Analyzing the influencing factors and developing Artificial Neural Network-based prediction model for water turbidity

The present study investigated the impact of settling time on the turbidity of treated water using a systematic approach by coupling jar test experiments, micro-scale investigations, and artificial neural network (ANN) modelling for polyalumunium chloride and Moringa oleifera coagulants. Nature-based materials like Moringa Oleifera are recognized for water treatment for their environmental sustainability, cost-effectiveness, simplicity and flexibility, and community engagement. The efficacy of the treatment process is dependent on the dosage, which is usually optimized through jar test experiments on a daily basis and imparts resource, time and energy wastage. This necessitates the development of artificial intelligence models that can predict turbidity removal based on influencing parameters, namely dosage, pH, and initial turbidity. However, the settling time also affects the turbidity removal, but has not been considered while modelling turbidity removal prediction. Thus, the study examined the characteristics of flocs and their rate of change during settling, elucidating the mechanism of coagulation and the floc dynamics. We developed three predictive models to validate the hypothesis: a linear regression model, a non-linear regression model, and an ANN model for two scenarios: Scenario I, which used initial turbidity, pH, and coagulant dose to predict final turbidity; and Scenario II, which also included settling time. The results indicated that the removal of turbidity and the resulting final water turbidity exhibit non-linear relation with settling time. The inclusion of settling time in the models improved their predictability. Among the three models, the ANN model demonstrated superior performance, achieving an R 2 value of 0.99 for polyaluminium chloride and 0.87 for Moringa oleifera . The ANN model for scenario II was less sensitive to fluctuations in pH, initial turbidity, and dosage for both coagulants.