Abstract:
The purpose of this research was to study the application of a mathematic model for substrate removal of SBR system. The study was designed to compare the predictive values from the model with the values from the actual experiment. The model was derived from mass balance of semi-batch process during and after feeding periods. The Monod expression was used in the model. In the experiment, a high concentrated pineapple syrup from a certain factory was diluted together with other necessary macronutrients and used as influent. The experiment was divided into two phases. At the first phase kinetic parameters were determined in order to apply in the model by using a completely mixed activated sludge system with varying sludge age of 3, 5, 9, 15 days. At the second phase, the filtered TOC and MLVSS profiles occurring in a cycle of SBR operation were monitored in order to compare with the predictive values from the model. Each cycle period used in the SBR system was 6 hours, consisting of fill period of 60 minutes, react period of 210 minutes, settle period of 45 minutes and draw period of 45 minutes. The influent volume was 4 litres per cycle and the reactor volume at the end of fill was 10 litres. Kinetic parameters from the first phase of the experiment were Y = 0.58 g. VSS/ g. TOC, K[subscript d] = 0.052 d⁻¹, µ[subscript m] = 1.62 d⁻¹, K[subscript s] = 29.63 mg. TOC/1. The derived model was tested and compared with results from the second phase of experiment. It was found that the concentration of MLVSS was decreasing during fill period and fairly constant after the end of fill. These changes were concurrent between the actual experiment and the model. The changes of concentration of MLVSS were affected mainly by the dilution of influent while the kinetic parameters expecially Y and Kd slightly affected the changes. Filtered TOC from the actual experiment was increasing during fill period and was decreasing until fairly constant after the end of fill. Experimental results showed that filtered TOC from the actual experiment expecially during the fill period were much lower than values from the model. It was possible that the substrate was utilized faster in the semi-batch reactor than in the continuous system. This might be due to the fact that semi-batch reactor had higher substrate concentration gradient, therefore the rate of growth of microorganism and the rate of adsorption were consequently higher.