Bed Expansion in Two-Phase Liquid–Solid Fluidized Beds with Non-Newtonian Fluids and ANN Modelling

Abstract

Bed expansion characteristics in non-Newtonian fluid–solid two-phase fluidized bed have been examined. Bed expansion measurements are carried with irregular shaped sized particles and non-Newtonian liquids (aqueous solutions of carboxymethyl cellulose). The bed voidage is found to be increasing with the increasing superficial liquid velocity. It has also been noticed that the value of this increase has been smaller as the particle size increases. Also, increase in the viscous non-Newtonian type flow behaviour has increased in bed voidage. A multilayer perceptron neural network trained with the backpropagation as well as the Levenberg-Marquardt algorithm has been attempted for the proposed analysis. Four types of common transfer functions are used with the use of a single hidden layer. The best performing ANN model has been the Levenberg-Marquardt algorithm applied with transfer function 4 having 11 processing elements within the hidden layer for the predictability related to the bed height ratio.