Comparative study of sonocatalytic process using MOF-5 and peroxydisulfate by central composite design and artificial neural network
The US/MOF-5/S2O82- system was used to degradation of Acid Blue 7 (AB7). The study was performed by the central composite design (CCD) and the obtained data were compared with the artificial neural network (ANN) results. The impacts of the main operational variables, like the initial AB7 and S2O82- concentrations, MOF-5 dosage, and process time on dye degradation efficiency (DE%), were investigated. The development of the ANN model was done by feed-forward back-propagation (BP) network along with topology of 4: 8:1, and trainbr algorithm. An acceptable agreement between the predicted amounts was attained for the DE% calculated by the CCD and ANN; the observed results exhibited correlation coefficients (R2) of 0.97 and 0.95, respectively. The findings revealed that the CCD model enjoyed slightly better accuracy for description of the nonlinear behavior of the sonocatalytic process. Based on Pareto analysis, the reaction time, initial AB7 concentration, S2O82- concentration, and MOF-5 dosage were found to be the most significant parameters on DE%. The maximum DE% (98%) was achieved at the optimized experimental conditions, including the initial 29.5 mg L-1 AB7 and 10 mg L-1 S2O82- concentrations, 0.8 mg L-1 MOF-5 dosage, and a 60-min process time.