Neural network modeling for separation of bentonite in tubular ceramic membranes

摘要: The neural network model is used for obtaining an estimation of permeate flux and rejection over the entire range process variables. This approach has been extended in this study applied to prediction sustainability membrane efficiency ceramic tubular membranes. Experimental results involving use turbulence promoters empty filtration have obtained are directly compared predicted values from black box model. Flux dependent on feed temperature, system pressure, concentration crossflow velocity. Neural networks also offer added advantage being quite straightforward its application. possibility using BPNN (back-propagation network) accurately predict variable effects included. Turbulence were experimentally significantly enhance during microfiltration dilute bentonite suspensions. Artificial can very real behaviour with relative errors reaching at most 5%. In order obtain data set necessary train different networks, three concentrations, pressures, temperatures one flowrate tested several operating conditions.