Applied Space Technology and Microgravity at University of Bremen · MAPEX Center for Materials and Processes at University of Bremen
Predicting the transition to turbulence in pipe flow remains a fundamental problem in fluid dynamics. We use a variational approach to compute nonlinear optimal perturbations to the laminar flow. Optimal perturbations are structurally similar for all Reynolds numbers, 1750 at most Re at most 10 4, considered. As Re increases, they increasingly localise in space with thickness scaling as Re-1/3. They grow via the Orr mechanism, followed by a phase of energy transfer across oblique waves of different axial wavenumber, characterised by a rapid delocalisation. The energy gain in the Orr and oblique-wave phases is independent of the initial perturbation energy, E 0. Subsequently, rolls grow by nonlinear mechanisms and streaks via the lift-up effect. The energy gain of the (inviscid) Orr phase is independent of the Reynolds number, whereas it increases linearly with Re during the initial oblique-wave phase and quadratically in the roll and lift-up phase. We find that, transition occurs if the energy of the optimal perturbation exceeds a maximum energy threshold during its temporal evolution. This energy is given by the edge state energy and found to be independent of Re. As a result, the minimum perturbation energy required to cause transition in pipe flow scales as E 0 crit proportional to Re-3, which is consistent with our computation of minimal seeds. We note that insufficiently long pipes lead to an exponent of smaller magnitude and that much larger numerical resolutions are required to accurately compute the energy gain than to compute the critical threshold.
Transcript
Introduction to the show: ident: Physics. The day in general physics: instruments, methods and results across the field.
Kai: I'm Kai.
Mira: And I'm Mira, here with Lev.
Kai: We're looking at the paper: "Scaling of the minimal energy for turbulence transition in pipe flow".
Mira: Here's the gist of the paper. Predicting the transition of turbulence in pipe flow remains a fundamental problem in fluid dynamics.
Kai: The minimum perturbation energy required to cause transition in pipe flow scales as O (Re-3).
Mira: Understanding and predicting the transition between laminar and turbulent states is a central problem in fluid dynamics since the pioneering work of Reynolds.
Kai: The difficulty in predicting transition stems from the linear stability of the system up to at least Re ≥ 107.
Title and authors: Kai: So this paper is called "Scaling of the minimal energy for turbulence transition in pipe flow". Mira, what's the real takeaway from that title?
Mira: It sounds very specific, focusing on how much energy you need to start turbulence in a pipe, and scaling with something. It suggests they are looking at the minimum amount of disturbance required to break laminar flow.
Lev: From an engineering standpoint, I wonder if this minimal energy is something we can actually measure in a real system without needing impossibly small initial perturbations.
Kai: Exactly, because usually these transition studies rely on linear stability theory which tells us that up to Reynolds numbers like one hundred seven the flow is stable <ref:2606.23269#pg2>. This paper seems to be pushing beyond that linear boundary to find where things actually break down experimentally.
Mira: Right, and the authors are using a variational approach to compute nonlinear optimal perturbations for pipe flow when the Reynolds number is less than or equal to five thousand <ref:2606.23269#pg1,a variational approach to compute nonlinear optimal perturbations>. That’s a specific range they’re testing.
Lev: If you're looking at real hardware, that Reynolds number constraint is important because we have to worry about how much computational overhead it adds when trying to model these nonlinear effects.
Kai: It seems the whole point is showing that even in this relatively low Reynolds number regime, we can still find the optimal way to trigger transition using these nonlinear perturbations.
The paper's summary: Mira: So, what’s the actual finding here? Basically, they found that transition happens when the energy of a perturbation crosses a certain threshold and it scales with Reynolds number in a specific way.
Kai: They found that this minimum required energy scales as O of Re to the power negative three. That's quite a specific scaling law they’ve derived for the threshold.
Lev: O of Re to the power negative three is something we need to check against our simulations because it dictates how small an initial seed perturbation needs to be if we want a certain transition probability.
Mira: The authors identify these optimal perturbations as nonlinear optimal perturbations, or NLOPs, which are special disturbances that combine the Orr and lift-up mechanisms with a nonlinear oblique phase.
Kai: They show that these NLOPs are what actually trigger transition efficiently because they outgrow the simpler linear optimal perturbations.
