WorksPhysics and Astronomy
Ideal incompressible axisymmetric MHD: Uncovering finite-time singularities
Abstract
We provide compelling numerical evidence for the development of (potential) finite-time singularities in the three-dimensional (3D) axisymmetric, ideal, incompressible magnetohydrodynamic (IMHD) equations, in a wall-bounded cylindrical domain, starting from smooth initial data, for the velocity and magnetic fields. We demonstrate that the nature of the singularity depends crucially on the relative strength C of the velocity and magnetic fields at the time of initialisation: (i) if C < 1, then the swirl components, at the wall, evolve towards square profiles that lead to the intensification of shear at the meridional plane (r = 1, z = L/2) and the development of a finite-time singularity; (ii) if C = 1, there is no temporal evolution; (iii) if C > 1, then the swirl components, at the wall, evolve towards a cusp-type singularity. By examining the spatiotemporal evolution of the pressure, we obtain insights into the development of these singularities.
Among the many interesting problems posed by the partial differential equations (PDEs) of fluid dynamics, perhaps the most challenging is proving the global regularity of their solutions. For the three-dimensional (3D) Navier–Stokes equation (NSE), which governs the motion of a Newtonian fluid [1, 2, 3, 4, 5, 6], the global-regularity problem is one of the Clay Millennium Prize Problems [7] [8]. A proof of global regularity (or lack thereof) should show that, starting with smooth initial data for the hydrodynamical fields, solutions of the PDE do not develop (or develop) singularities in finite time. For the two-dimensional (2D) NSE and Euler equation, global regularity has been proved for analytic initial data [9]. By contrast, the proofs of the global regularity of solutions of the 3D NSE and its inviscid forerunner, the 3D Euler equation, continue to be open problems. Furthermore, the breakdown of the regularity of solutions of the 3D Euler equation has been conjectured to be important for understanding anomalous dissipation for turbulent solutions of the 3D NSE, in the limit of vanishing viscosity [10, 11].
The global regularity problem remains open in both 2D and 3D [12] even for the ideal, incompressible MHD (IMHD) equations, which govern the dynamics of an incompressible, inviscid conducting fluid with zero magnetic diffusivity and are of relevance in astrophysics, space physics, geophysics, nuclear fusion, and liquid-metal flows [13, 14, 15, 16, 17, 18, 19, 20]. The existence (or nonexistence) of a finite-time singularity (FTS) in the solutions of the IMHD equations is of importance not only from a mathematical standpoint, but also from the point-of-view of physics [21]. Numerical studies of 3D incompressible MHD turbulence have shown that, at high Reynolds numbers and high magnetic Reynolds numbers, the rate of energy dissipation reaches a positive constant value [22, 23, 24]; this phenomenon is reminiscent of anomalous dissipation at high Reynolds numbers in 3D NS turbulence. Consequently, it has been conjectured that the breakdown of global regularity for the 3D IMHD equations might be related to anomalous dissipation in MHD turbulence [25].
Numerical investigations of both the 3D incompressible Euler [26, 27] and the 3D IMHD equations [21, 28, 29], in spatially periodic domains, have not yielded conclusive evidence for or against an FTS. By contrast, there is growing numerical evidence for a potential FTS in the 3D-axisymmetric incompressible Euler (3DAE) equations, in a wall-bounded cylindrical domain. This was first reported in the adaptive-mesh calculation of Ref. [30] which reported the development of a ring-like singularity on the wall of the cylinder in the meridional plane at an estimated singularity time . This potential FTS was re-examined and confirmed using Fourier-Chebyshev pseudospectral [31, 32] and Cauchy-Lagrange [33] methods. The role of pressure in the development of this FTS was elucidated in Ref. [31].
In this paper, we provide compelling numerical evidence for a similar FTS for the 3D-axisymmetric IMHD equations in a wall-bounded domain, with smooth initial data for both velocity and magnetic fields [akin to the initial condition in Refs. [30, 32] for the 3DAE]. Furthermore, we uncover a new type of FTS by tuning the ratio of the magnitudes of the initial magnetic and velocity fields. If , the evolution of the FTS is akin to that in the 3DAE [30, 32, 33]. In contrast, if , then the swirl-components of the fields develop cusp-like structures at the wall that lead, in turn, to a potential FTS. This novel cusp-type singularity has not been seen for any axisymmetric hydrodynamical partial differential equation [34] For both cases and , the vorticity and current fields show rapid growth.
The evolution of an incompressible, inviscid, perfectly conducting fluid is governed by the incompressible, ideal MHD equations (IMHD) below :
In cylindrical coordinates, the velocity field is written as . Here, and .; there is a similar expression for the magnetic field ; the superscripts , , and denote the radial, swirl, and axial components of the fields.
