Remoção de cor verdadeira de efluente têxtil real por biossorção: aplicação e modelagem por redes neurais artificiais

Abstract

During the textile industrial production processes generate effluents with strong color, fact that causes concern, given that this high level of color might affect the process of light absorption by plants and animals, interfering with photosynthesis process. The present research work aims to study the biosorption process for removing color from a real textile effluent using the bagasse of sugarcane as biosorbent material and modelling the process using artificial neural networks. The sugarcane bagasse undergo with a pre-treatment and then there was made a factorial design 2² in order to evaluate the influence of chemical treatment of the biosorbent with sulfuric acid on the color removal process by adsorption. It was verified that the best conditions for color removal occurred when the sugarcane bagasse was received chemical treatment, previously, the next stages of this work was carried out with sugarcane bagasse treated sugarcane bagasse (BCM). Also, it was studied the effect of pH solution, particle size, adsorbent’s amount, mixing time and color’s concentration from the effluent of the adsorption process. The best conditions were found in the acid medium using material of 0.7 mm, a dosage of 0.6 g, which reached removals of 100% color. The data from the experiments were used to feed an artificial neural network (ANN) with multilayer perceptron (MLP) architecture. It was found that the ANN structure which best modelled the experimental data was composed of 8 neurons in the hidden layer, with the MSE value equivalent to 0.013. The expected ANN data for color removal were fitted to the experimental data with an R = 0.94 value, thus ANN can be used to perform predictions for the final color of the effluent. Nonlinear adjustments were made to kinetic models and adsorption isotherms. To characterize the biosorvent, physical-chemical analyzes, zero load point (pHpzc), scanning electron microscopy (SEM), infrared spectroscopy by Fourier transform (FTIR) and surface area were performed. In the continuous flow test, it was found that the breaking point of the adsorption column was obtained at 15 min of operation, where the final effluent coloration was equivalent to 7.37 PtCo.L-1, compatible with the legislation. In general, it can be concluded that the BCM had presented a promising method for the treatment of the textile effluent.