Not logged in
PANGAEA.
Data Publisher for Earth & Environmental Science

Morón, Daniel; Avila, Marc (2024): Turbulent puffs in transitional pulsatile pipe flow at moderate pulsation amplitudes [dataset]. PANGAEA, https://doi.org/10.1594/PANGAEA.969126

Always quote citation above when using data! You can download the citation in several formats below.

Published: 2024-07-03 • DOI registered: 2024-08-01

RIS CitationBibTeX Citation Share

Abstract:
We show that, in the transitional regime of pulsatile pipe flow, at moderate-to-high amplitudes 0.5≲𝐴≲1, the first long-lived turbulent structures are localized and take the form of the puffs and slugs observed in statistically steady pipe flow. We perform direct numerical simulations at many pulsation frequencies (Wo), amplitudes, and Reynolds number (Re) and observe different dynamics of puffs and slugs. At certain flow parameters we find, using a causal analysis, that puffs actively make use of linear instabilities in the laminar Sexl-Womersley (SW) profile to survive the pulsation. Using all these lessons learned, we extend a low-order model by Barkley et al. [Nature (London) 526, 550 (2015)] to reproduce these dynamics. We find a good agreement between the extended model and our numerical results in a broad parametric space of pulsation amplitudes 0.5≲𝐴≲1, frequencies Wo≳5 and 2100≤Re≤3000. With the help of our numerical results, causal analysis and model, we determine that turbulence production has two sources at these flow parameters: the mean shear as in statistically steady pipe flow and the instabilities of the instantaneous pulsatile mean profile.
Keyword(s):
causality; direct numerical simulation; reduced-order model of turbulence; transitional regime
Related to:
Morón, Daniel; Avila, Marc (2024): Turbulent puffs in transitional pulsatile pipe flow at moderate pulsation amplitudes. 9, 24601, https://journals.aps.org/prfluids/abstract/10.1103/PhysRevFluids.9.024601
Funding:
German Research Foundation (DFG), grant/award no. AV 120/6-1: Instabilities, Bifurcations and Migration in Pulsating Flow (FOR 2688)
Parameter(s):
#NameShort NameUnitPrincipal InvestigatorMethod/DeviceComment
1 Binary ObjectBinaryMorón, DanielNumerical simulated
2 Binary Object (File Size)Binary (Size)BytesMorón, DanielNumerical simulated
3 FigureFigMorón, DanielNumerical simulated
4 TitleTitleMorón, DanielNumerical simulated
5 Analytical methodAnalyt methodMorón, DanielNumerical simulated
6 File nameFile nameMorón, DanielNumerical simulated
7 VariableVariableMorón, DanielNumerical simulated
8 File formatFile formatMorón, DanielNumerical simulated
9 DescriptionDescriptionMorón, DanielNumerical simulated
License:
Creative Commons Attribution 4.0 International (CC-BY-4.0)
Status:
Curation Level: Enhanced curation (CurationLevelC) * Processing Level: PANGAEA data processing level 2 (ProcLevel2)
Size:
123 data points

Data

Download dataset as tab-delimited text — use the following character encoding:

All files referred to in data matrix can be downloaded in one go as ZIP or TAR. Be careful: This download can be very large! To protect our systems from misuse, we require to sign up for an user account before downloading.