We solve the IMHD equations in an axisymmetric cylindrical domain: in the -direction, we use periodic boundary conditions for all fields and in the direction, we impose no-flow condition for , and the perfectly conducting condition for , at the wall, i.e., ; (ii) at the axis, we use the pole conditions for all fields, namely, that their radial derivatives vanish at [32, 35].
We initialize the velocity field with as in Refs. [30, 32]; and the initial magnetic field is chosen to be a multiple of the initial velocity where controls the strength of relative to , at the initial time.
We first analyse the IMHD flows evolved from the aforementioned initial data using the Caflisch-Klapper-Steele (CKS) criterion [25] in Figure 1(a). The rapid growth of , the integrand in the CKS criterion (10), indicates the generation of FTSs for various values of . However, since we employ a fixed-grid spectral scheme, we do not rely completely on the CKS criterion for singularity detection. We now carry out a detailed study of these potential FTSs using the analyticity-strip method of Ref. [36] which is especially well-suited for spectral schemes.

Figure 1. (a) Plot vs. of the CKS criterion for different values of given in the legend. We see rapid growth of the CKS integrand in time, which occurs earlier for values of far from . (b) Semi-log plots versus of the Fourier energy spectra (red) and (blue) at the wall (at ) for four different times . For early times, these spectra (shown in faded colours) have exponential tails, with large negative slopes, so our system exhibits spectral convergence. At later times, the spectra flatten out and begin to thermalise (see text); the flow is initiated with the representative value and a resolution of .
In order to track the nearest complex-space singularities using the analyticity-strip method, we consider the Fourier spectra and of the velocity and magnetic fields computed at the wall , where we expect the potential FTS to precipitate, in Figure 1(b).
At large-, the spectra exhibit exponentially decaying tails, so we use the asymptotic relation (11) to extract the slope of the semi-log plot in Figure 1(b), which yields our estimate for the distance of the nearest complex singularity from the real line, at time [37]. As time progresses, we find that , i.e., the complex singularity approaches the real line, and a potential FTS develops.
Clearly the rate of decay of depends on . In Figure 2(a), we present log-log plots of versus ; the data from our numerical simulations are shown by filled circles; the linear regimes in these plots are consistent with the power-law forms , which we show via solid lines. These plots provide strong evidence for FTSs in the 3D-axisymmetric IMHD model for the flows that we initialise with . We display the -dependence of the singularity time in Figure 2(b) [38]. The case is special because the kinetic and magnetic terms on the RHS of Eqs. (6) cancel each other, so the solution remains stationary and there is no singularity; decreases as moves away from .
Figure 2. Plots versus of (a) the analyticity-strip width (and in the inset) obtained by performing the nonlinear fit (11) of the total energy Fourier spectrum at the wall () for flows initiated with different choices of in the initial data. (b) Log-log plot versus of and the corresponding power-law fit for flows initiated with different choices of in the initial data. (c) Plot versus of the estimate for the time of singularity that we obtain from the power-law fit of from (a); we use (, ).
Here, we note that there are a finite number of Fourier-Chebyshev modes in any practical numerical simulation, so there is energy-backflow from high- modes, near the dealiasing cut-off , and our Galerkin-truncated system moves towards thermalisation [39, 40], which is reflected in the flattening of and in Figure 1(b) at late times . In physical space, this leads to the emergence of localised, oscillatory structures in the kinetic and magnetic fields, as we discuss below). These structures are referred to as tygers [32, 40, 41]; they are precursors to thermalisation [40] and often appear just prior to an FTS [32, 40].
To understand the development of the singularity in physical space, we now examine the streamlines and secondary flows of the velocity and magnetic fields, for two illustrative values of and ; henceforth, we refer to the resulting flows as Flow-A and Flow-B, respectively. The observations that we make for Flow-A (Flow-B), hold true for flows initialised with other choices for , with ().
Profile at wall: In Flow-A, the velocity fields are initially stronger than their magnetic counterparts. In Figure 3(a), we observe that at early times (), has a nearly sinusoidal profile at the wall, which develops into a square profile at later times (), . This in turn results in the intensification of shear at the meridional plane () and tygers appear near at this time. [For similar plots of other fields, see Figure F2 in the Supplementary Information.] In Flow-B [] the magnetic field is initially stronger than the velocity field. In Figure 3(b) the plots of at the wall (), for various representative times, show that the initial sinusoidal profile of (yellow line) evolves into a tent-like profile (pink line) at intermediate times, losing differentiability at . At later times [e.g., ], this is damped (blue line) as the flow decelerates along the wall, and we see the development of cusps in at , which are distinctly different from the square-type profile in Figure 3(a) for Flow-A. Tygers appear too, but away from the positions of the cusps.

Figure 3. Plot versus of the swirl velocity superposed for various times listed in the legend, for flows initiated with (a) and (b) . Here, we use a resolution of (). At later times, we see the development of localized oscillatory structures called tygers in these fields. See the Supplementary Information for similar plots of .

