Research Information System
International Journal of Environmental Science and Technology, 2024 (SCI-Expanded)
Conventional and microbubble pump dissolved air flotation systems have become proven technologies in the removal of oil, grease and suspended solids. The treatment performance of microbubble pump dissolved air flotation systems is high, and they offer advantages such as requiring less space, requiring less mechanical equipment and easier operation. Therefore, they are particularly favored for the pre-treatment of wastewater from milk and dairy plants, slaughterhouses, metal processing and coating plants, food and beverage factories and the petrochemical industry. In the presented study, the performance of conventional and microbubble pump dissolved air flotation systems was compared and the effect of influent flow rate and pressure changes on the performance was investigated. The treatment efficiency was determined based on oil/grease, organic matter, and suspended solids removal. An artificial neural network was created using the obtained data and variables. The neural network was trained using the Levenberg–Marquardt backpropagation algorithm. When all outputs were considered in relation to the prediction performance of the designed neural network, a match was found, as demonstrated by a high correlation coefficient (R = 0.98966), with a mean error of 7.330% between the predicted and actual (obtained in the field) values. As a result, when comparing the values obtained from the field with the values predicted by the model for the concentrations of oil/grease, organic matter, and suspended solids, it was found that the artificial neural network presented has a good predictive capability.