betaFeedback

WorksPhysics and Astronomy

Optimizing helical disc dynamo

Abstract

We present an optimized design of our recently realized helical disc dynamo. Like the original set-up, the optimized dynamo consists of a flat multi-arm spiral coil and a co-axially placed disc which is connected to the former by sliding liquid metal contacts. In contrast to the original set-up, the disc and the coil in the optimized design have different sizes. This allows the disc to capture more of the high-density magnetic flux generated in the inner part of the coil and to avoid the reverse flux in the outer part of the coil. By optimizing the coil and dics radii, the critical magnetic Reynolds number can be reduced from Rm≈34.6{\mathit Rm}\approx34.6 when the disc and coil have equal inner and outer radii with the ratio ri/ro≈0.36r_{i}/r_{o}\approx0.36 to Rm≈11.6.{\mathit Rm}\approx11.6. This lowest possible disc dynamo threshold is attained when the disc and coil have relatively narrow widths. Using a slightly suboptimal but more practical set-up with the inner and outer radii of the disc and coil equal to to (0.3,0.9)(0.3,0.9) and (0.74,1),(0.74,1), respectively, self-excitation is expected at Rm≈14.6.{\mathit Rm}\approx14.6.

Introduction

The disc dynamo is a very simple physical model that is often used to illustrate the self-excitation of the magnetic field by moving conductors [10]. This is known as the dynamo action, which is currently the most likely explanation for the origin of the magnetic fields of the Earth, the Sun, and other cosmic bodies [5]. The basic disc dynamo model, which is due to Bullard [6], consists of a solid metal disc that rotates about its central axis, and a wire that is twisted around and connected through sliding contacts to the rim and the axis of the disc. If the disc spins sufficiently fast and in the right direction, such a set-up can amplify electric current circulating in the system and, thus, the associated magnetic field. This happens when the rotation rate of the disc exceeds a certain critical threshold above which the potential difference induced across the disc exceeds the voltage drop caused by the ohmic resistance of the system. Then the current starts to grow exponentially in time, resulting in the self-excitation of the magnetic field. The growth stops when the braking electromagnetic torque becomes so strong that it slows the disc.

Bullard dynamo has several extensions and modifications. The most well-known is perhaps the so-called Rikitake dynamo [19], which consists of two coupled disc dynamos. This dynamo can generate an oscillating magnetic field with complex dynamics [15]. The latter study uses a modified disc dynamo model proposed by [11]. In this model, the disc is radially segmented, and azimuthal current is enforced at the rim. The aim of these modifications is to eliminate the unphysical growth of the magnetic field in the limit of perfectly conducting disc. We used this model to show that the dynamo can be excited by the parametric resonance mechanism at a substantially reduced rotation rate when the latter contains harmonic oscillations in certain frequency bands [17].

Despite its simplicity, the implementation of the disc dynamo is faced with severe technical difficulties. The most challenging problem is the sliding electrical contacts which are required to convey the current between the rim and the axis of the rotating disc. The electrical resistance of such contacts, which are usually made of solid graphite brushes, can significantly exceed that of the rest of the set-up. This results in an unrealistically high rotation rate required for the dynamo to operate. Therefore, in contrast to fluid dynamos, which have been performed in several laboratory experiments using liquid metal [8, 21, 12] ([20] in preparation), the disc dynamo was generally considered technically unfeasible [18, 7, 9].

To overcome this problem, we proposed a feasible disc dynamo design with sliding liquid-metal electrical contacts [16]. This design was realized in a set-up consisting of a copper disc with a radius of 30 cm and thickness of 3 cm which was placed coaxially beneath a flat multi-arm spiral coil of the same size and electrically connected to it electrically at the centre and along the circumference by sliding liquid-metal contacts [4]. The slits make the conductivity of the coil anisotropic, which allows this essentially axially symmetric dynamo to generate an axially symmetric magnetic field [14, 2]. In this set-up, we observed a dynamo action when the disc rotation rate reached Ω≈7\Omega\approx7 Hz [3].

