its not a property of molten aluminum to freeze in 6 hours, it depends on the thermal insulation.

what I proposed was to have a secondary sealed thermal bath of aluminum which gets molten during cheap power and which returns heat during expensive power to maintain the temperature of the primary (work) batch of aluminum so that the work batch doesn't freeze

the heat can be transfered with an appropriate working fluid.

since the secondary thermal bath aluminum doesn't need to satisfy client demands (specific alloy etc) the alloy can be chosen so as to have a suitable melting point for such a setup