Figure F2. Plot versus of (a,b) , (c,d) and (e,f) at , superposed for various times listed in the legend. The flows have been initiated with (a,c,e) and (b,d,f) . Here, we use a resolution of . At later times, we see the development of localized oscillatory structures called tygers in these fields.
Secondary flows and streamlines: In order to understand the spatiotemporal evolution of at the wall described above, we now study the secondary flows in the plane. We present , for Flow-A, (i) plots of the streamlines for the velocity field , in 3D [Figure 4(a)] and (ii) plots of the streamlines of the secondary flow in the plane [blue-green solid lines in Figure 4(b)], superposed on the heat-map of for time . The secondary flow consists of the fluid rotating, within convection-type cells, along the wall; the direction of rotation of the cells is such that the fluid is driven away from along the wall; thereby , the initial maxima of at these locations are similarly advected outwards by .

Figure 4. Flow-A: (a) Plot of the streamlines of the axisymmetric velocity field seeded from a straight line (shown on the left side, inside cylinder) at time . We plot the heat-map of on the back-face of the cylinder. At early times, the streamlines (blue) are pushed outward towards the wall by the secondary flows; at later integration times, they flatten on the wall, leading to the square-type profile of shown in Figure 3(a). (b) Plots of streamlines (blue-green) of the secondary flows driven by in the plane, superposed over the heat-map of at the same time . We initiate the flow with and use for this run.
In Figure 4(a), we see that the streamlines, which have been started out along a straight line (blue), move into (out of) the plane given the negative (positive) sign of in the upper (lower) half of the cylinder. They bulge outwards towards the wall, because of the influence of the secondary flows, near . At later times, the streamlines (yellow), which have started from near , fan out at the wall; this is nothing but the advection of outwards from . For this time, we show the heat-map of on the back-face of the cylinder (colorbar of ) from which we can deduce the square-profile of at late times [cf. Figure 3(a)]; this leads gradually to uniform blue-green (orange) colouring of the heat-map in the upper (lower) half of the cylinder.
For flows initiated with , real-space dynamics of the velocity field closely follow the dynamics summarised above for Flow-A with ; however, as increases, decreases. [Figure 2(b)]. Importantly, the mechanism that we report here for Flow-A is akin to that seen for the potentially singular flow in the 3D-axisymmetric wall-bounded Euler equation [30, 31, 32] ( here). Furthermore, the spatiotemporal evolution of the magnetic field is similar to that of ; in particular, we see the development of a square-profile of at the wall [see Figure F2 in the Supplemental Information].
For Flow-B, we now discuss the secondary flows in the plane that lead to the formation of the tent-like profile of in Figure 3. As for Flow-A, we visualise (i) the streamlines for the axisymmetric velocity field [Figure 5 (a)] and (ii) the streamlines of the secondary flow in the plane [Figure 5 (b)], at time .

Figure 5. Flow-B: (a) Plot of the streamlines of the axisymmetric velocity field seeded from a straight line (shown on the left side, inside cylinder) at time . We plot the heat-map of on the back-face of the cylinder. The streamlines (red) are twisted and pulled inward by the secondary flows at , close to the singularity time and this leads to the cusp-like profile of shown in Figure 3(b). (b) Plots of streamlines (red) of the secondary flows driven by in the plane, superposed over the heat-map of at the same time . We initiate the flow with and use for this run.
The red streamlines [Figure 5 (b)] of the secondary flow in the plane, superposed on the heat-map of , show rotating cells of fluid, similar to those for Flow-A [Figure 4(b)], but with the opposite direction of rotation. Here, the secondary flow is such that the fluid is driven towards on the wall. This leads to the pinching of the initially sinusoidal profile of , because of its advection by , which now points inward at in these cells.
The deceleration of the flow along the wall can be seen clearly in the streamline plot [Figure 5(a)], which is seeded as in Figure 4 (a). Here, Flow-B is distinctly different from Flow-A: the sheet formed by the streamlines in each half of the cylinder, folds in on itself at (dark red) because of the advection of the slower-moving fluid inward, at these points, by the secondary flow. Thus, at later times (yellow), the streamlines seeded near move towards in the upper half of the cylinder; we see similar dynamics in the lower half where the fluid rotates with positive swirl velocity. Correspondingly, the heat-map of , shown on the back-face of the cylinder, displays an intensification of velocity solely at [in contrast with the heat-map of Figure 4(a), where the flow is accelerated everywhere along the wall]. Thus, our examination of secondary flows elucidates the magnetic-field-induced suppression of the swirl velocity at the wall and the formation of cusps in Flow-B.
Role of pressure: We now demonstrate that the pressure plays a crucial role in establishing the secondary flows discussed above, which in turn lead to the formation of FTS in the IMHD flows. We obtain the pressure from the divergence of Eq. (1a) that yields the Poisson-type equation (9). For our IMHD flows, we define the effective pressure , which we compute for Flow-A and Flow-B using the Fourier-Chebyshev Tau Poisson solver developed by us [32].
