Flow parameters.
Abstract
Vortex definition and identification are extremely important for the study of fluid dynamics research. Liutex is a newly proposed concept that correctly represents vortex. Liutex is a vector whose direction is the local rotation axis and whose magnitude is twice the angular speed. To identify the unique structure of a vortex, a method known as the Liutex Core Line method has been developed, which displays the rotational core axis of a vortex. However, the original method is a manual method, which is not practical for real application, and an automatic algorithm is required for practical usage. Xu et al. proposed an algorithm by selecting the best line from a group of candidate lines, which is an important progress. In this chapter, from another perspective to solve this problem, a new algorithm is introduced based on forward Liutex magnitude gradient lines. Since gradient lines have the feature that they advance to the local maximums, the route will still result in a unique line, which avoids the process to find the best line. This algorithm has achieved some success for the Lambda vortex in early boundary layer transition.
Keywords
- Liutex
- vortex
- vortex core
- vortex identification
- Liutex lines
- Liutex core line
1. Introduction
Defining and visualizing vortices have been a challenge in fluid mechanics for several decades, even though the vortex is the essential part of formulating the turbulent flow that consists of countless vortices in different sizes and strengths. Turbulence generation and sustenance are still a mystery after intensive research efforts of centuries. Kaczorowski et al. [1] and Xi et al. [2] performed deep theory and experiment research on this topic, especially the Rayleigh–Bénard (RB) convection. In 2019, Zhou et al. elaborated the hydrodynamic instabilities that induced turbulent mixing in wide areas including inertial confinement fusion, supernovae, and their transition criteria. Zhou [3, 4] described in detail Rayleigh–Taylor and Richtmyer–Meshkov instability and some related models. Zhou’s analysis is systematical, comprehensive, and sophisticated for flow instability and vortex generation, covering long history and state-of-the-art advances in turbulence research, which has clearly shown guidance for further and deeper scientific research. Since vortex is the sinus and muscles of turbulence, it is of great importance to find out the definition of vortex. Human’s understanding of vortex has gone through three stages [5].
The first generation is vorticity-based methods which have several weaknesses. One severe weakness is the contamination by shear. This shear contamination can be easily seen in the near-wall laminar boundary layer where there are high levels of shear with small amounts of rotational motion. High vorticity magnitude is seen in areas of high shear and low rotation, while areas with stronger rotation can be found to have relatively lower vorticity magnitude values [6]. Though the misconception that a vortex is defined by vorticity is still being published in textbooks and research papers, we know this is not appropriate. Then, the second generation that uses eigenvalue-based vortex identification criteria was developed. The second generation of vortex identification methods mainly used the eigenvalues of the velocity gradient tensor. Among the second generation, there are some popular methods such as Q criterion [7],
Since the invention of Liutex, many researchers have used and tested it and reported that it is a strictly mathematical vortex definition and the currently best vortex identification method. Cuissa et al. [12] found that Liutex matches the analytical result of Lamb-Osceen vortex. Shen et al. [13] commented that Liutex is the only method that fully identifies small-scale vortices in vertical slit fishways simulation. Xu et al. found Liutex similarity, i.e., both the frequency and wavenumber spectrum of Liutex match the -5/3 law [14, 15, 16]. More reports on Liutex application can be found from [17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29] and four published books [30, 31, 32, 33]. Liutex is considered as the vortex definition in this chapter.
Since Liutex is a vector-based concept that provides the direction as well as the magnitude of the rigid rotation, we can integrate the Liutex vector passing through a point within the vortex region to obtain a special Liutex core line. The concept of Liutex core line was first proposed in [34]. In short, Liutex core lines are formed by points where local maximal Liutex magnitude is reached. When calculating Liutex core lines using computer programs, some numerical difficulties are raised. Because of the numerical errors, the points selected by Liutex core line definition are a cluster of points rather than a single point. To solve this problem, Xu et al. [35] proposed seed point selection method. Li et al. [36] further improved this method by choosing the best line among the candidate lines generated by seed points. In this chapter, a new method, from another perspective, is reported to calculate the Liutex core lines based on the idea that the forward Liutex magnitude gradient lines converge to the Liutex core lines. So, instead of drawing two-directional Liutex magnitude lines, we only draw the forward lines. This leads to one advantage that the converging lines are unique so that people do not need to select one line from several candidate lines. This method achieves preliminary success for identifying Liutex core lines of the Lambda vortex in early boundary layer transition.
