Artificial intelligence -driven insights into bisphenol A removal using synthesized carbon nanotubes

Water quality nowadays, under climate change, has become a risk and challenging problem to save water from deterioration. Advanced solutions such as nanomaterials and artificial intelligence for simulation have become some of the best and essential solutions. Therefore, this study assessed the artificial intelligence models' accuracy in simulating the elimination of Bisphenol A (BPA) using synthesized carbon nanotubes (CNTs). We concluded that the pseudo-second-order model's (R2) correlation coefficient is (0.999) significantly higher than the other models. Because the findings between the Model and Actual Values are so accurate, the adsorption of BPA on CNT could be modeled using the pseudo-second-order model, qe = 144.928(mg/g) and K2 = 0.0016. The correlation coefficient of Pseudo-First-Order model's (R2) is (0.825) qe = 27.107(mg/g) and K1 = 0.0161, and the Intraparticle diffusion model's (R2) is (0.821),qe = 151.98(mg/g) and Kd = 2.4. The Langmuir model performed the best in isothermal experiments, with correlation coefficients of R2 = 0.9441, qm = 181.81, and RL = 0.0375. Based on the information provided, we may conclude that the Langmuir model accounts for more BPA adsorption than the other models. We employed the feedforward neural network (FFNN) and the recurrent neural network (RNN). The FFNN achieved a coefficient of 0.971, while the RNN obtained a higher correlation coefficient of 0.98. The proposed methodology in this study. [Display omitted] •Water quality preservation is crucial amidst climate change, necessitating advanced solutions like nanomaterials and AI.•This study evaluated the accuracy of AI models in simulating Bisphenol A (BPA) elimination using carbon nanotubes (CNT).•This study employed the feedforward neural network (FFNN) and the recurrent neural network (RNN).•The FFNN achieved a coefficient of 0.971, while the RNN obtained a higher correlation coefficient of 0.98.