 |
Computational Modelling
Numerical Simulations
Mathematical Modelling |  |
|
Dr. Georgi Djambazov
Senior Research Fellow
|
|
Containerless Melting of Reactive Metals
(G. Djambazov's work October 1998 - July 2001)
EPSRC Grant No:- GR/L97483
TIFF pictures of a restarted 3-dimensional
simulation of a melting case with a 2-layer, 4-turn coil, 3400 A, 6000 Hz,
where an equilibrium shape of the liquid metal body is established:
1. Difference of the vertical component of 3D magnetic induction
from axisymmetric
2. Difference of the radial component of the magnetic induction
from axisymmetric
3. Radial, and
vertical component of magnetic force
4. Temperature contours for melting alumunium
5. Instantaneous contours of the liquid fraction,
pressure, and
velocity magnitude
6. Velocity vectors in y=0 plane,
zoom in corner
- Centre line of 3-dimensional coil
(Leads are parallel to x-axis.)
- Difference of radial force from axisymmetric
- Difference of radial force from axisymmetric
(bottom chill block taken into account)
- Another 3D coil (4 turns in 2 layers)
- Difference of vertical magnetic induction from axisymmetric
- Velocity field towards the end of melting
- Another simulation with the 3D, 2-layer coil (4 turns)::
(starting from a liquid state and allowing the bottom to solidify,
simulation time of 1 s in 1000 time steps.
An equilibrium shape is established with flow inside. Coil leads
are parallel to axis x.)
- Difference of magnetic induction vector in yz plane
from axisymmetric (Some instability of the surface
tracking is visible on one side.)
- Radial component of magnetic force
- Verical component of magnetic force
- Velocity vectors in yz plane
- Turbulent viscosity contours
- Liquid fraction contours
- Dynamic shape and velocities in radial section
(3D simulation, single layer conical coil; 4 pages )
- (2+2)*3400 A, 6000 Hz, melting Aluminium
(Axisymmetric simulation; 12 pages )
- (!) 100 pages
- Movement history of selected points on the surface
- A bad case where the metal tries to spill out, and the
sumulation breaks up because of the distorted mesh
(30 pages, time step 0.2 ms, 3D simulation)
- Algorithm for surface tracking and moving mesh
- Volume conservation method for tracking the free surface
- Mesh adaptation by sliding the outermost nodes
- Mesh adaptation by deforming the whole mesh
- Typical configuration of semilevitation device
- Typical presure field in semilevitated molten metal
- Typical radial force constricting the molten metal
- Typical vertical force and velocity field inside the melt
- Calculating the electromagnetic induction field
(Finite Volume formulation)
- Quasi-steady equations (with source terms)
- Boundary conditions on the surface
- Test case on a sphere: comparison of computed (finite
volume) and analytic electric current density in a metal
sphere with a single current loop around equator.
- Finest mesh used for the sphere
- 3D picture
- First Report.
- Second Report.
- G. Djambazov, K. Pericleous, V. Bojarevics.
Free surface and mesh control for the finite volume
simulation of semi-levitation melting.
In 8th Annual Conference of ACME, pp. 42-45, April 2000.
- G. Djambazov, K. Pericleous, V. Bojarevics.
Adapting mesh and magnetic field for simulations of
levitation melting.
In ECCOMAS CFD Conference, Swansea, September 2001.
Email: G.Djambazov@gre.ac.uk