Holdup prediction in inverse fluidization using non-Newtonian pseudoplastic liquids: Empirical correlation and ANN modeling

Abstract

The bed expansion characteristics in inverse fluidization using non-Newtonian liquids are reported. Experiments have been carried out using single and binary systems of four different polymeric solids and four different non-Newtonian pseudoplastic liquids in two different columns. Empirical correlation has been developed to determine the bed expansion characteristics as a function of physical and dynamic variables of the system. A multilayer perceptron trained with backpropagation and Levenberg–Marquardt algorithm have been used for the Artificial Neural Network (ANN) analysis. Four different standard transfer functions in a single hidden layer are used. The ANN model with Levenberg–Marquardt algorithm with transfer function 2 having 12 processing elements in hidden layer gives good predictability of the bed height. Statistical analysis indicates that both the empirical correlation and the ANN prediction give acceptable results.