Simulation Of The Viscosity Of Ceramic Mineral Suspensions By Neural Networks

The mineral suspensions do not present a newtonian behavior as their continuous phase; they can be pseudoplastic, dilatant, or both. The viscosity becomes as the most important rheological variable because the different industrial processes involve a wide variety of rheological scenes, with a specific rheological behavior for each one. Some authors have tried to model the viscosity of suspensions considering the volumetric fraction of solids and others have also kept in mind the size of the particles (Einstein, 1906; Batchelor, 1977; Jinescu, 1974; Krieger and Dougherty, 1959; Tangsathitkulchai and Austin, 1988). Others consider the shear rate (Cross, 1965; Hershel-Bulckley, 1926; Bingham, 1916, etc). However, none of the models is able to simulate the viscosity involving the physical, chemical and mechanic factors at the same time for crowded suspensions. A new model based on Neural Networks is presented, taking into account all the variables of ceramic clay suspensions and chemical additives. Keywords: viscosity, clay minerals, neural networks, simulation, mathematical function