Tertiary treatment using ultrafiltration in an existing sewage treatment plant for industrial reuse: a modelling approach using artificial neural network with uncertainty estimation

Navi Mumbai Municipal Corporation of Maharashtra state, India, unified a tertiary treatment plant (TTP) of 20 × 103 m3/day capacity with ultrafiltration technology in an existing Koparkhairane sewage treatment plant (STP) for producing effluent quality usable for industrial purposes. As prior art, an artificial neural network-genetic algorithm (ANN-GA) along with uncertainty estimation using prediction interval is employed to model secondary treated effluent (STE) flow rate (QT) and other quality parameters such as biochemical oxygen demand, chemical oxygen demand, and total suspended solids (TSS) to conclude the reliability of the range in which the input available to TTP. ANN-GA model provides a coefficient of determination above 0.90 for all STE parameters modelled other than TSS. Inferring that a good quantity and quality of 20 × 103 m3/day STP treated water is currently available, where a decreasing trend of QT is also noticed and highlighted. Further, the Wilcoxon signed-rank test on the quality parameter of effluent TTP for industrial reuse standard infers TSS shows infringement during the initial period but started adhering to standards over time. The research delineates at the outset of exploring water reuse policy in India, emphasizing Maharashtra state, modelling STE using ANN-GA and performance evaluation of TTP. HIGHLIGHTS Secondary treated effluent modelling using the artificial intelligence algorithm.; Artificial neural network integrated with the genetic algorithm is used in this study.; Uncertainty estimation using the prediction interval technique.; Ultrafiltration technology is checked for treatment compliance.; The Wilcoxon signed-rank test is used for compliance analysis.;