Pistachio Processing Wastewater treatment by Peroxi-Electrocoagulation Method and Artificial Neural Network Modelling

Abstract In this study, the peroxi-electrocoagulation process was investigated for the treatment of pistachio processing wastewater. The electrolysis cell was configured with 5 × 5 iron plate electrodes at a fixed 5 mm spacing. Electrolysis time of 90 min, 25 °C temperature, and 200 rpm stirring speed were maintained constant throughout the experiments. The effects of H 2 O 2 dose (94–658 mM), initial pH (2.0–6.0), and current density (1–6 mA/cm 2 ) on the removal of TOC, COD, and total phenol (TPh) were evaluated. Optimum operation conditions within the studied range were found as 564 mM H 2 O 2 , 3.0 initial pH, 3 mA/cm 2 current density, resulting in 88.9% TPh, 80.7% COD, and 65% TOC removal efficiencies. Kinetic study showed that the PEC process followed second-order kinetics, with R 2 of 0.82, 0.82, and 0.80 for TPh, COD, and TOC, respectively. Reaction rate constants increased with increasing current density and H 2 O 2 concentration, whereas reaction rates decreased at pH below and above 3.0. The highest reaction rates were observed at pH 3.0, 6 mA/cm 2 CD, and 658 mM H 2 O 2 . A single hidden-layer backpropagation network with a topology of 4:6:3 was developed following hidden layer neurons optimization to predict TOC, COD, and TPh concentrations. The final ANN model performance yielded RMSE of 150.4 ± 31, 725.7.3 ± 156.9, and 197.2 ± 37.9, and R 2 of 0.98, 0.97, and 0.95 for TPh, COD, and TOC, respectively. The high correlation between measured and predicted concentrations (R 2 ; TPh:0.97, COD:0.97, TOC:0.92), confirmed the reliability of the model. Under selected conditions, the specific energy consumption and operating cost were calculated as 16.2 kWh/kg COD and 65.9 $/m 3 , respectively. Graphical Abstract