1 
Binary		2 
Binary (Size) [Bytes]		3 
Fig		4 
Title		5 
Analyt method		6 
File name		7 
Variable		8 
File format		9 
Description
CODE1_nsPipe_Pulsatile.zip480.5 kBytesCode for the direct numerical simulation (DNS) of the Navier-Stokes equations NSE15014CODE1_nsPipe_Pulsatile.zipSources of the Fourier pseudospectral DNS code solving the NSE in a rigid pipe driven with the desired pulsatile bulk velocity. See github.com for more information about the methods. The code has been modified so the user can input with an in_cond.txt file the initial (laminar) flow and on top an axially localized perturbation. The code also allows the user to define a forcing term to perturb the flow locally. To run the case compile the code and save the executable in the same folder as the desired nsPipeFlow.in, and the *.txt files.
CODE2_TGA.zip23.4 kBytesCode in Matlab to perform Transient Growth Analysis (TGA) in pulsatile pipe flows15014CODE2_TGA.zipCode in Matlab to perform transient growth analysis or integrate the linearized NSE (LNSE). The codes in the folder: compute the TGA of a laminar profile driven with the desired waveform (A_MAIN_STHIN); and integrate the LNSE (P_MAIN_STHIN)
CODE3_Loc_Pert.zip424.1 kBytesCode in Matlab to generate the initial conditions for the DNS simulations15014CODE3_Loc_Pert.zipCode in Matlab to generate the localized initial perturbation used in the DNS. It reads an optimum perturbation computed with the TGA and then it localizes it axially in a desired length. An example is included, ready to be run
CODE4_TGA_MASTER_SLAVE.zip7.3 MBytesCode in Matlab to perform TGA on the artificial (slave) laminar profiles and the actual (master) profile15014CODE4_TGA_MASTER_SLAVE.zipCode in Matlab to perform transient growth analysis or integrate the linearized NSE (LNSE). The codes in the folder: compute the TGA of a laminar profile driven with the desired waveform (A_MAIN_STHIN); and integrate the LNSE (P_MAIN_STHIN). It considers the actual laminar pulsatile pipe flow (master) and an artificially imposed profile (slave).
CODE5_EBM.zip14.9 kBytesCode in Matlab of the extended Barkley Model15014CODE5_EBM.zipCode in Matlab that extends the reduced order model of puffs in pipe flow initially developed by Barkley et al. [Nature (London) 526, 550 (2015)] to pulsatile pipe flow. It integrates two non-linearly coupled one dimensional reaction-advection-diffusion equations with a multiplicative noise term, using an explicit Euler method, and finite differences for the spatial derivatives. The user can select the desired stencil length of the spatial derivatives.
CODE6_nsPipe_CUDA_MASTER_SLAVE.zip35.2 kBytesVersion of nsPipe in CUDA that integrates simultaneously the master-slave DNS15014CODE6_nsPipe_CUDA_MASTER_SLAVE.zipSources of the Fourier pseudospectral DNS code solving the NSE in a rigid pipe driven with the desired pulsatile bulk velocity. See github.com for more information about the methods. This version of the code is written in CUDA and runs in GPUs. Additionally, this version integrates simultaneously a master simulation (full DNS of pulsatile pipe flow) and the slave simulation, that is a full DNS except for the mean profile, that is artificially imposed.
SIM_SETUP.zip6.1 GBytesSimulation set-ups for all the DNS of pulsatile pipe flows and master-slave DNS15014SIM_SETUP.zipIt includes two folders. Both have all the files required to reproduce all the DNS of pulsatile pipe flow: one for the full DNS (nsPipe_CPU_setUp), the other for the master-slave DNS (nsPipe_GPU_setUp). As an example each folder includes the output of one of the simulations, and Matlab routines to postprocess them.
Fig1.zip375.2 MBytes1Space-time diagram of turbulence indicator for different pulsatile pipe flows15014Fig1.zipCross-section averaged axial vorticity squaremat files and .m filesAxial and time resolved cross-section averaged axial vorticity square for four different pulsatile pipe flow DNS, driven with a single harmonic pulsation and different Amplitudes, Reynolds and Womersley numbers. The folder includes the function Fig1.m to generate Fig 1 of the paper.
Fig2.zip4.6 GBytes2Phase-dependent statistics of turbulence in pulsatile pipe flow15014Fig2.zipVolume averaged axial vorticity squaremat files and .m filesPhase-dependent statistics of the volume averaged axial vorticity square of pulsatile pipe flow at Re=2100 and either Wo=15, A=1 or Wo=11, A=0.5. The folder includes the function Fig2.m to generate Fig 2 of the paper.