In Figure 6(a) and (b), we show, respectively, the heat-maps of , computed at the early time close to the wall, in the lower half of the cylinder, for (a) Flow-A and (b) Flow-B; we superpose the streamlines of the corresponding secondary flows.

Figure 6. Streamline plots of the secondary flows in the – half-plane superposed on the heat-map of the effective pressure for (a) Flow-A and (b) Flow-B at time . We see that the flow moves outward from the pressure maximum at along the wall for Flow-A. For Flow-B, the flow moves into the pressure minimum at the same point.
From Figure 6(a) we can deduce that, for Flow-A, the pressure is lowest at the axis and it increases as ; along the wall , a local pressure maximum develops at . As a result, we see that the fluid moves along the pressure gradient on the wall, i.e., it moves outwards from the pressure maximum at in the -direction; and, in the -direction, it moves away from the wall at . At , the fluid is forced to move against the local pressure gradient by virtue of incompressibility, thereby enforcing the direction of rotation of the fluid in the cells. Thus, the secondary flow rotates anti-clockwise for and clockwise for .
In contrast, for Flow-B [Figure 6(b)], the effective pressure is lower at the wall and increases as ; now, there is a local minimum of at [cf. the maximum of for Flow-A in Figure 6(a)]. As a result, the fluid moves towards along the wall, and then it moves outward radially, towards the wall at .
We can relate the pressure field, which develops at initial times, to the initial data by considering the radial-pressure balance at (on the wall), as shown in Eq. (3). In order to obtain the radial force-balance, we consider the -component of the 3DA-IMHD equations (6) and we assume that the flow is instantaneously stationary right before initialisation; thus, the LHS vanishes and we get, along :
We find that the gradient of has to balance the centrifugal force that points outward along , minus the curvature forces because of the tension of the magnetic field lines , inward along , at the wall. For Flow-A, and ; thus, the centrifugal force is dominant, giving a resultant force that points outwards along . In order to achieve force-balance, is set up in a way such that its radial gradient points outward along to balance the resultant force above [note the sign in in Eq. (3)]. This is verified by the heat-map of shown in Figure 6 (a), with increasing as ; the pressure maximum occurs at as and assumes its largest value here.
By contrast, for Flow-B where , the magnetic-tension forces dominate over the centripetal forces, and the resultant force points radially inwards. The pressure field is now such that its radial gradient points inwards. This is seen in Figure 6 (b), where the pressure decreases as and the local minimum occurs at , where the attains its largest value.
In conclusion, the global-regularity of solutions of 3D hydrodynamical PDEs, such as the 3D IMHD equations, continues to challenge physicists, fluid dynamicists, and mathematicians. Numerical studies play an important role in the quest for FTSs in such PDEs. We have taken the first crucial step that is required for uncovering FTSs in the 3D IMHD equations: In particular, we have presented compelling numerical evidence for the development of (potential) finite-time singularities in these equations in a wall-bounded axisymmetric domain. Furthermore, we have obtained deep insights into these FTSs by examining secondary flows and their dependence on the effective pressure at initial times .
Our study brings out clearly the importance of emergent secondary flows, arising from differential rotation, and the role played by the effective pressure in their development. Had the system been a viscous and resistive fluid, then the boundaries would have led to Ekman pumping, well-studied in the contexts of magnetic stirring of liquid metals [18, 42] and the persistence of geodynamos [43, 44]. In the potential FTS flows that we consider [45], the differential rotation arises by virtue of the initial data.
The singular solutions we uncover are of two types – either they display developing discontinuities or cusp– depending on the relative magnitudes of the kinetic and magnetic fields at the initial time; the cusp-type singularities have not been anticipated in the types of PDEs we study. Both these types of singularities will be of relevance and interest to plasma-physicists, fluid-dynamicists, and mathematicians. In particular, the potential FTSs we find should aid in the development of rigorous proofs for the global-regularity (or lack thereof) of solutions to the 3D IMHD PDEs, with the initial data we have proposed.
Additionally , we note that our work reports the first evidence for tyger-type oscillations [32, 40] in a Galerkin-truncated pseudospectral study of the 3D IMHD PDEs. Thus, our work has implications for the general problem of thermalisation in spectrally truncated ideal hydrodynamical systems. The tygers we have found are the precursors of thermalisation in the Galerkin-truncated 3D IMHD system; and they generalise considerably analogous results for ideal-fluid hydrodynamical equations [32, 39, 40, 46, 47].
