Comparative studies on modelling and optimization of hydrodynamic parameters on inverse fluidized bed reactor using ANN-GA and RSM

Optimization of hydrodynamic parameters in counter flow inverse fluidized bed reactor (IFBR) is studied in this paper. Comparative analysis was made between artificial neural network (ANN) and response surface methodology (RSM) to evaluate the parameters. The effects of operating variables such as bed volume (300-1200 cm(3)), superficial liquid velocity (0.37-1.84 cm/s) and superficial gas velocity (0.07-0.59 cm/s) on percentage bed expansion, liquid holdup, gas holdup, solid holdup and average pressure drop were evaluated using three-factorial Box-Behnken design (BBD). The same was utilized to train a feed forward multilayer perceptron (MLP), ANN with back-propagation algorithm. The predicted values of the both the methodologies were compared with error functions such as root mean square error (RMSE), mean square error (MSE), mean absolute error (MAE), model predictive error % (MPE), chi-square (chi(2)) and correlation coefficient (R-2). Optimization of operating conditions was obtained through Derringer's desirability function (RSM) and genetic algorithm (GA), and the results were compared. It is ascertained that well trained ANN-GA has provide a high sensitive results. (C) 2018 Production and hosting by Elsevier B.V. on behalf of Faculty of Engineering, Alexandria University.