ABSTRACT
This study evaluates the effectiveness of an artificial neural network-genetic algorithm (ANN-GA) artificial intelligence (AI) model in the prediction of behavior and optimization of an electro-oxidation pilot press-type reactor, which treats a synthetic wastewater prepared with a dye. The ANN was built from real experimental data using as input the following variables: time, flow, j, dye concentration, and as output discoloration. The performance of the ANN was measured with MAPE (8.3868%), calculated from real and predicted values. The coupled AI model was used to find the best operational conditions: discoloration efficiency (above 90%) at j = 27 mA/cm2 and dye concentration of 230 mg/L.
Subject(s)
Artificial Intelligence , Neural Networks, Computer , Wastewater , Algorithms , Oxidation-ReductionABSTRACT
Comparative degradation of the industrial dyes Blue BR, Violet SBL and Brown MF 50â¯mgâ¯L-1 has been studied by the electrochemical oxidation (EOx), electro-Fenton (EF), photoelectro-Fenton (PEF) process based on BDD electrode. Each dye was tested in 0.05â¯mM Na2SO4 with 0.5â¯mM Fe2+ at pH 3.0, and electrolyzed in a stirred tank reactor under galvanostatic conditions with 2.0, 5.0, 7.0, 11.0 and 18.0â¯mAâ¯cm-2. Dyes were oxidized via hydroxyl radicals (OH) formed at the BDD anode from water oxidation coupled with Fenton's reaction cathodically produced hydrogen peroxide (H2O2). Under Na2SO4 medium close to 100% the decolorization was achieved. Through the color abatement rate the dyes behavior was analyzed at the beginning of the oxidation process. Dissolved Organic Carbon (DOC) was tested to evaluate the degradation. From DOC removal, it was established an increasing relative oxidation power of the EOxâ¯<â¯EFâ¯<â¯PEF, according with their decolorization trend. This study highlights the potential of the electrochemical/BDD process for the degradation of industrial dyes found in wastewaters under appropriate experimental conditions.