Rosa J., Prado K., Alves W., Pereira F., Santana J., Tambourgi E., 2013, Applying of a Neural Network in Effluent Treatment Simulation as an Environmental Solution for Textile Industry, Chemical Engineering Transactions, 32, 73-78.
Considerable attention has been given to issues associated with the presence of colored compounds in aqueous wastewater generated from textile industries, since water is the only efficient carrier for dyes and other compounds that are used in the dyeing and finishing processes. An average textile finishing company uses 100,000 – 150,000 liters of water per ton of textile material treated. This work aimed asimulation of the biodegradation of a dye in textile process using a perceptron multilayer neural network. A 24-1 experimental design has been drawn up to study the effect of dye (0.01-0.18 g/L), glucose (0.66-2.43%, w/v), microorganism (0.16-1.84 mL/L) concentrations and pH (5.3-8.7), on dye biodegradation index. Pseudomonas oleovorans was used in biodegradation of Remazol Brilliant Blue R. Experimental results used to validate the proposed numerical approach. The best conditions have found at 0.15 g/L of dye, pH 8, 1.5 mL/L of microorganism and 1 g/100 mL of glucose, while a dye biodegradation index of 96% was achieved. The perceptron multilayer neural network has been efficiently in simulation of dye biodegradation on reduced conditions of data set.