It is important to note that this observed critical rotation rate is somewhat lower than the theoretically predicted threshold Ωc≈8.2\Omega_c \approx8.2 Hz which corresponds to the minimal critical magnetic Reynolds number Rm≈35Rm \approx35 for a negligible contact resistance [16]. The most likely reason for this difference is the very approximate nature of the current sheet model, which we used to evaluate the dynamo threshold [16]. In particular, our original disc dynamo model is based on the assumption that the magnetic flux through the disc is the same as that through the coil. In reality, however, the effective inner radius of the disc is somewhat smaller than that of the coil. Due to the relatively high strength of the magnetic field in the central part of the helical coil, the total magnetic flux through the disc may be larger than theoretically predicted. This hints at the possibility, which is investigated theoretically in the present note, that the critical rotation rate can be reduced further by optimizing the inner and outer radii of the disc and coil.

Mathematical model

The principal set-up of the disc dynamo considered in this study is shown in Figure 1. It consists of a thin rotating disc with inner radius rir_i and outer radius rdr_d, and a coaxial coil of similar shape with the same inner radius rir_i but a larger outer radius ro>rdr_o > r_d. In contrast to our original set-up [16], now the coil is placed below the disc and sectioned by spiral slits starting only from the radius rc>rir_c > r_i. The sectioned part of the coil, which generates an axial magnetic field, extends to the outer radius ror_o. Between the radii rdr_d and ror_o, the coil has a solid upper part that connects its rim to that of the disc by a sliding liquid-metal contact. The circuit made of the disc and the coil is closed by the second sliding liquid-metal contact which is located at the inner radius rir_i.

Image not extracted

Figure 1. Schematic bottom view of the set-up which consists of a coil (1) and a disc (2) with inner radius rir_i and the outer radius rdr_d. The coil is sectioned by spiral slits, which extend from the inner radius rcr_c to the outer radius ror_o, and has two solid parts: the lower one between radii rir_i and rcr_c and the upper one between rdr_d and ror_o, which are electrically connected to the disc by sliding liquid-metal contacts (3).

As before, to simplify the analysis, the thickness of the disc and the coil is assumed to have the same thickness dd, which is much smaller than the radial dimensions of the system. The same assumption applies also to the vertical separation between the disc and the coil. Thus, the disc and the coil are treated as vertically collocated thin sheets with effective linear conductance σˉ=σd\bar{\sigma}=\sigma d.

Now consider an equilibrium dynamo state in which the disc rotating with a fixed angular velocity Ω\Omega in the magnetic field B\mathbf{B} generated by a constant electric current I0I_0 sustains this current. In the solid disc this stationary current flows is purely radial with the linear density

Jr=I02πr,J_r=\frac{I_0}{2\pi r},

which decreases due to charge conservation inversely with the cylindrical radius rr. This current flowing in the opposite radial direction through the coil is deflected sideways by the spiral slits, which thus produce an azimuthal current component proportional to the radial one:

Jϕ=−Jrβ=I0β2πr,(1)J_\phi=-J_r\beta=\frac{I_0\beta}{2\pi r}, \tag*{(1)}

where arctan⁡β\arctan\beta the pitch angle of the current lines relative to the radial direction. The shape of slits is governed by

JϕJr=r dϕdr=−β.\frac{J_\phi}{J_r}=\frac{r\,d\phi}{dr}=-\beta.

The general solution of this ordinary differential equation

ϕ(r)=ϕ0−βln⁡r(2)\phi(r)=\phi_0-\beta\ln r \tag*{(2)}

describes logarithmic spirals. The electric potential distribution in the sectioned part of the coil can be found from the Ohm’s law

J=I02πr(−er+βeϕ)=−σd∇φ,(3)\mathbf{J}=\frac{I_0}{2\pi r}(-\mathbf{e}_r+\beta\mathbf{e}_\phi)=-\sigma d\nabla\varphi, \tag*{(3)}

as

φ(r,ϕ)=I02πσd(ln⁡r−βϕ).\varphi(r,\phi) = \frac{I_0}{2\pi\sigma d}(\ln r-\beta\phi).

