Evaluating urban wastewater remediation efficiency of the Hydroponic Vetiver System through predictive modelling using Artificial Neural Network

Abstract

Urban wastewater management requires alternative cost-efficient, energy-efficient sustainable technologies as conventional electro-mechanical engineering systems are unable to keep pace with the wastewater infrastructure requirements of the rapidly urbanizing world, especially in developing countries. Sustainable urban wastewater management remains a distant goal as none of the existing systems address the treatment of the urban water bodies, which are the last mile of the urban wastewater journey. Literature review shows that among the various potential decentralized solutions, the Hydroponic Vetiver system (HVS) shows good performance in remediation of different types of wastewater. The paper aims to evaluate the general quantitative wastewater remediation potential of the HVS by formulating a predictive model using Artificial Neural Network (ANN). The water quality parameters (WQP) under discussion are Biochemical oxygen demand (BOD) and Chemical oxygen demand (COD). The four significant predictor variables of inlet concentration, plant density, hydraulic retention time, and pH value were analysed from secondary sources of laboratory experiments conducted worldwide. The trained, tested, and validated neural network models for BOD and COD showed high predictive accuracy and goodness-of-fit. The HVS predictive model will enable municipal policymakers, urban planners, and water managers to measure the wastewater remediation potential of any given urban water body. • Urban water body management is the key to sustainable urban wastewater management. • Hydroponic application of Vetiver is effective for phytoremediation of wastewater. • Predictive model of the Hydroponic Vetiver System using ANN is valid and precise. • Degree of urban water body remediation using HVS can be planned and monitored.