Liquid metal magnetohydrodynamic flow driven by a system of rotating permanent magnets in a container of orthogonal cross-section has been studied. The main objective of the work is to research the impact of magnetic forcing parameters (magnetic field value, magnets arrangement, and angular velocity of their rotation) on the generated hydrodynamic structures and flow modes. On this basis, we contemplate realizing required flow features by setting certain parameters of the driving magnetic system. A numerical study of the problem in the induction-free approximation without taking into account the effect of the variable component of electromagnetic force is presented. The parameters of spin-up modes and steady-state flow regimes have been calculated by three-dimensional direct numerical simulation based on COMSOL Multiphysics 4.3a software and experimentally verified on a specially designed setup using noninvasive Doppler ultrasound technique.
The limits of the axisymmetric ideal inductor approach and the low magnetic Reynolds number (Rem) approach are demonstrated for the case study of a flow driven by a rotating magnetic field induced by an explicit-pole non-ideal inductor. These effects are studied for an intermediate range of the magnetic Reynolds number 0.1 ≤ Rem ≤ 1. It is shown that phenomena that do not exist under the ideal inductor approach change the distribution of the electromagnetic body force (EMBF) from axial symmetry to rotational symmetry. However, it is found that the magnetic field advection (MFA), which is neglected under the low Rem approach, induces body force that changes the distribution of the EMBF so that it effectively becomes closer to being axisymmetric. Therefore, it is suggested that the ideal inductor approach should not be applied in low Rem applications; this approach can be applied in high Rem applications. It is also found that the effect of MFA on the averaged velocity field is negligible for higher values of Rem than traditionally assumed, but MFA has a significant effect in reducing the turbulent kinetic energy. The reduction is found to be a result of the MFA-induced EMBF reducing the turbulence production.
Small-scale turbulence measurements in liquid metals are generally conducted using potential difference probes (PDPs). However, the use of PDPs in alternating magnetic field-driven flows is limited as the driving magnetic field (DMF) induces a noise signal, which cannot be separated from the turbulence signal using a filtering algorithm that does not take into consideration the physics of both the noise and the flow. Therefore, we developed a post-measurement method for the rejection of DMF-induced signals. The method was demonstrated for a case study of rotating magnetic field (RMF)-driven flow in which the interaction between the PDP and DMF was investigated. The probe was successfully aligned so that the DMF-induced signal could be linearly separated from the turbulence signal. Consequently, our method enables the properties of both the cyclostationary induced signal and turbulence to be utilized to reconstruct DMF-induced signals. The induced signal is reconstructed using a moving ensemble (or ergodic) averaging algorithm that removes the turbulence contribution to the original signal. The algorithm was tested by applying it to velocity measurements that were conducted inside liquid Ga–In–Sn alloy subjected to RMF driven flow, and it was found that the developed method enabled the successful rejection of the DMF-induced signal. The rejection of the induced signal revealed an unknown spectral distribution of the turbulent kinetic energy for RMF-driven flow at a high magnetic Taylor number. This distribution could not have been identified without the rejection of the induced signal.