This chapter is organized as follows: The definitions of Liutex and Liutex core lines are introduced in Section 2 and 3, respectively. In Section 4, the manual method and Xu’s algorithm are reviewed. Then, the new proposed algorithm is introduced in Section 5. In Section 6, the proposed automatic method is applied to find the Liutex core lines from the data obtained from a Direct Numerical Simulation (DNSUTA) [37] of flow transition in a boundary layer validated by researchers from UTA and NASA Langley. The results of the DNSUTA were also compared to other DNS results where the DNSUTA was shown to have remarkable consistency with [38]. Conclusions are made in Section 7.
2. Liutex
Liutex
where its direction
3. Liutex core line
The Liutex core line [34] (or the vortex rotation axis) can be defined as a Liutex line passing through the points that satisfy the condition:
where
The Liutex core concept comes from the idea that Liutex core is located at the position where Liutex reaches the maximum in the plane perpendicular to the local Liutex direction as shown in Figure 1. Since it is the local maxima in the plane, the projection of the Liutex gradient in the plane is zero, and thus, its only possible direction is perpendicular to the plane, which is parallel to the local Liutex direction. So, Eq. (3) is used to define Liutex core.
4. Manual method and Xu’s algorithm to calculate Liutex core lines
After defining the Liutex core line in [34], a manual method was proposed simultaneously in the same paper based on visualization software, for instance, Tecplot, Paraview, and so forth. After loading data files into these software, slices, iso-surfaces, and streamlines can be drawn, which are used in the manual method. The steps of the manual method are listed below.
Manual method:
Step 1: Draw the iso-surface of Liutex to get a general idea of the vortex structure.
Step 2: Set a slice that exhibits the contour of Liutex magnitude.
Step 3: Determine the concentration of line of some Liutex magnitude gradient lines passing through the slice.
Step 4: Find the intersection point of the concentration line and the slice and create a Liutex line passing through the intersection point.
Use Burgers vortex as an example. The Bugers vortex is an analytical solution of the Navier-Stokes governing equations and is a typical widely used vortex model. Its velocity field can be described in the cylindrical coordinates as follows:
where
Some streamlines of the Burgers vortex is shown in Figure 2. It can be seen that Burgers vortex is like an upside-down cyclone where the flow converges into the center and then stays in the center and is stretched. As stated in the step 1, the iso-surface of Liutex is shown in Figure 3, which shows there is a vortex center inside the iso-surface. Then, according to step 2 and step 3, a slice cutting the iso-surface is created, and gradient of Liutex magnitude lines is drawn, as shown in Figure 4. Clearly, all gradient lines converge to one point, and that point is the seed point. Next, we draw the Liutex line passing through the seed point, and this line represents the vortex center, and the result is shown in Figure 5.
The Burgers vortex is a simple vortex model. Then, a more complicated case is tested. We will use the direct numerical simulation of the flat plate boundary layer transition. Similarly, we can first draw the iso-surface of Liutex (shown in Figure 6) and find the vortex structures are much more complicated than the Burgers vortex. Next, we display Liutex gradient lines and find the intersection point of the lines and the slice as shown in Figure 7. Then, we draw the Liutex line passing through that point. In Figure 8, colors are used to represent the Liutex magnitude and we can see the rotation strengths.
Obviously, the manual method is low in efficiency and impractical for generating all Liutex core lines for complicated vortex. So, many scientists aim to find an automatic way. Li and Xu et al. proposed an algorithm to automatically calculate Liutex core lines in [36].
Xu’s algorithm:
Step 1: Read position and velocity data.
Step 2: Compute Liutex vector.
Step 3: Search vortex core points by the criterions
where
Step 4: Compute vortex core lines based on selected points in step 3.
Step 5: Group the vortex core lines.
Step 6: Delete the false and inaccurate vortex core lines by the third criterion.