Lev: But I have to ask, how robust is this finding? Does it hold up when we try to run these simulations on actual hardware instead of just on a pseudo-spectral code like nsPipe-GPU?
Mira: Well, they set up a checkpointing procedure using the direct adjoint optimization procedure in the GPU version of their Navier Stokes code. They found that for a pipe length L equal to fifty the energy gain becomes independent of L, which is good for them <ref:2606.23269#pg3>.
Kai: And they verified their spatial resolution by simulating the last optimal perturbation inside the laminar attractor at Reynolds number four thousand with a spatial resolution of one hundred forty-four radial points, one hundred twenty azimuthal modes, and five hundred forty axial modes <ref:2606.23269#pg3,the last optimal perturbation inside the laminar attractor>.
The paper's improvements: Kai: One big improvement they point out is how the optimal perturbations behave as the Reynolds number increases. They say that while the structure stays similar, it starts to localize more, with their radial thickness scaling as delta r proportional to Re to the power negative one over three.
Mira: So, as flow gets faster and more turbulent, these optimal structures don't just get bigger; they get squeezed in space. That localization is key for understanding where the energy goes.
Lev: If the thickness scales that way, it means that for higher Reynolds numbers, you need a much sharper initial seed perturbation to maintain that optimal structure before it triggers transition.
Kai: They also discuss how the energy gain during the Orr phase increases linearly with Re but is independent of your initial energy E0. That part seems quite powerful because it suggests the mechanism itself drives the growth initially.
Mira: And then there’s that lift-up phase, where they say the energy gain scales as proportional to Re to the power two. That nonlinear interaction between oblique waves and lift-up is what really pushes things over the edge.
Lev: So, if we were trying to build a real hardware system for this, we'd need to manage that transition from linear growth driven by Re to this quadratic growth phase driven by E0 and Re squared. That’s a lot of dynamics for any controller.
Conclusion: Kai: So, wrapping up the Scaling of the minimal energy for turbulence transition in pipe flow paper, we see that NLOPs are the key drivers here because they combine those initial linear Orr growth with that subsequent strong nonlinear interaction from oblique waves and lift-up.
Mira: The main conclusion they draw is that to find this minimal seed, you need an initial perturbation energy E crit zero that scales as Etr multiplied by Re to the power negative three, where Etr is approximately zero point one three <ref:2606.23269#pg1>.
Lev: For hardware design, this means if we want to reach a certain transition probability at a high Reynolds number, we need to ensure our seed energy is proportional to Re negative three.
Kai: And they also noted that this minimal seed structure remains similar but localizes as the Reynolds number increases, with that radial thickness decreasing as delta r proportional to Re negative one over three.
Mira: It’s interesting because they pinpoint exactly how the nonlinearities enhance the growth via that lift-up mechanism, which strongly depends on E0 and Re squared.
Lev: I just want to say that while their scaling relationship is helpful for prediction, we still have to figure out how much overhead that actual nonlinear optimization procedure adds when you try to run it on real hardware instead of a simulator.
Kai: Yeah, so we’ve seen what they’ve done with this specific case in the Scaling of the minimal energy for turbulence transition in pipe flow paper. That's where we are for now.
More episodes
- 2610.34250-Yb/Sr optical clock comparison at 2.08E-18 uncertainty over a 1350-km fiber link
- 2610.36514-Integer electron-count cuts in variational second-order reduced-density-matrix theory. I. Molecular dissociation with fixed real-space and orbital partitions
- 2611.14082-Two atoms with strong light-induced dipolar interactions
- 2611.14040-Characterization of BNLs 20 and 30 micrometer thick LGADs for the Electron Ion Collider
- 2611.13399-Lift-element-based dynamic network of vortex gust-airfoil interactions
- 2611.13476-Hopfion Superlattices in 3D Twisted Moir'e Light Fields
- 2611.13642-Investigation of the timing performance of double-sided p-type 3D silicon sensors produced by CNM
- 2611.13830-SiPM optical readout of a PICOSEC Micromegas precise timing detector
- 2611.14152-Where are the shortest-path nodes in spatial networks?
- 2611.14219-Sub-Ice: semi-automated mapping of basal channels from ice shelf surface elevation data