New Theory on Turbulence Generation and Structure



题目:New Theory on Turbulence Generation and Structure

报告人:Prof. Chaoqun LiuUniversity of Texas at Arlington, Arlington, Texas, USA





    Dr. Chaoqun Liu received his BS in 1968 and MS in 1981 from Tsinghua University, Beijing, China and PhD in 1989 from University of Colorado at Denver, Colorado, USA. He is currently the Distinguished Professor, a highest honoured academic title, and Director of Center for Numerical Simulation and Modelling at University of Texas at Arlington, Arlington, Texas, USA. He has worked on high order direct numerical simulation (DNS) and large eddy simulation (LES) for flow transition and turbulence since 1990. As PI, he has been awarded with 47 federal research grants of over 5.7 million US dollars since 1990 in the United States. As Co-PI, he just received an AFOSR MURI grant of $7.3 millions for next 5 years. He has published 7 books, 93 journal papers and 134 conference papers. As a principal lecturer, he just held TGS2015 Workshop, New Theory of Turbulence Generation and Sustenance, Tsinghua University, Beijing, China, June 4-6, 2015, sponsored by 20 research institutions with over 220 audiences. He also chaired first and third AFOSR international conference on DNS/LES. He is a well- established expert in high order DNS/LES for flow transition, turbulence, and shock and boundary layer interaction.


1. New Theory

After 28-years research. Chaoqun Liu at University of Texas at Arlington presents new theories on turbulence generation and structure, which are consistent without any self-contradiction and well explain turbulence generation and structure.

1Turbulent flow has unique and deterministic solution. Turbulence is not generated by vortex “breakdown”, but vortex “buildup”, 2) Vorticity and vortex are totally different concepts. A newvortex identification method has been given by Liu et al., 3)The nature of turbulence generation is that fluid cannottolerate high shear stress and vorticity must transfer to rotation, 4) The role of linear unstable modes is to push the vorticity up from the wall (rollup). When leaving away from the wall, the streamwise vortex legs and spanwise vortex rings will form due to the trend from shear to rotation. The rotation is a stable state with minimized dissipation (deformation is very small.), 5) There is no such a process that the-vortex self-deformsto hairpin vortex. The-vortex root and ring head are formed separately by different mechanisms and ring is not part of -vortex which is a pair of open rotation cores and is not a vortex tube, 6)A momentum deficit zone (low speed zone) is formed above the -vortex and further generates a high shear due to the vortex root ejection. The vortex rings are generated by the high shear layer (K-H type) instability,7)The second level vortices are generated by second level shear layer which is caused by first level vortex rings through sweeps, ejections, positive and negative spikes, 8)Multiple vortex rings are all formed by shear layerinstability which is generated by momentum deficit, 9)The vortex structure is quite stable,10)All small vortices are generated by shear layerinstability without exception. In other words, “shear layer instability is the mother of turbulence.” There is no way that small vortices (turbulence) are generated by “large vortex breakdown”, 11)The multiple level shear layers are generated by vortex sweeps and ejections. They form the multiple level shear layers and multiple vortex rings. The second level smaller vortices are conjugated with the first level mother vortex which does not support the vortex breakdown mechanism by Feyhmann while the large vortex is live when the small vortices are generated, 12)The energy transport channel is that the high energy isbrought down to the lower boundary layer by multiple level sweeps. Without these sweeps, all small vortices (turbulence) would be dissipated quickly. Large vortex cannot pass energy to smaller vortices through “vortex breakdown” which was never observed by any experiment or DNS. It is easy to make any faked vortex breakdown by changing threshold,13) The disordering of flow structure is mainly not caused by the background noise or non-symmetric spanwise boundary conditions, but internal property of the multiple level vortex structure. The non-symmetry starts in the middle of the vortex packages, 14) Richardson eddy cascade revisit: It is not found, 15) Kolmogorov hypothesis revisit: There is no proofthat large vortex passes energy to small vortices through vortex breakdown, 16)Turbulence bursting and intermittency: The term is generated by misunderstanding of turbulence package self-motion and relative motion. Turbulence cannot suddenly burst and suddenly disappear.

2. Vortex and wave

Vortex is linked to fluid particle but wave is linked only to physical quantity where no fluid particle is involved. The linearized Euler equation has 5 eigenvalues, u, u, u, u+c, u-c. The first three eigenvalues are linked with vortex and last two are linked with waves.

3. Vorticity, vorticity flux and vorticity line length andvolume vorticity

Like other physical quantity, vorticity has transport equation in a flow field and changes all the time. However, vorticity flux will keep constant as vorticity line cannot be stopped inside the flow field. For example, if we have 100 vorticity tubes comein from front side, we then must have exactly 100 vorticity tubes come out from backside. All vorticity lines are straight in a laminar boundary layer which is minimized in length. We define a term called “volume vorticity” which is product of vorticity flux and the vorticity tube length. The vorticity line length and the volume vorticity are minimized in laminar inflow (like Blassius solution). Since the vorticity lines are stretched, tilted, and tangled during the transition process, the vorticity length is increased and thus the volume vorticity is increased simultaneously. Volume vorticity is a right quantity to mark the flow status, i.e. Laminar, transition, or turbulent.

4. Instantaneous velocity in boundary layer

In time and spanwise averaged sense, velocity is larger in main flow and smaller in boundary layer. However, according to our DNS, the maximum instantaneous velocity is located inside the boundary layer, in particular near the vortex ring neck due to the fast rotation of vortex ring. The maximum velocity is larger than the main flow.

5. Fast rotation of vortex core

According to our DNS and experiment, the vortex rings have very fast rotation core with thousands of circles per second.

6. Turbulence modeling

Currently widely used eddy viscosity model is based on Boussinesq assumption that Reynolds stress is proportional to mean rate of resolved strain tensor. The assumption has no scientific foundation.

7. Transition and turbulent flow control

Basically, there are two types of flow control technology: modes control and structure control. Theoretically modes control is powerful, but there are rarely reports about success of modes control although tens of millions dollars were spent in US and Europe. The reason is quite simple that the turbulence generation is a non-linear process and the base flow is changing all the time. Our flow control method development should be focused on structure control, e.g. blowing, riblet, surface vibration, etc.The near wall structure determines the surface drag, but the near wall structure is determined by the large and medium vortexstructure in the log area and outer boundary layer. Therefore, we must understand the turbulence structure before we can efficiently reduce the surface drag.

8. Vista of turbulence research

There must be breakthrough in turbulence research with fast development of computational and experimental technology including the physics of turbulence generation and sustenance, turbulence structure, theoretic proofs, turbulence control, drag reduction, and new turbulence models based on DNS.