Abstract
In order to optimize the process and control the variables in the washing tanks of the aluminum production process, the design of a mixing tank has been considered, which handles incoming liquid flows, both hot and cold, for manipulate the outlet temperature of the tank, as well as the torque and simultaneously visualize the associated variables. It is expected that this system not only increases the efficiency of the mixture but also reduces the risk involved in the handling of substances at high temperatures. When designing intelligent supervision, the strategies of use of variables in real time and robust use models against disturbances. The system was created in Matlab© using a neuronal Backpropagation model. We worked on density, torque, turbidity, tank level and mud level in the thickener tanks, to analyze the quality of the product and its link to the Bayer cycle.