Methods
Axisymmetric formulation of the 3D IMHD equations
The pressure field in the IMHD equations (1) can be eliminated using the vorticity and the current density and rewriting the above equations in the following vorticity and current formulation:
In 3D, the stream function and magnetic potential are vector quantities that are related to the primitive variables in Eqs. (1a)-(4) via and . We represent vector fields in the 3D axisymmetric domain as follows: , where , and are the basis vectors in the cylindrical coordinate system. We introduce the transformed variables [30, 32]
and we rewrite eq:1 as follows:
The Poisson equations for and are given below:
The radial and axial components of the and fields are obtained from and , respectively:
The solution of the above equations, in terms of the transformed variables [see Eq. (6)], ensures that the corresponding fields remain smooth at the coordinate singularity on the axis of the cylinder (). We obtain the pressure by taking divergence of Eq. (1a), which yields the following Poisson-type equation:
Numerical method
We use Fourier-Chebyshev pseudospectral methods to solve Eqs. (6). The domain is discretised using the collocation grid . In the periodic direction, the collocation points are distributed uniformly. In the wall-bounded direction, the collocation nodes are the roots of , the order- Chebyshev polynomial of the first kind; these roots are spaced non-uniformly and they cluster near and . Given a function , its Fourier-Chebyshev approximation on the grid is
If , the spacing between Chebyshev nodes, near and , is much smaller than the spacing between the Fourier nodes. Therefore, we use ; specifically, in our direct numerical simulation (DNS), we use and , which suffice to uncover the FTSs we explore and the associated tyger oscillations. We use a fourth-order Runge-Kutta scheme, with an adaptive time step that ensures Courant–Friedrichs–Lewy (CFL) stability in both the Fourier-discretised -direction and the Chebyshev-discretised -direction. We compute the derivatives in spectral space and, subsequently, the nonlinear terms in physical space. In order to remove aliasing errors, in both Fourier and Chebyshev spectral approximations, we use the 2/3-dealiasing method [48]. We use our Tau Poisson solver [32] to solve the Poisson equations (11) and (12) with the boundary conditions. We use the Tau Poisson solver for solving the pressure-Poisson problem (9).
Singularity-detection criteria
We use two criteria to check for FTSs in solutions of the 3D-axisymmetric IMHD equations (6): (a) the Caflisch-Klapper-Steele (CKS) criterion and (b) the analyticity-strip method [49]; we discuss the details below.
(a) Caflisch-Klapper-Steele (CKS) criterion: The Beale-Kato-Majda (BKM) criterion is regularly employed to look for FTSs for the 3D Euler equations [50]. It was generalised to ideal MHD flows by Caflisch, Klapper, and Steele [25] as follows:
A solution of the 3D IMHD equations remains smooth in the time interval provided that
For our spectral simulations, we define the global maximum (resp. ) as the maximal value of (resp. ) on the computational grid . In order to use this criterion effectively, we must capture singular structures accurately, e.g., by using an adaptive-meshing scheme (see, e.g., Refs. [30, 51, 52]). For our fixed-grid Fourier-Chebyshev pseudospectral scheme, we employ the analyticity-strip method for it requires the spectral approximations of the fields; these are readily available to us from our DNS.
(b) Analyticity-strip method: This method was developed in Ref. [49] for tracking complex-space singularities when Fourier pseudospectral schemes are used [21, 32, 49, 53, 54, 55, 56, 57, 58]. At a given instant of time , the real-space solution , of the PDE, can be analytically continued to complex-space variables , inside the analyticity-strip: . If at this time, the complex singularity, nearest to the real axis , is present at , then the Fourier spectrum of the velocity field behaves as
here is the exponent of the singularity and . The presence of multiple competing singularities close to gives rise to oscillatory behavior in the large- regime. In Ref. [32], we have extended the application of this method to pseudospectral studies in bounded domains, where we use Chebyshev discretisation in the bounded direction. Here, we employ our analyticity-strip method [32] to analyse the Fourier-Chebyshev spectra of the velocity and magnetic fields obtained from our DNS of the 3D-axisymmetric IMHD model (6). For our analysis, we define the Fourier-Chebyshev spectra , , and , for the fluid kinetic energy, the magnetic energy, and the total energy , respectively, as follows:
here, and . At a given time and radial coordinate , we perform the nonlinear fit of the total energy spectrum using the form given in Eq. (18); we can then obtain from the slope of the exponential tail in the large- regime. If at some time , , then the nearest complex singularity hits the real axis, at this time, and the flow develops an FTS in physical space.
Any practical implementation of a pseudospectral DNS has a finite spatial resolution, so the extraction of can only be accurate up until the time at which the spectral approximation is relatively free from errors [41]; beyond this time, spectral convergence is lost, so we stop our analysis. For the 3DAE [32] and for the 3D-axisymmetric IMHD flows that we discuss here, we see the emergence of localised oscillatory structures, called tygers [40], around this time; at later times, the energy spectra show signs of thermalization [32, 39, 46].
