Overcoming instabilities in aluminium reduction cells: a route to cheaper aluminium

The amount of energy currently used to reduce alumina to aluminium in elect rolysis cells is staggering, around 10(11) kWh/year. Yet much of this energ y (almost one-half) is lost in the form of (IR)-R-2 heating of the highly r esistive electrolyte. Strenuous efforts have been made to minimise these lo sses by reducing the volume of electrolyte in the cells. However, the alumi nium industry has come up against a fundamental problem: when the depth of the electrolyte is reduced below a critical threshold (around 4-5 cm), the liquids in the cell start to 'slosh around' in an uncontrolled fashion. Thi s is an instability, fuelled by the intense currents which pass through the liquids. At present, cells operate just above the critical electrolyte dep th, but if this depth were reduced from, say, 4.5 to 4 cm, then the annual savings would exceed pound 10(8). After a number of false starts, we now ha ve a clear understanding of the physical mechanisms which underpin the inst ability, and it turns out that these are remarkably simple. These mechanism s are described here and it is shown that, although the cell geometry is to o complex to produce an accurate mathematical model, the underlying mechani sms are so simple that it becomes clear how to suppress the instability, Th us, after two decades of research, we are finally in a position to design i nherently stable cells. For example, it is shown that slow, small movements of the anode assembly can lower the critical electrolyte depth to around 2 cm, Such a control system could be retrofitted to most existing cells. MST /4501.