Thus, the potential difference along the current line over the sectioned part of the coil is

Δφco=[φ(r,ϕ(r))]rcro=I02πσd(1+β2)ln⁡rorc.(4)\Delta\varphi_c^o = [\varphi(r,\phi(r))]_{r_c}^{r_o} = \frac{I_0}{2\pi\sigma d}(1+\beta^2)\ln\frac{r_o}{r_c}. \tag*{(4)}

Over the solid part of the coil (ri≤r≤rcr_i \le r \le r_c), where the current is purely radial, which corresponds to β=0\beta=0, we respectively have

Δφic=I02πσdln⁡rcri.\Delta\varphi_i^c = \frac{I_0}{2\pi\sigma d}\ln\frac{r_c}{r_i}.

The potential difference across the disc, which rotates as a solid body with the azimuthal velocity vϕ=rΩv_\phi=r\Omega, is defined by the radial component of Ohm’s law for a moving medium

Jr=I02πr=σd(−∂rφ+vϕBz),J_r = \frac{I_0}{2\pi r} = \sigma d(-\partial_r\varphi+v_\phi B_z),

where BzB_z is the axial component of the magnetic field. Integrating the expression above over the upper part of the set-up from rir_i to rodiscr_o^{\mathrm{disc}}, which includes the disc and the solid upper part of the coil, where Ω=0\Omega=0, we have

I02πln⁡rori=σd(−Δφio+ΩΦid),(5)\frac{I_0}{2\pi}\ln\frac{r_o}{r_i} = \sigma d(-\Delta\varphi_i^o+\Omega\Phi_i^d), \tag*{(5)}

where Δφio=[φd(r)]riro\Delta\varphi_i^o=[\varphi_d(r)]_{r_i}^{r_o} is the potential difference across the upper part of the set-up and Φid=∫rirdBzr dr\Phi_i^d=\int_{r_i}^{r_d}B_z r\,\mathrm{d}r is the magnetic flux through the disc. Using the relation Bz=r−1∂r(rAϕ)B_z=r^{-1}\partial_r(rA_\phi), the latter can be expressed in terms of the azimuthal component of the magnetic vector potential AϕA_\phi as

Φid=[rAϕ]r=rird.(6)\Phi_i^d=[rA_\phi]_{r=r_i}^{r_d}. \tag*{(6)}

In the stationary state, which is assumed here, the potential difference induced by the ro tating ring in Eq. (5) is supposed to balance that over the coil defined by Eq. (4) as well as the potential drop over the liquid-metal contacts with the effective resistance R\mathcal{R}:

Δφio=Δφic+Δφco+RI0.(7)\Delta\varphi_i^{o}=\Delta\varphi_i^{c}+\Delta\varphi_c^{o}+\mathcal{R}I_0. \tag*{(7)}

This equation implicitly defines the critical rotation rate at which a steady current can sustain itself.

To complete the solution, we need to evaluate the magnetic flux (6) through the rotating disc. The azimuthal component of the vector potential that appears in Eq. (6) is generated by the respective component of the electric current which is present only in the coil. Thus, we have

Aϕ(r,z)=μ04π∫02π∫rcroJϕ(r′)cos⁡ϕ r′ dr′ dϕr′2−2r′rcos⁡ϕ+r2+z2.A_{\phi}(r,z)=\frac{\mu_0}{4\pi}\int_{0}^{2\pi}\int_{r_c}^{r_o}\frac{J_{\phi}(r')\cos\phi\,r'\,\mathrm{d}r'\,\mathrm{d}\phi}{\sqrt{r'^{2}-2r'r\cos\phi+r^{2}+z^{2}}}.

where zz is the axial distance from the coil ring carrying the azimuthal current JϕJ_{\phi} defined by Eq. (1). Note that the poloidal currents with radial and axial components circulating through the rings and liquid-metal contacts produce a purely toroidal magnetic field, which is parallel to the velocity of the rotating ring, and thus do not interact with the latter. In the plane of the ring (z=0z=0), the double integral above can be evaluated analytically as

Aϕ(r,0)=μ0βI08π2[F(rc/r)−F(ro/r)].A_{\phi}(r,0)=\frac{\mu_0\beta I_0}{8\pi^2}\left[F(r_c/r)-F(r_o/r)\right].