Conserved quantities
The 3D IMHD equations (1) have three quadratic invariants; these are the total (kinetic + magnetic) energy , magnetic helicity , and the cross helicity . For the 3D-axisymmetric system in Eq. (6), these are given below:
In Figure F1 in the Supplementary Information, we plot versus time , the numerical conservation of the invariants of Eq. (13) achieved using our Fourier-Chebyshev pseudospectral DNS for different values of specified in the legend. In panel (a), we plot the percentage relative error in total energy with respect to ; thus, total energy is conserved with an accuracy of for early times, until ; the accuracy is gradually lost as the system nears the potential FTS. In panel (b), we plot the magnetic helicity in Eq. (13b), and in panel (c), the cross helicity in Eq. (13c), with time . We observe that conservation of magnetic helicity is extremely good; however, the numerical conservation of cross helicity deteriorates for .

Figure F1. Plots versus of (a) the percentage change in the total energy , (b) the magnetic helicity , (c) the cross helicity , and (d) the CKS integrand , for flows initialised using different choices of in the initial data; these values of are specified in the legend.
Acknowledgments
We thank T. Hertel, N. Besse, J.K. Alageshan, M. Brachet and T. Matsumoto for useful discussions; and we thank the Anusandhan National Research Foundation (ANRF), the Science and Engineering Research Board (SERB), and the National Supercomputing Mission (NSM), India, for support, and the Supercomputer Education and Research Centre (IISc), for computational resources.
References
Potential finite-time singularities of the 3D-axisymmetric ideal, incompressible, magnetohydrodynamical equations
Sai Swetha Venkata Kolluru; Rahul Pandit
This Supplemental Information contains information about (a) the numerical conservation of the invariants of the 3D axisymmetric Ideal Magnetohydrodynamics (IMHD) equation [ in the main paper], (b) computation of magnetic helicity and (c) extra figures that show the evolution of , and at the wall at .
Conservation of invariants
Computation of magnetic helicity
Magnetic helicity is given by:
where is the vector-valued magnetic potential defined as .
In order to determine , we first write the components of in terms of those of as follows:
Note that we compute all components of while solving Eq. (6) and (14). From Eq. (8b) and associated boundary conditions ( and ), we solve for using the Tau Poisson solver.
We can determine and from in Eq. (S2a). Additionally, we use the Coulomb Gauge , since the cylinder wall is considered to be a magnetic surface and since there is no mean-field in the periodic -direction [1]:
We can reduce the above system of equations to a Poisson equation for given below:
We can re-cast this equation in terms of as follows:
The above equation when supplemented with appropriate boundary conditions is a well-posed Poisson type problem which can be solved using a Tau Poisson solver akin to the one used for the solution of Eq.(8b). Once we have computed , we use Eq.(6c) to compute as follows:
This equation can be inverted in Fourier-Chebyshev space; we require one boundary condition for this problem to be well-posed.
Boundary conditions
We now obtain boundary conditions for and at and . Again, we use the Coulomb gauge :
For vector potential , we have from Refs. [2, 3]:
Once again using the Coulomb gauge (S5a) and Eq. (S6a), we get:
Thus, we solve the following problems for and . This allows us to compute the magnetic helicity in Eq. (13b), shown in panel (b) of Figure F1.
A: For :
B: For :
Extra figures
References
References
- [1]C. R. Doering and J. D. Gibbon, Applied analysis of the Navier-Stokes equations. Cambridge university press, 1995.
- [2]G. P. Galdi, “An introduction to the navier-stokes initial-boundary value problem,” in Fundamental directions in mathematical fluid mechanics, pp. 1–70, Springer, 2000.DOI
- [3]C. Foias, O. Manley, R. Rosa, and R. Temam, Navier-Stokes equations and turbulence, vol. 83. Cambridge University Press, 2001.
- [4]J. C. Robinson, J. L. Rodrigo, and W. Sadowski, The three-dimensional Navier–Stokes equations: Classical theory, vol. 157. Cambridge university press, 2016.
- [5]P. G. Lemarié-Rieusset, The Navier–Stokes problem in the 21st century. Chapman and Hall/CRC, 2018.
- [6]J. C. Robinson, “The navier-stokes regularity problem,” Philosophical Transactions of the Royal Society A, vol. 378, no. 2174, p. 20190526, 2020.DOI
- [7]J. A. Carlson, A. Jaffe, and A. Wiles, The millennium prize problems. American Mathematical Soc., 2006.
- [8]To quote from Ref. [7]: “Show that the Navier–Stokes equations on Euclidean 3-space have a unique, smooth, finite energy solution for all time greater than or equal to zero, given smooth, divergence-free, initial conditions which “decay rapidly at large distances.” Alternatively, show that there is no such solution.
- [9]V. I. Yudovich, “Non-stationary flow of an ideal incompressible liquid,” USSR Computational Mathematics and Mathematical Physics, vol. 3, no. 6, pp. 1407–1456, 1963.DOI
- [10]L. Onsager, “Statistical hydrodynamics,” Nuovo Cimento, vol. 6 (Suppl.), p. 279–287, 1949.
