Abstract:
This manuscript aims to improve the efficiency of wastewater treatment at the Aïn Oulmène treatment plant by developing a predictive model to estimate the optimal coagulant dose used during the coagulation-flocculation process. This work is based on artificial neural networks application as an intelligent tool able of predicting the required chemical dosage based on raw water characteristics.
Data were collected over a twelve-week period and covered four key parameters: pH, total suspended solids (TSS), chemical oxygen demand (COD), and biochemical oxygen demand over five days (BOD₅). These data were processed and integrated into a model built using MATLAB 2020 a, with the actual doses applied (between 2.5-3 mg/L) used as the learning reference.
The results showed good agreement between the predicted and observed values, with a correlation coefficient of approximately 0.95, indicating satisfactory model accuracy. This approach represents a promising alternative for adjusting doses more flexibly and efficiently, while reducing costs associated with chemical use.
