This study investigates how to simulate alternating nonuniform magnetic field effects on ferrofluid flow using ANSYS Fluent. Ferrofluids are special liquids containing tiny magnetic nanoparticles (usually Fe3O4) suspended in water or oil that respond strongly to external magnetic fields. When engineers place ferrofluids in alternating magnetic fields, the fluid movement changes dramatically as the magnetic force repeatedly attracts and releases the fluid, creating unique flow patterns. These nonuniform magnetic fields generated by strategically placed magnetic dipoles create powerful Kelvin body forces that disturb the thermal boundary layer and significantly boost heat transfer rates. Based on groundbreaking research paper by Goharkhah and Ashjaee (2014), this simulation captures how ferrofluid magnetization changes with temperature and field strength, causing the fluid to accelerate near heated surfaces. Our model incorporates full ferrohydrodynamic equations to predict the complex interaction between magnetic body forces, fluid viscosity, and thermal gradients inside channels. This approach has shown impressive heat transfer enhancement of up to 13.9 times normal rates at specific Reynolds numbers and optimal field frequencies, making alternating magnetic fields a revolutionary technique for thermal management in microelectronics, biomedical devices, and energy systems.

• Reference [1]: Goharkhah, Mohammad, and Mehdi Ashjaee. “Effect of an alternating nonuniform magnetic field on ferrofluid flow and heat transfer in a channel.” Journal of magnetism and magnetic materials362 (2014): 80-89.

Figure 1: Schematic of the studied problem [1]

Simulation Process

For our ferrofluid simulation, we created a simple 2D channel geometry based on the research paper’s setup. The channel has two parallel walls where constant heat flux is applied. We used structured meshing in ANSYS to create high-quality rectangular cells with refinement near the walls where temperature gradients are highest. The key to this simulation was writing a custom User-Defined Function (UDF) that implements the complex magnetic field effects. This UDF calculates the magnetic body forces on the fluid using the exact Kelvin force formulas from the paper. We also created a User-Defined Scalar (UDS) to track how the magnetic susceptibility changes with temperature. The UDF implements eight magnetic dipoles (four on top, four on bottom) that turn on and off following a rectangular wave pattern, creating the alternating magnetic field effect. Additionally, we employed single-phase approach to model ferrofluid nanofluid.

Post-processing

The simulation proves dramatic flow disturbances caused by magnetic forces! Looking at the colored contours of User-Defined Memory-0, we see how the magnetic field strength varies from 0 to 133 around each dipole. This creates a powerful and highly non-uniform force field that pulls the ferrofluid toward areas of highest field gradient. When dipoles alternate between on and off states, they create pairs of counter-rotating vortices near the channel walls. These vortices break up the thermal boundary layer and mix the fluid, explaining why heat transfer improves so dramatically. The simulation perfectly captures how these magnetic effects decrease with distance from each dipole source – notice how the field rapidly weakens within just a few millimeters from each dipole location.

Figure 2: Magnetic Field Distribution and Flow Disturbance

Our UDF perfectly implements the alternating magnetic field pattern! The second image shows the square wave function that controls dipole activation, with clear connection time (Δt₂) and disconnection time (Δt₁) periods. This pulsing magnetic field creates a unique flow situation where fluid is repeatedly pulled toward walls and then released. The simulation reveals that finding the optimal frequency is critical – too slow and the fluid doesn’t mix enough, too fast and the fluid doesn’t have time to respond to magnetic forces. At Reynolds number 2000, a frequency of 20 Hz achieved the maximum heat transfer enhancement of 13.9 times compared to non-magnetic conditions. The pressure drop increased by only 6 times at these conditions, showing that magnetic field-induced mixing is far more efficient than trying to achieve similar mixing through higher flow rates.

Figure 3: Square Wave Function Implementation