Removal of boron from mining wastewaters by electrocoagulation method: Modelling experimental data using artificial neural networks

Excess boron in drinking and irrigation water is a serious environmental and health problem because it can be toxic to many crops and lead to various diseases in humans and animals upon long-term consumption. In this work, the removal of boron from aqueous solutions was achieved by electrocoagulation using aluminium as the anode and cathode. The operating parameters influencing the efficiency of boron removal, namely, initial pH (pH0), current density, and treatment time, were investigated. An optimum removal efficiency of 70% was achieved at a current density of 18.75 mA/cm2 and pH0 = 4 within 90 min of treatment time. An artificial neural network (ANN) was utilised for modelling the experimental data. The model with a topology of 3-10-1 (corresponding to input, hidden, and output neurons, respectively) provided satisfactory results in the identification of the optimal conditions. The sum of squared error and correlation coefficient (R2) were 0.616 and 0.973, respectively, confirming the good fit of the ANN model.