where F(x)=(1−x)K(m+)+(1+x)E(m+)+sgn⁡(1−x)[(1+x)K(m−)+(1−x)E(m−)]F(x)=(1-x)K(m_+)+(1+x)E(m_+)+\operatorname{sgn}(1-x)\left[(1+x)K(m_-)+(1-x)E(m_-)\right] is defined in terms of the complete elliptic integrals of the first and second kind, K(m±)K(m_{\pm}) and E(m±)E(m_{\pm}), of the parameter m±=±4x(1±x)2m_{\pm}=\frac{\pm4x}{(1\pm x)^2} [1]. Using the above definition, the magnetic flux (6) can be written as

Φid=μ0βI0ro8π2Φ‾(λi,λc,λd)(8)\Phi_i^d=\frac{\mu_0\beta I_0r_o}{8\pi^2}\overline{\Phi}(\lambda_i,\lambda_c,\lambda_d) \tag*{(8)}

where

Φ‾(λi,λc,λd)=λd[F(λc/λd)−F(λd−1)]−λi[F(λc/λi)−F(λi−1)](9)\overline{\Phi}(\lambda_i,\lambda_c,\lambda_d)=\lambda_d\left[F(\lambda_c/\lambda_d)-F(\lambda_d^{-1})\right]-\lambda_i\left[F(\lambda_c/\lambda_i)-F(\lambda_i^{-1})\right] \tag*{(9)}

is a dimensionless magnetic flux and λi=ri/ro\lambda_i = r_i/r_o, λc=rc/ro\lambda_c = r_c/r_o, λd=rd/ro\lambda_d = r_d/r_o are the dimensionless counterparts of the corresponding radii which are defined using the outer coil radius ror_o as the length scale for this problem. Substituting the relevant parameters into Eq. (7), we obtain the critical magnetic Reynolds number

Rm=μ0σdroΩ=4π(Rˉ−2ln⁡λi−β2ln⁡λc)βΦ(λi,λc,λd),(10)Rm = \mu_0\sigma d r_o\Omega= \frac{4\pi(\bar{R} - 2\ln\lambda_i - \beta^2\ln\lambda_c)}{\beta\Phi(\lambda_i,\lambda_c,\lambda_d)}, \tag*{(10)}

which defines the dynamo threshold depending on the dimensionless contact resistance Rˉ=2πσdR\bar{R} = 2\pi\sigma d\mathcal{R}, the spiral pitch β\beta, and three radii ratios λi,λc,λd\lambda_i,\lambda_c,\lambda_d.

46Results

In the following, we assume Rˉ=0\bar{R} = 0, which corresponds to a relatively small contact resistance, as in our previous set-up, and search for an optimal set the remaining four parameters which yield the lowest possible RmRm. First, an optimal β\beta is found in terms of the radii ratios by solving the stationarity condition ∂Rm∂β=0\frac{\partial Rm}{\partial\beta} = 0, which reduces to

2β−2ln⁡λi−ln⁡λc=0,(11)2\beta^{-2}\ln\lambda_i - \ln\lambda_c = 0, \tag*{(11)}

and straightforwardly yields

β=2ln⁡λi/ln⁡λc.(12)\beta= \sqrt{2\ln\lambda_i/\ln\lambda_c}. \tag*{(12)}

Substituting this back into Eq. (10), we obtain

Rm=16πln⁡λiln⁡λc2 Φˉ(λi,λc,λd)(13)Rm = \frac{16\pi\sqrt{\ln\lambda_i\ln\lambda_c}}{\sqrt{2}\,\bar{\Phi}(\lambda_i,\lambda_c,\lambda_d)} \tag*{(13)}

depending on λi,λc,λd\lambda_i,\lambda_c,\lambda_d.

With λd=1\lambda_d = 1 and λi=λc\lambda_i = \lambda_c, which correspond to rc=rir_c = r_i and rd=ror_d = r_o, as in our original model, the lowest Rm≈34.6Rm \approx34.6 is obtained in the radii ratio λ=ri/ro≈0.36\lambda= r_i/r_o \approx0.36 (see Figure 2) [16]. First, this basic set-up can be improved by optimazing the inner coil radius λc\lambda_c so that to minimize RmRm for the given inner disc radius λi\lambda_i. The optimal λc\lambda_c along with the corresponding

Image not extracted

Figure 2. Marginal RmRm versus the inner disc radius λi\lambda_i for the same-size inner coil radius λc=λi\lambda_c=\lambda_i and for optimal λc\lambda_c which is shown on the axis at the right; here λd=1\lambda_d=1 and Rˉ=0\bar{R}=0.

