background image
MHD AND POTSHELL MECHANICAL DESIGN OF A 740 KA CELL

Marc Dupuis, Jonquière
Valdis Bojarevics, Greenwich
Daniel Richard, Montréal
In 2005 in another article published in ALUMINIUM
[1], it was demonstrated that as far as the thermal
balance aspect of the cell design is concerned, there is
no limit to the size of cells that could be designed.
In order to substantiate that declaration, the thermal-
electric design of a 740 kA cell was presented. But
what about the cell MHD stability and the potshell
mechanical design for that 9 risers, 26.2 meters long
cell?
The present article covers both those aspect of the cell
design. Results suggest that as far as the MHD cell
stability aspect of the cell design is concerned, there is
no limit to the size of cells that could be designed.
Finally, as far as the potshell mechanical design is
concerned, after taking care of the potshell thermal
deformation issue [2], there is no limit in sight to the
size of cells that could be designed either.
500 kA cell MHD design
In his TMS 2005 paper [3], Urata clearly indicates
that, in addition to the average magnitude, it is the
gradient of the vertical component of the magnetic
field (Bz) in the longitudinal direction of the cell that
is responsible for the main instability mechanism in a
modern side by side high amperage cell.
So it is not surprising that in their 1987 busbar patent
[4], Pechiney researchers describe how they minimized
that Bz longitudinal gradient (see figure 2 of [4]) by
using compensation busbars running along both sides
of the row of cells carrying current in the same
direction as the potline (see figure 3 of [4]). According
to the Pechiney patent, this compensation busbar
configuration will work well up to 500-600 kA cells.
This compensation busbar configuration has been
analyzed using MHD cell stability modeling tool [5, 6
and 7] on a 500 kA cell (see figure 1 for the busbar
configuration). It indeed reduced the Bz magnetic field
gradient in the longitudinal direction to about 40 Gauss
over 17.3 meters (from 20G to ­20G, see figure 2)
despite the fact that all 6 positive busbars are running
under the cell.
Yet, even with that relatively low Bz longitudinal
gradient, the MHD model is predicting that a coupled
(2,0) and (0,1) bath metal interface deformation wave,
the exact type of wave predicted in Urata's paper, will
grow in the cell (see figure 3 and 4).
Of course, there was no attempt to really optimize the
compensation busbar configuration, so for sure, a
better setup can exist. But already in order to reduce
the Bz longitudinal gradient to 40 Gauss over 17.3
meters, the compensation busbar on the side of the
return line must carry 250 kA and the compensation
busbar on the opposite side must carry 145 kA. So in
the case presented here, a total of 395 kA or 79% of
the potline current is been carry by the 2 compensation
busbars. It is very doubtful that this compensation
busbar configuration could be extend to a 740 kA cell.
In order to stabilize very high amperage cells, a new
compensation busbar configuration has been
developed. The resulting average Bz is so close to 0
(0.0003 T) that no distinctive low frequency wave can
develop in the cell (see figure 5 and 6).
740 kA cell MHDdesign
Notably, this new compensation busbar configuration
is working for any length of potshell and any number
of positive busbars running under the cell. The 9 risers,
740 kA cell design presented before [1] has too a
resulting Bz so close to 0 that no distinctive low
frequency wave can develop in the cell (see figure 7
and 8).
Extrapolation to a 2380 kA cell design
This new compensation busbar configuration will
equally well work for any reasonable length of potshell
and any number of risers, for example a 85.8 meter
long, 30 risers, 2380 kA cell [1] could too be
magnetically compensated using the same approach.
So it is possible to conclude that as far as the MHD
cell stability aspect of the cell design is concerned,
there is no limit to the size of cells that can be
designed.