Furnace Stirrer 3D

Drag to orbit, scroll or pinch to zoom. The glowing layer is an alloy addition sitting on top of the melt. The stirrer under the floor covers only part of the bath width, so the melt also flows sideways. 2D side view · Vortex toy

x field direction · y up · z across the bath

This view needs WebGL 2. Try a recent Chrome, Safari or Firefox, or the 2D side view.

Frequency5.0 Hz

Power100 %

Bath depth0.80 m

Stirrer width60 %

Alloy opacity3.0

Reversal

Resolution

Homogeneity0%

To 95%n/a

Sim time0s

Skin depth0mm

Peak flow0m/s

Cross flow0m/s

Model notes

Force: a traveling-wave push along the floor under the left half of the bath, decaying with height as exp(−2d/δ). Skin depth δ = √(1 / (π f μ₀ σ)) with σ = 3.8 MS/m for liquid aluminum (approximate). The cyan box over the stirrer is drawn δ tall.
Flow: 3D incompressible solver (stable fluids) on a 56 × 22 × 28 grid with 36 mm cells (Fast) or 84 × 34 × 42 with 24 mm cells (Fine). Above about 20 Hz the skin depth is only one or two Fast cells thick, and Fast over-predicts mixing; Fine is closer to converged. Walls and floor are no-slip and the top is free-slip. Vorticity confinement and weak drag stand in for turbulence.
At high frequency δ is smaller than one cell. The solver keeps the total thrust but cannot resolve the thin layer, so treat high-frequency results as rough.
Cross flow is the fastest motion across the bath width. It comes from the stirrer being narrower than the bath and from friction on the side walls. A 2D model has none.
Wall force density is held constant. Frequency changes penetration depth only. Stirrer width, bath length and the 0.20 m pole pitch are assumptions, not machine data.
Left out: free-surface motion, thermal convection, the refractory gap, and any calibration against real installations. Qualitative only. Do not use it to size or predict a real installation.