RmRm found numerically using Mathematica [22] are plotted in Figure 2 against λi\lambda_i. As seen in Figure 3, for this partially constrained set-up with λd=1\lambda_d=1, the lowest Rm≈14.9Rm\approx14.9 is attained at λi≈0.78\lambda_i\approx0.78 and λc≈0.94\lambda_c\approx0.94. Optimizing also λd\lambda_d, we find the lowest possible Rm≈11.6Rm\approx11.6 at λd≈0.98\lambda_d\approx0.98, λi≈0.88\lambda_i\approx0.88 and λc≈0.96\lambda_c\approx0.96.

Image not extracted

Figure 3. Isocontours of the magnetic Reynolds number RmRm in the plane of radii ratios (λi,λc)(\lambda_i,\lambda_c) for λi≤λc\lambda_i \le\lambda_c and λd=1\lambda_d = 1.

It is important to note that the coil in this optimal set-up represents a very narrow ring with the relative width 1−λc≈0.041-\lambda_c\approx0.04. First, such a coil may be difficult to make. Second, it may not be adequately described by the thin-sheet approximation used in our mathematical model. The same applies largely also to the previous (suboptimal) set-up with λd=1\lambda_d=1. On the other hand, Figure 2 indicates that the increase in RmRm caused by the reduction of λi\lambda_i below its optimal value remains relatively small down to λi≈0.3\lambda_i\approx0.3. Choosing this value for the inner disc radius, and minimizing RmRm with respect to λc\lambda_c and λd\lambda_d, we obtain Rm≈14.6Rm\approx14.6 at λc≈0.74\lambda_c\approx0.74 and λd≈0.9\lambda_d\approx0.9. The spiral pitch angle, which corresponds to the optimal β\beta and is defined by Eq. (12), is arctan⁡β≈70.5∘\arctan\beta\approx70.5^\circ.

Conclusion

We found a feasible disc dynamo set-up with a critical magnetic Reynolds number Rm≈15Rm\approx15 that is more than twice lower than that for our original design. For a set-up with an outer radius Ro=15R_o=15 cm and a thickness of d=1.5d=1.5 cm, that is twice smaller than in our original set-up, dynamo is expected to work at the critical rotation frequency

f=Ω2π=Rm2πμ0σCudRo≈15 Hz,f = \frac{\Omega}{2\pi} = \frac{Rm}{2\pi\mu_0\sigma_{\mathrm{Cu}}dR_o} \approx15\ \mathrm{Hz},

where μ0=4π×10−7 H/m\mu_0 = 4\pi\times10^{-7}\ \mathrm{H/m} is the vacuum permeability and σCu=5.96×107 S/m\sigma_{\mathrm{Cu}} = 5.96 \times10^{7}\ \mathrm{S/m} is the conductivity of copper. For sliding contacts with a gap of width δ=0.25 mm\delta= 0.25\ \mathrm{mm} filled with the eutectic alloy of GaInSn, as in our original set-up [3], the relative contact resistance can be estimated as [16]

R‾≈σCuσGaInSnδRo(1λi+1λd)≈10−2,\overline{R} \approx\frac{\sigma_{\mathrm{Cu}}}{\sigma_{\mathrm{GaInSn}}}\frac{\delta}{R_o}\left(\frac{1}{\lambda_i}+\frac{1}{\lambda_d}\right) \approx10^{-2},

where σGaInSn=3.3×106 S/m\sigma_{\mathrm{GaInSn}} = 3.3 \times10^{6}\ \mathrm{S/m} [13]. This confirms the assumed negligibility of R‾\overline{R}.