- [11]G. L. Eyink and K. R. Sreenivasan, “Onsager and the theory of hydrodynamic turbulence,” Reviews of modern physics, vol. 78, no. 1, p. 87, 2006.
- [12]M. Paicu and N. Zhu, “Global regularity for the 2d mhd and tropical climate model with horizontal dissipation,” Journal of Nonlinear Science, vol. 31, no. 6, p. 99, 2021.
- [13]D. Biskamp, Magnetohydrodynamic turbulence. Cambridge University Press, 2003.
- [14]A. R. Choudhuri, The physics of fluids and plasmas: an introduction for astrophysicists. Cambridge University Press, 1998.
- [15]J. H. Goedbloed and S. Poedts, Principles of magnetohydrodynamics: with applications to laboratory and astrophysical plasmas. Cambridge university press, 2004.
- [16]J. P. Freidberg, ideal MHD. Cambridge University Press, 2014.
- [17]S. Galtier, Introduction to modern magnetohydrodynamics. Cambridge University Press, 2016.
- [18]P. A. Davidson, Introduction to magnetohydrodynamics, vol. 55. Cambridge university press, 2017.
- [19]D. A. Gurnett and A. Bhattacharjee, Introduction to plasma physics: With space, laboratory and astrophysical applications. Cambridge University Press, 2017.
- [20]J. P. Goedbloed, H. Goedbloed, R. Keppens, and S. Poedts, Magnetohydrodynamics: of laboratory and astrophysical plasmas. Cambridge University Press, 2019.
- [21]M. E. Brachet, M. Bustamante, G. Krstulovic, P. D. Mininni, A. Pouquet, and D. Rosenberg, “Ideal evolution of magnetohydrodynamic turbulence when imposing taylor-green symmetries,” Physical review E, vol. 87, no. 1, p. 013110, 2013.
- [22]V. Dallas and A. Alexakis, “The signature of initial conditions on magnetohydrodynamic turbulence,” The Astrophysical Journal Letters, vol. 788, no. 2, p. L36, 2014.
- [23]P. D. Mininni and A. Pouquet, “Finite dissipation and intermittency in magnetohydrodynamics,” Physical Review E—Statistical, Nonlinear, and Soft Matter Physics, vol. 80, no. 2, p. 025401, 2009.
- [24]M. F. Linkmann, A. Berera, W. D. McComb, and M. E. McKay, “Nonuniversality and finite dissipation in decaying magnetohydrodynamic turbulence,” Physical review letters, vol. 114, no. 23, p. 235001, 2015.
- [25]R. E. Caflisch, I. Klapper, and G. Steele, “Remarks on singularities, dimension and energy dissipation for ideal hydrodynamics and mhd,” Communications in Mathematical Physics, vol. 184, no. 2, pp. 443–455, 1997.
- [26]J. Gibbon, “The three-dimensional euler equations: how much do we know?,” Physica D: Nonlinear Phenomena, vol. 237, no. 14-17, pp. 1894–1904, 2008.
- [27]T. Y. Hou, “Blow-up or no blow-up? a unified computational and analytic approach to 3d incompressible euler and navier-stokes equations,” Acta Numerica, vol. 18, pp. 277–346, 2009.
- [28]R. M. Kerr and A. Brandenburg, “Evidence for a singularity in ideal magnetohydrodynamics: implications for fast reconnection,” Physical review letters, vol. 83, no. 6, p. 1155, 1999.DOI
- [29]R. Grauer and C. Marliani, “Current-sheet formation in 3d ideal incompressible magnetohydrodynamics,” Physical Review Letters, vol. 84, no. 21, p. 4850, 2000.
- [30]G. Luo and T. Y. Hou, “Potentially singular solutions of the 3D axisymmetric Euler equations,” Proceedings of the National Academy of Sciences, vol. 111, no. 36, pp. 12968–12973, 2014.
- [31]D. Barkley, “A fluid mechanic’s analysis of the teacup singularity,” Proceedings of the royal society of london. Series A, Containing papers of a mathematical and physical character, vol. 476, 2020.
- [32]S. S. V. Kolluru, P. Sharma, and R. Pandit, “Insights from a pseudospectral study of a potentially singular solution of the three-dimensional axisymmetric incompressible euler equation,” Physical Review E, vol. 105, no. 6, p. 065107, 2022.
- [33]T. Hertel, N. Besse, and U. Frisch, “The cauchy-lagrange method for 3d-axisymmetric wall-bounded and potentially singular incompressible euler flows,” Journal of Computational Physics, vol. 449, p. 110758, 2022.
- [34]Cusp singularities have been attributed to interfaces between different fluids, as discussed, e.g., in Refs. [59, 60].
- [35]T. Y. Hou and R. Li, “Blowup or no blowup? the interplay between theory and numerics,” Physica D: Nonlinear Phenomena, vol. 237, no. 14-17, pp. 1937–1944, 2008.
