Parametric optimization and modeling of continuous electrocoagulation process for the removal of fluoride: Response surface methodology and machine learning approach

The main objective of this study was to assess the continuous electrocoagulation process effectiveness for removing fluoride from potable water. The effect of different parameters like applied potential, electrode spacing, and feed flow rate was optimized for the continuous removal of fluoride from potable water. Response surface methodology (RSM) was used to examine the impact on essential operational factors such as voltage, concentration, and pH for fluoride removal as a response. The results demonstrate that all the parameters had a significant effect on removal efficiency. The quadratic model accurately predicted the optimal parameters for maximal fluoride removal efficiency with the association of desirability 1.0, which was discovered to be voltage 2.38 V, feed concentration 5.52 mg/L, and pH 6.45. According to the analysis of variance, R 2 of the proposed quadratic model is higher (0.9877). Moreover, the difference between the predicted R 2 of 0.9258 and the adjusted R 2 of 0.9767 was less than 0.2. The model adequacy was also studied based on residual plot, perturbation plot, and box-cox plot. The RSM was best modeling techniques use to predict data than the multilayer perceptron and linear regression due to high accuracy. Finally, the generated flocs were characterized by scanning electron microscopy, energy-dispersive X-ray, X-ray diffraction, and Fourier transform infrared spectroscopy instrumental techniques. The outcomes demonstrate that a newly designed continuous electrocoagulation process is a promising alternative for the removal of fluoride from potable water.