ADSORPTIVE REMOVAL OF HEXAVALENT CHROMIUM USING RESPONSE SURFACE METHODOLOGY AND ARTIFICIAL NEURAL NETWORK

Abstract

Fe–SCC adsorbent was used in a batch procedure to remove Cr(VI) from an aqueous solution. An RSM and an ANN model were developed using data gathered from 30 batch trials, which were then utilized to optimize and accelerate the absorption processes. As a result of a three-level, four-factors central composite design (CCD) in RSM, the impacts of operational factors such as Cr(VI) concentration, contact duration, the dosage of adsorbent, and pH of solution were evaluated. The suggested quadratic model had a coefficient of determination (R 2 ) value of 0.996 and a Fisher F-value of 264.18, which indicated an excellent match of the experimental data. When it came to figuring out how important the various variables were in determining the best process conditions, response surface plots came in handy. Assuming ideal operating circumstances, the maximum removal of Cr(VI) was determined to be 98.3% when the test variables stayed unchanged at a maximum desirable value of 0.978: 20 mg/L initial Cr(VI) concentration plus 0.1 g Fe–SCC dosage, pH 8, and 13 min of contact time. The same architecture was used to construct an ANN model that predicted Cr(VI) adsorption with acceptable accuracy (R 2 = 0.962). The R 2 coefficient of determination and the order of relevance of the operational parameters were used to compare the two models. The experimental datasets were well–suited to both models, as seen by the overall findings.