- [36]C. Sulem, P.-L. Sulem, and H. Frisch, “Tracing complex singularities with spectral methods,” Journal of Computational Physics, vol. 50, no. 1, pp. 138–161, 1983.
- [37]We have used the LMFIT package from Python to perform the nonlinear fitting.
- [38]From the fit window, we exclude the range for which δ < Δz, where Δz is the smallest grid spacing in X₍₂₅₆,₅₁₂₎.
- [39]P. C. Di Leoni, P. D. Minnini, and M. E. Brachet, “Dynamics of partially thermalized solutions of the burgers equation,” Physical Review Fluids, vol. 3, no. 1, p. 014603, 2018.
- [40]S. S. Ray, U. Frisch, S. Nazarenko, and T. Matsumoto, “Resonance phenomenon for the galerkin-truncated burgers and euler equations,” Phys. Rev. E, vol. 84, p. 016301, Jul 2011.
- [41]S. S. V. Kolluru, N. Besse, and R. Pandit, “Novel spectral methods for shock capturing and the removal of tygers in computational fluid dynamics,” Journal of Computational Physics, vol. 519, p. 113446, 2024.
- [42]P. Davidson, D. Kinnear, R. Lingwood, D. Short, and X. He, “The role of ekman pumping and the dominance of swirl in confined flows driven by lorentz forces,” European Journal of Mechanics-B/Fluids, vol. 18, no. 4, pp. 693–711, 1999.
- [43]A. Ojha, M. Anas, A. Ranjan, P. Joshi, and M. K. Verma, “Helicity segregation by ekman pumping in laminar rotating flows with gravity orthogonal to rotation,” Physical Review Fluids, vol. 7, no. 3, p. 034801, 2022.
- [44]A. Ranjan and P. Davidson, “The spatial segregation of kinetic helicity in geodynamo simulations,” Geophysical Monograph Series, p. 241, 2024.
- [45]For the 3DAE cf. Refs. [30–32].
- [46]C. Cichowlas, P. Bonaiti, F. Debbasch, and M. Brachet, “Effective dissipation and turbulence in spectrally truncated euler flows,” Physical review letters, vol. 95, no. 26, p. 264502, 2005.
- [47]S. D. Murugan, U. Frisch, S. Nazarenko, N. Besse, and S. S. Ray, “Suppressing thermalization and constructing weak solutions in truncated inviscid equations of hydrodynamics: Lessons from the burgers equation,” Physical Review Research, vol. 2, no. 3, p. 033202, 2020.
- [48]M. Uhlmann, The need for de-aliasing in a Chebyshev pseudo-spectral method. PIK, 2000.
- [49]C. Sulem, P.-L. Sulem, and H. Frisch, “Tracing complex singularities with spectral methods,” Journal of Computational Physics, vol. 50, no. 1, pp. 138–161, 1983.
- [50]J. T. Beale, T. Kato, and A. Majda, “Remarks on the breakdown of smooth solutions for the 3-d euler equations,” Communications in Mathematical Physics, vol. 94, no. 1, pp. 61–66, 1984.
- [51]R. Grauer and C. Marliani, “Geometry of singular structures in magnetohydrodynamic flows,” Physics of Plasmas, vol. 5, no. 7, pp. 2544–2552, 1998.
- [52]R. Grauer, C. Marliani, and K. Germaschewski, “Adaptive mesh refinement for singular solutions of the incompressible euler equations,” Phys. Rev. Lett., vol. 80, pp. 4177–4180, May 1998.
- [53]M. E. Brachet, D. I. Meiron, S. A. Orszag, B. Nickel, R. H. Morf, and U. Frisch, “Small-scale structure of the taylor–green vortex,” Journal of Fluid Mechanics, vol. 130, pp. 411–452, 1983.
- [54]S. Kida, “Study of complex singularities by filtered spectral method,” Journal of the Physical Society of Japan, vol. 55, no. 5, pp. 1542–1555, 1986.
- [55]M. Brachet, M. Meneguzzi, A. Vincent, H. Politano, and P. Sulem, “Numerical evidence of smooth self-similar dynamics and possibility of subsequent collapse for three-dimensional ideal flows,” Physics of Fluids A: Fluid Dynamics, vol. 4, no. 12, pp. 2845–2854, 1992.
- [56]U. Frisch, T. Matsumoto, and J. Bec, “Singularities of euler flow? not out of the blue!,” Journal of statistical physics, vol. 113, no. 5-6, pp. 761–781, 2003.
- [57]M. D. Bustamante and M. Brachet, “Interplay between the beale-kato-majda theorem and the analyticity-strip method to investigate numerically the incompressible euler singularity problem,” Phys. Rev. E, vol. 86, p. 066302, Dec 2012.
- [58]C. Cichowlas and M. Brachet, “Evolution of complex singularities in kida–pelz and taylor–green inviscid flows,” Fluid Dynamics Research, vol. 36, no. 4-6, p. 239, 2005.