References

  1. [1]M. Abramowitz and I. A. Stegun. Handbook of Mathematical Functions. Dover, New York, 1972.
  2. [2]T. Alboussière, F. Plunian, and M. Moulin. Fury: an experimental dynamo with anisotropic electrical conductivity. Proc. R. Soc. Lond. A, 478:20220374, 2022.DOI
  3. [3]A. R. Avalos-Zúñiga and J. Priede. Realization of Bullard’s disc dynamo. Proc. R. Soc. Lond. A, 479(2271):20220740, 2023.
  4. [4]R.A. Avalos-Zúñiga, J Priede, and CE Bello-Morales. A homopolar disc dynamo experiment with liquid metal contacts. Magnetohydrodynamics, 53:341–348, 2017.arxiv.org/abs/1703.00467
  5. [5]R. Beck, A. Brandenburg, D. Moss, A. Shukurov, and D. Sokoloff. Galactic magnetism: recent developments and perspectives. Annu. Rev. Astron. Astrophys., 34(1):155–206, 1996.DOI
  6. [6]E. Bullard. The stability of a homopolar dynamo. Proc. Camb. Phil. Soc., 51(4):744–760, 1955.DOI
  7. [7]E. Dormy and A. M. Soward. Mathematical aspects of natural dynamos. CRC, 2007.
  8. [8]A. Gailitis, O. Lielausis, E. Platacis, S. Dement’ev, A. Cifersons, G. Gerbeth, T. Gundrum, F. Stefani, M. Christen, and G. Will. Magnetic field saturation in the riga dynamo experiment. Phys. Rev. Lett., 86(14):3024, 2001.
  9. [9]P. Lorrain, F. Lorrain, and S. Houle. Magneto-fluid dynamics: fundamentals and case studies of natural phenomena. Springer, 2006.DOI
  10. [10]H. K. Moffatt. Magnetic field generation in electrically conducting fluids. Cambridge monographs on mechanics and applied mathematics. Cambridge, 1978.
  11. [11]H. K. Moffatt. A self-consistent treatment of simple dynamo systems. Geophys. Astrophys. Fluid Dyn., 14(1):147–166, 1979.DOI
  12. [12]R. Monchaux, M. Berhanu, M. Bourgoin, M. Moulin, Ph. Odier, J.-F. Pinton, R. Volk, S. Fauve, N. Mordant, F. Pétrélis, A. Chiffaudel, F. Daviaud, B. Dubrulle, C. Gasquet, L. Marié, and F. Ravelet. Generation of a magnetic field by dynamo action in a turbulent flow of liquid sodium. Phys. Rev. Lett., 98:044502, Jan 2007.DOI
  13. [13]U. Müller and L. Bühler. Magnetofluiddynamics in Channels and Containers. Springer, 2001.
  14. [14]F. Plunian and T. Alboussière. Axisymmetric dynamo action is possible with anisotropic conductivity. Phys. Rev. Res., 2(1):013321, 2020.DOI
  15. [15]F. Plunian, Ph. Marty, and A. Alemany. Chaotic behaviour of the rikitake dynamo with symmetric mechanical friction and azimuthal currents. Proc. R. Soc. Lond. A, 454(1975):1835–1842, 1998.DOI
  16. [16]J. Priede and R. Avalos-Zúñiga. Feasible homopolar dynamo with sliding liquid-metal contacts. Phys. Lett. A, 377(34-36):2093–2096, 2013.DOI
  17. [17]J. Priede, R. Avalos-Zúñiga, and F. Plunian. Homopolar oscillating-disc dynamo driven by parametric resonance. Phys. Lett. A, 374(4):584–587, 2010.DOI
  18. [18]K.H. Rädler and M. Rheinhardt. Can a disc dynamo work in the laboratory? Magnetohydrodynamics, 38(1-2):211–217, 2002.DOI
  19. [19]T. Rikitake. Oscillations of a system of disk dynamos. Math. Proc. Cambridge Philos. Soc., 54(1):89–105, 1958.DOI
  20. [20]F. Stefani, A. Gailitis, G. Gerbeth, A. Giesecke, Th. Gundrum, G. Rüdiger, M. Seilmayer, and T. Vogt. The dresdyn project: liquid metal experiments on dynamo action and magnetorotational instability. Geophys. Astrophys. Fluid Dyn., 113(1-2):51–70, 2019.DOI
  21. [21]R. Stieglitz and U. Müller. Experimental demonstration of a homogeneous two-scale dynamo. Phys. Fluids, 13(3):561–564, 2001.
  22. [22]S. Wolfram. The Mathematica Book. Wolfram Media, 2003.

Paper details

vlnc:W_vpmu5p95
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents