Final Report Summary - PIPEEDGE (Analysis of coherent states at the laminar-turbulence boundary in pipe flow)
Natural turbulent flows exhibit irregular eddy motion in a wide range of temporal and spatial scales. This complexity can arise due to numerous driving mechanisms, e.g. rotation, temperature and density gradients, shear stresses or the presence of walls. Due to the difficulties of the general problem, there has been a need to seek fundamental laboratory flows from which the important physics can be extracted. The flow in a circular pipe is a canonical shear flow, where despite the simple geometry turbulence is encountered in its full complexity. In general, the transfer of fluids in pipes is energetically far more efficient if the fluid motion is smooth, since friction losses are much lower. Therefore, the transition from laminar to turbulent flow poses a challenge of great practical and theoretical importance. The aim of this project is to advance the understanding of the onset of turbulence in pipe flow by investigating:
(i) the boundary between the laminar and the turbulent flow, which is believed to be governed by coherent flow structures, so-called edge states; and
(ii) the mechanisms that sustain and expand turbulence contaminating the laminar flow.
We performed laboratory experiments to investigate the recurrence of coherent flow structures associated with the edge boundary. The laminar base flow was perturbed using a jet-like perturbation through a tiny hole in the pipe wall. The perturbation resulted in wavy patterns that grew in size as they were advected by the flow and eventually broke down to turbulence downstream. The breakdown to turbulence implies that the phase-space boundary between laminar and turbulent state must have been crossed and the observed flow features may resemble close visits to the edge boundary. Further analysis of wave propagation speeds and frequency showed that we have found a varicose instability (i.e. arising from an inflection point in the streamwise velocity profile) that gives rise to a wave-train of hairpin vortices. Secondary instabilities at higher amplitudes and Reynolds numbers made a systematic analysis rather difficult in the sense that power spectra showed a number of peaks rather than a single well defined maximum. The next step is to use selected cases as initial conditions for numerical edge tracking algorithms to further elucidate the connection between the observed flow and the dynamics of the edge boundary.
We used data from direct numerical simulations and analysed the interfaces bounding localised turbulent flow regions to shed light on their propagation mechanism. We found that that the bounding edges of a turbulent puff propagate mostly through outward and inward motion of continuous surfaces that propagate relative to the fluid velocity. At the same time, they are advected by the flow and stretched out by the straining of large eddies and by the laminar velocity profile, resulting in very large surface areas. In addition to the continuous propagation mechanism relative to the fluid, surface parts may locally detach giving rise to small disconnected blobs of laminar fluid enclosed in the puff and patches of turbulent fluid disconnected from the puff. The former are expected to play the key role in the splitting process. With growing computer power, the method may be applied at higher Reynolds number and shed light on the differences between puffs and slugs. It is expected that at higher Re the leading edge will be mostly entraining fluid and propagate faster.
During fall 2011, results were presented at the European turbulence conference (ETC 13) in Warsaw. A scientific publication entitled 'Lagrangian to laminar-turbulence interfaces in transitional pipe flow' is currently in review in the Journal of Fluid Mechanics. EC contribution was adequately acknowledged.
During the course of the project, the fellow was offered a group leader position at ETH Zurich. Crucial for obtaining the position were new scientific skills, as well as mentoring and project management abilities acquired during the project period. Indeed, within the framework of the new group leader position the fellow will strongly benefit and intends to further strengthen the research network established during this project.
We expect that a better understanding of the mechanisms that induce and sustain turbulence potentially has a large impact on society, as fluid flows play an important role in a large number of applications. In natural and industrial processes, turbulence can have unfavourable and even detrimental effects, so means of suppressing the accompanying vigorous velocity and pressure fluctuations would have important practical implications. For example, the energy dissipation of turbulent flows is much larger than that of laminar ones and consequently, it is far more costly to transport fluid through a vessel or to propel a vehicle if the flow is turbulent. In oil pipelines, the pressures required to pump the fluid is typically more than thirty times larger than would be necessary if the flow could be held laminar. Additionally, in applications problems arise in the transitional regime where pressure fluctuations are caused by the intermittent switching between laminar and turbulent flow, e.g. puffs and slugs can disrupt production processes and the accompanying pressure changes can damage equipment. Our observations of wave-like solutions that govern the dynamics during transition to turbulence underline its potential relevance for turbulence control.
(i) the boundary between the laminar and the turbulent flow, which is believed to be governed by coherent flow structures, so-called edge states; and
(ii) the mechanisms that sustain and expand turbulence contaminating the laminar flow.
We performed laboratory experiments to investigate the recurrence of coherent flow structures associated with the edge boundary. The laminar base flow was perturbed using a jet-like perturbation through a tiny hole in the pipe wall. The perturbation resulted in wavy patterns that grew in size as they were advected by the flow and eventually broke down to turbulence downstream. The breakdown to turbulence implies that the phase-space boundary between laminar and turbulent state must have been crossed and the observed flow features may resemble close visits to the edge boundary. Further analysis of wave propagation speeds and frequency showed that we have found a varicose instability (i.e. arising from an inflection point in the streamwise velocity profile) that gives rise to a wave-train of hairpin vortices. Secondary instabilities at higher amplitudes and Reynolds numbers made a systematic analysis rather difficult in the sense that power spectra showed a number of peaks rather than a single well defined maximum. The next step is to use selected cases as initial conditions for numerical edge tracking algorithms to further elucidate the connection between the observed flow and the dynamics of the edge boundary.
We used data from direct numerical simulations and analysed the interfaces bounding localised turbulent flow regions to shed light on their propagation mechanism. We found that that the bounding edges of a turbulent puff propagate mostly through outward and inward motion of continuous surfaces that propagate relative to the fluid velocity. At the same time, they are advected by the flow and stretched out by the straining of large eddies and by the laminar velocity profile, resulting in very large surface areas. In addition to the continuous propagation mechanism relative to the fluid, surface parts may locally detach giving rise to small disconnected blobs of laminar fluid enclosed in the puff and patches of turbulent fluid disconnected from the puff. The former are expected to play the key role in the splitting process. With growing computer power, the method may be applied at higher Reynolds number and shed light on the differences between puffs and slugs. It is expected that at higher Re the leading edge will be mostly entraining fluid and propagate faster.
During fall 2011, results were presented at the European turbulence conference (ETC 13) in Warsaw. A scientific publication entitled 'Lagrangian to laminar-turbulence interfaces in transitional pipe flow' is currently in review in the Journal of Fluid Mechanics. EC contribution was adequately acknowledged.
During the course of the project, the fellow was offered a group leader position at ETH Zurich. Crucial for obtaining the position were new scientific skills, as well as mentoring and project management abilities acquired during the project period. Indeed, within the framework of the new group leader position the fellow will strongly benefit and intends to further strengthen the research network established during this project.
We expect that a better understanding of the mechanisms that induce and sustain turbulence potentially has a large impact on society, as fluid flows play an important role in a large number of applications. In natural and industrial processes, turbulence can have unfavourable and even detrimental effects, so means of suppressing the accompanying vigorous velocity and pressure fluctuations would have important practical implications. For example, the energy dissipation of turbulent flows is much larger than that of laminar ones and consequently, it is far more costly to transport fluid through a vessel or to propel a vehicle if the flow is turbulent. In oil pipelines, the pressures required to pump the fluid is typically more than thirty times larger than would be necessary if the flow could be held laminar. Additionally, in applications problems arise in the transitional regime where pressure fluctuations are caused by the intermittent switching between laminar and turbulent flow, e.g. puffs and slugs can disrupt production processes and the accompanying pressure changes can damage equipment. Our observations of wave-like solutions that govern the dynamics during transition to turbulence underline its potential relevance for turbulence control.