Step 7: Output data of vortex core lines.
To make it simple, Xu’s algorithm indeed proposes a series of criterions on how to select the seed points. His method first selects some seed point candidates by applying some criterions. And then, it groups the seed point candidates of the same vortex together. After that, it selects the best one from the seed point candidates in each group.
5. New algorithm to calculate Liutex core line
As explained in Section 3, Liutex direction is parallel to Liutex magnitude gradient direction at vortex core positions, so theoretically, Liutex lines and Liutex gradient lines will be the same in the location where the Liutex core lines pass. However, in practice, due to numerical errors, the points that the numerical program finds will not be exactly located on the same single Liutex line, which requires selecting the best one by a specified algorithm in Xu’s paper. Another perspective to solve this problem is to make use of the property of gradient lines. Gradient lines have a good property that they can converge to the local maximums. In Figure 9, people can see that even though the seed points are much away from the vortex center, the generated Liutex gradient lines will converge to a unique line finally. So, even though selected points have numerical errors, we will still get a unique line. Based on this idea, a new algorithm is proposed as follows.
Step 1: Set the Liutex threshold and exclude the non-vortex points whose Liutex magnitudes are below the threshold.
Step 2: Find all seed points satisfying
where
Step 3: Draw forward Liutex gradient lines starting from the selected seed points.
The proposed method has the following merits. First, this method does not have a high demand on the accuracy of the seed points. Xu’s method already shows that it is hard to get enough accurate seed points and to ensure a unique line, people need to pick one seed point from seed point candidates. But the proposed method can tolerate some errors on the seed point selection. Secondly, since the proposed method can tolerate some errors on the seed point selection, the parameters in this method are not case-sensitive because people do not rely on the criterions to filter out all fake seed points.
6. Test case
In this section, the proposed automatic method is applied to find the Liutex core line for the data obtained from a Direct Numerical Simulation [6] (DNSUTA) of flow transition in a boundary layer, which has been validated by researchers from UTA and NASA Langley. The mesh of this simulation includes
Re | |||||||
---|---|---|---|---|---|---|---|
0.5 | 1000 | 300.79 | 273.15 K | 273.15 K |
This simulation uses the 3D compressible Navier–Stokes equations in curvilinear coordinates, which can be expressed as the following,
where
Compact schemes are used to do the spatial discretization. It has the following form:
where
The total variation diminishing (TVD) third order Runge–Kuntta method is used for the time discretization. The equations are:
The simulation result is shown in Figure 11. Spanwise vortex appears first followed by the Lambda vortex. We will use the Lambda vortex region to test the proposed method.
Applying step 1 and step 2 of the proposed method, we can get the seed points as shown in Figures 12–14. Drawing forward Lituex gradient lines starting from these seed points, we get Figures 15–17. If we draw Liutex lines from these seed points, we will get several separate lines rather than a unique line as shown in Figures 18–20. The proposed new algorithm obtains the unique vortex core line and does not need to select the best line from a group of lines generated by the seed points. Even though the obtained seed points have numerical errors, the output will still be convergent to a unique line. Figures 21–23 show Liutex magnitudes at the vortex core positions by colors. They show the rotation strength of the vortex core line is not uniform and is varied at different locations. This is what people cannot get by using the iso-surface methods as iso-surface is a surface where the magnitudes are equal to the threshold.
7. Conclusion
This study reports a method to automatically draw Liutex core lines from a new perspective. The proposed method is based on the property of gradient lines that they are converging toward to the local maximum points. Instead of Liutex lines, Liutex gradient lines are used in the proposed method and give this method two major merits. Firstly, the uniqueness comes from the property of gradient lines. Secondly, the proposed method can tolerate some numerical errors as these errors can easily be eliminated during the process that gradient lines converge to the center. However, this is not a mature method, and some work still needs to be done in the future, for example, developing it into a universal method that works for different cases.
Acknowledgments
The project is currently supported by NSF under Grant #2300052. The authors are thankful for the support by the UTA Department of Mathematics, which houses the UTA Vortex and Turbulence Research Team. The authors are also grateful to Texas Advanced Computing Center (TACC) for providing computation time.
