Herramientas avanzadas en el proceso de biosorción de ciprofloxacina en columna

Abstract

In this work, the removal of ciprofloxacin in synthetic waters was investigated in fixed bed column at laboratory and pilot plant scale using sugarcane bagasse (SCB) and corn cob (CC) as bioadsorbents. Advanced tools such as artificial neural networks(ANN) and specialized software were used for the modeling and simulation of the adsorption of ciprofloxacin (CFX) with both agro-industrial wastes. On a laboratory scale, corncob has an adsorption capacity of 1.98 mg∙g-1 while sugarcane bagasse has 1,28 mg∙g-1 under the same operating conditions (CFX concentration=5 mg∙L-1, flow rate = 7 mL∙min-1, column diameter = 2.2 cm). At the pilot plant scale (flow rate= 28 mL∙min-1) the adsorption capacity for the SCB is 1,097 mg∙g-1 as for the CC is 0.786 mg∙g-1, being lower in both cases. The empirical mathematical model that best reproduced the CFX breakthrough curve in SCB and CC at both scales was the Dose-Response model. The modeling of the rupture curves with ANN presented very high correlation coefficients (R2), it was determined that the network created in Tensorflow is more robust than the network in Matlab for both scales and both biosorbent materials by means of a sensitivity analysis. The dynamic simulation at laboratory scale for the SCB with Aspen Adsorption generated a breakthrough curve with better fit to the experimental data than using Comsol Multiphysics, however, using CC the fit is greater in Comsol Multiphysics for both scales.