(583aa) Adsorption and Kinetic Studies of Using Entrapped Sewage Sludge Ash in the Removal of Chemical Oxygen Demand from Domestic Wastewater, with Artificial Intelligence Approach

Huge amounts of sludge are produced daily from municipal wastewater treatment plants. These solid organic wastes produce active ash after complete combustion which can be used in many applications. This study explores different adsorption and kinetic models that can describe the adsorption process of chemical oxygen demand (COD) into sewage sludge ash (SSA) and implementation of the obtained result in original domestic wastewater sample. The SSA and stander COD solution were prepared in the laboratory. X-Ray Diffraction (XRD) was used for SSA characterization. The effect of SSA in alginate on COD removal was studied at different absorbent dose, contact time, stirring rate, pH, and initial COD concentration. The results indicated that SSA is effective in the removal of COD from aqueous solution, where removal efficiencies of 61 and 91% were achieved for 800 and 100 mg/L initial COD concentration, respectively, after 60 min of contact time using dose 8 g/L at pH 5 with fixed stirring rate 200 rpm. The equilibrium isotherm of COD was determined using Freundlich, Langmuir, Koble–Corrigan, Toth, Hill, Khan, Redlich and Elovich models. The kinetic adsorption of COD onto SSA was determined by Avramin, pseudo first order, Intraparticulate, and pseudo second order models. The adsorption data of COD fitted well to Hill isotherm (lowest error summation: 7.81) and Avramin kinetic model (lowest error summation: 2.14). A removal efficiency of 63% was achieved for initial COD concentration of 167 mg/L at previously mentioned optimum conditions. Artificial neural network (ANN) with a structure of 5 – 10 – 1 was used to predict the COD removal efficiency. The proposed ANN was found to be effective in simulating the performance of SSA for COD removal, where a high R-value was recorded to be 0.99095, 0.99995, and 0.99999 for training, validation and test plots, respectively.