Data availability
The data that supports the findings of this study are available from the corresponding author upon reasonable request.
References
- 1.
Kaczorowski M, Chong K-L, Xia K-Q. Turbulent flow in the bulk of Rayleigh–Bénard convection: Aspect-ratio dependence of the small-scale properties. Journal of Fluid Mechanics. 2014; 747 :73-102 - 2.
Xi H-D, Zhang Y-B, Hao J-T, Xia K-Q. Higher-order flow modes in turbulent Rayleigh–Bénard convection. Journal of Fluid Mechanics. 2016; 805 :31-51 - 3.
Zhou Y. Rayleigh–Taylor and Richtmyer-Meshkov instability induced flow, turbulence, and mixing. I. Physics Reports. 2017; 720 :1-136 - 4.
Zhou Y. Rayleigh–Taylor and Richtmyer–Meshkov instability induced flow, turbulence, and mixing. II. Physics Reports. 2017; 723 :1-160 - 5.
Liu C, Gao Y-S, Dong X-R, Wang Y-Q, Liu J-M, Zhang Y-N, et al. Third generation of vortex identification methods: Omega and Liutex/Rortex based systems. Journal of Hydrodynamics. 2019; 31 :205-223. DOI: 10.1007/s42241-019-0022-4 - 6.
Wang Y, Yang Y, Yang G, Liu C. DNS study on vortex and vorticity in late boundary layer transition. Communications in Computational Physics. 2017; 22 :441-459. DOI: 10.4208/cicp.OA-2016-0183 - 7.
Hunt JCR, Wray AA, Moin P. Eddies, stream, and convergence zones in turbulent flows. Center for Turbulent Research Report CTR-S88. 1988:193-208 - 8.
Chong MS, Perry AE, Cantwell BJ. A general classification of three-dimensional flow fields. Physics of Fluids A: Fluid Dynamics. 1990; 2 :765-777. DOI: 10.1063/1.857730 - 9.
Zhou J, Adrian RJ, Balachandar S, Kendall TM. Mechanisms for generating coherent packets of hairpin vortices in channel flow. Journal of Fluid Mechanics. 1999; 387 :353-396. DOI: 10.1017/s002211209900467x - 10.
Jeong J, Hussain F. On the identification of a vortex. Journal of Fluid Mechanics. 1995; 285 :69-94 - 11.
Gao Y, Liu C. Rortex and comparison with eigenvalue-based vortex identification criteria. Physics of Fluids. 2018; 30 :085107. DOI: 10.1063/1.5040112 - 12.
Cuissa JC, Steiner O. Innovative and automated method for vortex identification-I. Description of the SWIRL algorithm. Astronomy & Astrophysics. 2022; 668 :A118 - 13.
Shen C, Yang R, Wang M, He S, Qing S. Application of vortex identification methods in vertical slit fishways. Water. 2023; 15 :2053 - 14.
Xu W, Wang Y, Gao Y, Liu J, Dou H-S, Liu C. Observation on Liutex similarity in the dissipation subrange of turbulent boundary layer. Computers & Fluids. 2022; 246 :105613 - 15.
Xu W-Q, Wang Y-Q, Gao Y-S, Liu J-M, Dou H-S, Liu C. Liutex similarity in turbulent boundary layer. Journal of Hydrodynamics. 2019; 31 :1259-1262. DOI: 10.1007/s42241-019-0094-1 - 16.
Yan B, Wang Y, Liu C. Liutex-represented vortex spectrum in turbulence. Entropy. 2022; 25 :25 - 17.
Trieu XM, Liu J, Gao Y, Charkrit S, Liu C. Proper orthogonal decomposition analysis of coherent structure in a turbulent flow after a micro-vortex generator. Applied Mathematical Modelling. 2022; 104 :140-162 - 18.
Dong X, Hao C, Liu C. Correlation between vorticity, Liutex and shear in boundary layer transition. Computers & Fluids. 2022; 238 :105371 - 19.
Dong X, Hao C, Dong Y, Liu C, Li Y. Investigation of vortex motion mechanism of synthetic jet in a cross flow. AIP Advances. 2022; 12 :035045 - 20.