Fig3.zip6.4 kBytes3Parametric study of phase lag and turbulence front speed in pulsatile pipe flow15014Fig3.zipTime averaged phase lag and front speedmat files and .m filesTime averaged phase lag between the driving bulk velocity and turbulence intensity, and time averaged upstream front speed of turbulent patches. Results correspond to all the full DNS and master DNS considered in this study. It includes Fig3.m to generate Fig 3 of the paper.
Fig4.zip475.7 kBytes4Comparison between mean profiles in master and slave simulations15014Fig4.zipPhase-averaged mean flow profilemat files and .m filesMean profiles of master and slave DNS at Re=2100, Wo=11, A=0.5 both in the case of laminar and turbulent flows. It includes Fig4.m to generate Fig 4 of the paper.
Fig5.zip6.9 kBytes5Maximum energy growth of perturbations on top of master-slave laminar profiles15014Fig5.zipEnergy growthmat files and .m filesEnergy growth as the ratio between maximum energy and initial energy of the optimum perturbation on top of different laminar pulsatile pipe flows (master) and artificial profiles (slave). The results correspond to different combinations of pulsation amplitudes and Womersley numbers at Re=2100. It includes Fig5.m to generate Fig 5 of the paper.
Fig6.zip162.5 MBytes6Space-time diagram of turbulence indicator for two different pairs of master-slave simulations15014Fig6.zipCross-section averaged cross-section kinetic energymat files and .m filesAxial and time resolved cross-section averaged cross section kinetic energy for two pairs of master-slave DNS, at Re=2100, Wo=11 and A=0.1, A=0.5. The folder includes the function Fig6.m to generate Fig 6 of the paper.
Fig7.zip1.6 GBytes7Space-time diagram of turbulence indicator for different pairs of DNS and EBM results15014Fig7.zipTurbulence indicator in the EBM; cross-section averaged axial vorticity squaremat files and .m filesIt includes results of DNS and the EBM. In the case of the DNS it has the axial and time resolved cross-section averaged axial vorticity square for three different pulsatile pipe flows, driven with a single harmonic pulsation and different Amplitudes, Reynolds and Womersley numbers. For the case of the EBM it has the turbulence indicator (q in the model) for the same cases as in the DNS. The folder includes the function Fig7.m to generate Fig 7 of the paper.
Fig8.zip1.8 GBytes8Space-time diagram of turbulence indicator for two DNS and EBM results with different parameters15014Fig8.zipTurbulence indicator in the EBM; cross-section averaged axial vorticity squaremat files and .m filesIt includes results of DNS and the EBM. In the case of the DNS it has the axial and time resolved cross-section averaged axial vorticity square for two different pulsatile pipe flows, driven with a single harmonic pulsation and different Amplitudes, Reynolds and Womersley numbers. For the case of the EBM it has the turbulence indicator (q in the model) for the same case as in the DNS. It includes results of the fitted EBM, but also of the EBM with one of the parameters set to 0. The folder includes the function Fig8.m to generate Fig 8 of the paper.
Fig9.zip46.3 kBytes9Colormaps of turbulence upstream front speed in the EBM at different flow parameters15014Fig9.zipUpstream front speed; decay/survive binary datamat files and .m filesColormaps of turbulence upstream front speed in the EBM at different Reynolds and Womersley numbers, at A=0.5, A=0.75 and A=1. The results also include the survive/decay behavior of DNS determined according to heuristic thresholds. The folder includes the function Fig9.m to generate Fig 9 of the paper.
Fig10.zip6.9 kBytes10Laminar pulsatile pipe flow and the corresponding instantaneous instability15014Fig10.zipLaminar flow profile; maximum eigenvalue.m filesTime-resolved laminar flow profile at Re=2100, Wo=11, A=0.9, and corresponding maximum eigenvalue, as a proxy of the instantaneous instability of the laminar flow profile. This has been computed assuming the laminar flow profile to be quasi-steady. The folder includes Fig10.m to produce Fig 10 in the paper.
Fig11.zip2.9 GBytes11Space-time diagram of turbulence indicator for a DNS and EBM results with different parameters15014Fig11.zipTurbulence indicator in the EBM; cross-section averaged axial vorticity squaremat files and .m filesIt includes results of DNS and the EBM. In the case of the DNS it has the axial and time resolved cross-section averaged axial vorticity square for a steady pipe flow at Re=2400. For the case of the EBM it has the turbulence indicator (q in the model) for the same case as in the DNS, but for three different noise strengths in the model. The folder includes the function Fig11m to generate Fig 11 of the paper.