Artificial intelligence-based control for membrane bioreactor in sewage treatment

Abstract

Recently, membrane bioreactors (MBRs) have emerged as a promising approach for sewage treatment because of their high efficiency in removing contaminants. However, they are prone to membrane-fouling and computational loading. To resolve these issues, this research article presents an innovative control strategy combining both artificial bee colony optimization (ABC) and recurrent neural network (RNN) to regulate the performance of MBR in sewage treatment. Initially, the influent wastewater data were collected and pre-processed using the regression imputation approach. RNN architecture was designed and trained using the pre-processed data to forecast the performance of the MBN system. Further, the ABC algorithm was applied to optimize the function of MBR by adjusting the control variables. The developed model was validated with the publically available wastewater treatment plan dataset and the effectiveness of the developed model was validated by performing intensive performance and comparative assessment. The performance evaluation demonstrates that the proposed methodology attained greater results of 98.59% effluent quality, 98.70% of nutrient removal efficiency, less computational time of 2.87 s, and a low membrane-fouling index of 1.23%. The comparative analysis illustrates that the presented approach achieved improved performances than the existing methodologies.