Artificial Neural Network and Response Surface Design for Modeling the Competitive Biosorption of Pentachlorophenol and 2,4,6-Trichlorophenol to Canna indica L. in Aquaponia

Abstract

The continuous exposure of the environment to carcinogenic wastes and toxic chlorophenols such as pentachlorophenol (PCP) and 2,4,6-trichlorophenol (TCP) resulting from industrial production activities is become a great concern. The search for cost efficient and ecofriendly approach to phytoremediation of water will guarantee sustainability. The present research work is concerned with cost benefit evaluation, and the optimization modeling of the competitive biosorption of PCP and TCP from aqueous solution to Cana indica. L (CiL-plant) using response surface methodology (RSM) and artificial neural network (ANN) model. The predictive performances of the ANN model and the RSM were compared based on their statistical metrics. The antagonistic and synergetic effect of significant biosorption variables (pH, initial concentration, and exposure time) on the biosorption process were studied at p-values ≤0.005. The optimized output transcends to PCP and TCP removal rates of 90% and 87.99% efficiencies at predicted r-squared ≤0.9999, at 95% confidence interval. The cost benefit evaluation established that at the optimum conditions, the cost of operating the removal of TCP from aqueous solution will save $ 7.72 compared to PCP. The reliability of the optimization model based on design of experiment was proven to be more sustainable compared to the one-factor-at-a-time methodologies.

Virtual chemical analysis and machine learning-based prediction of polyethylene terephthalate nanoplastics toxicity on aquatic organisms as influenced by particle size and properties

“…The hidden layer's nodes include mathematical functions that define the relationships. Once the connections (mathematical functions) have been es tablished, the tes ting data will be used to validate them [38].…”

“… (BA) es t : is the es timated value of the binding affinity in kcal mol -1 by ANN model;  (BA) exp , is the experimental value of the binding affinity in kcal mol -1 ; and  : is the average value of binding affinity in kcal mol -1 The goal should be to achieve the lowes t error with (RMSE, MAE, and SSE) and the greates t with (R 2 ) correlations to obtain the optimal ANN model [37,38].…”

Cefixime removal via WO3/Co-ZIF nanocomposite using machine learning methods

“…We also use optimization algorithms such as GA and SOLVER to find the best combination of parameters for achieving the highest removal efficiency of cefixime 32 . The optimized conditions identified through this research can be directly applied in real-world settings, aiding environmental remediation efforts and solving water pollution problems caused by pharmaceutical pollutants 33 , 34 . By accurately predicting the optimal conditions, we can reduce the consumption of catalysts, energy, and time required for the photocatalytic process.…”