OPTIMIZATION BY ARTIFICIAL NUERAL NETWORK OF A FLUIDIZED BED FENTON SYSTEM FOR THE TREATMENT OF REVERSE OSMOSIS CONCENTRATE

Abstract

Fluidized-bed Fenton (FBR-Fenton) is a novel Fenton-type advanced oxidation process, which consists of various design and operation components. In this study, operational parameters (HRT, chemical dosages and bed expansion) of FBR-Fenton were modelled with artificial neural network (ANN), for the degradation of reverse osmosis concentrate (ROC). The performance of the model in predicting total organic carbon (TOC) removal and sludge production was evaluated and compared with conventional modelling techniques. The ANN model achieved a high accuracy for predicting TOC removal (RMSE = 0.0263; R2 = 0.8680) and sludge production (RMSE = 0.0355; R2 = 0.8182) under different operation conditions. Subsequently, the design parameter (carrier use) was studied to further enhance the degradation performance of FBR-Fenton. GAC had a better removal efficiency (DOM: 66.7%, COD: 67%) compared to conventional sand carrier (DOM: 47%, COD: 33%). Novel mechanism was subsequently proposed based on literature review and experimental findings for FBR-Fenton/GAC system.