Liu C, Yu Y, Gao Y-S. Liutex based new fluid kinematics. Journal of Hydrodynamics. 2022; 34 :355-371 - 21.
Liu C, Liu Z. New governing equations for fluid dynamics. AIP Advances. 2021; 11 :115025. DOI: 10.1063/5.0074615 - 22.
Liu C. New ideas on governing equations of fluid dynamics. Journal of Hydrodynamics. 2021; 33 :861-866 - 23.
Yu Y, Shrestha P, Alvarez O, Nottage C, Liu C. Investigation of correlation between vorticity, Q, λci, λ2, Δ and Liutex. Computers & Fluids. 2021; 225 :104977. DOI: 10.1016/j.compfluid.2021. 104977 - 24.
Charkrit S, Shrestha P, Liu C. Liutex core line and POD analysis on hairpin vortex formation in natural flow transition. Journal of Hydrodynamics. 2020; 32 :1109-1121 - 25.
Yu Y, Shrestha P, Alvarez O, Nottage C, Liu C. Correlation analysis among vorticity, Q method and Liutex. Journal of Hydrodynamics. 2020; 32 :1207-1211. DOI: 10.1007/s42241-020-0069-2 - 26.
Dong X-R, Cai X-S, Dong Y, Liu C. POD analysis on vortical structures in MVG wake by Liutex core line identification. Journal of Hydrodynamics. 2020; 32 :497-509 - 27.
Wang Y-Q, Gao Y-S, Xu H, Dong X-R, Liu J-M, Xu W-Q, et al. Liutex theoretical system and six core elements of vortex identification. Journal of Hydrodynamics. 2020; 32 :197-211. DOI: 10.1007/s42241-020-0018-0 - 28.
Yu Y, Shrestha P, Nottage C, Liu C. Principal coordinates and principal velocity gradient tensor decomposition. Journal of Hydrodynamics. 2020; 32 :441-453. DOI: 10.1007/s42241-020-0035-z - 29.
Tran CT, Long X-P, Ji B, Liu C. Prediction of the precessing vortex core in the Francis-99 draft tube under off-design conditions by using Liutex/Rortex method. Journal of Hydrodynamics. 2020; 32 :623-628 - 30.
Wang Y, Gao Y, Liu C. Liutex and Third Generation of Vortex Identification:Workshop from Aerospace and Aeronautics World Forum 2021. Singapore: Springer Nature; 2023 - 31.
Liu C, Wang Y. Liutex and Third Generation of Vortex Definition and Identification: An Invited Workshop from Chaos 2020. Cham: Springer Cham; 2021 - 32.
Liu C, Xu H, Cai X, Gao Y. Liutex and Its Applications in Turbulence Research. London: Academic Press; 2020 - 33.
Liu C, Gao Y. Liutex-Based and Other Mathematical, Computational and Experimental Methods for Turbulence Structure. Singapore: Bentham Science Publishers; 2020 - 34.
Gao Y-S, Liu J-M, Yu Y-F, Liu C. A Liutex based definition and identification of vortex core center lines. Journal of Hydrodynamics. 2019; 31 :445-454. DOI: 10.1007/s42241-019-0048-7 - 35.
Xu H, Cai X-S, Liu C. Liutex (vortex) core definition and automatic identification for turbulence vortex structures. Journal of Hydrodynamics. 2019; 31 :857-863. DOI: 10.1007/s42241-019-0066-5 - 36.
Li H, Liu Y, Wang D, Xu H. Liutex (vortex) core and tube identification and automatic generation algorithms. Computers & Fluids. 2023; 250 :105731. DOI: 10.1016/j.compfluid.2022.105731 - 37.
Yan Y, Liu C. Shear layer stability analysis in later boundary layer transition and MVG controlled flow. 51st AIAA Aerospace Sciences Meeting including the New Horizons Forum and Aerospace Exposition. 2013 - 38.
Bake S, Meyer D, Rist U. Turbulence mechanism in Klebanoff transition: A quantitative comparison of experiment and direct numerical simulation. Journal of Fluid Mechanics. 2002; 459 :217-243