Predictive Machine Learning Framework for Optimizing ZnO-Modified TFN Membranes in RO Water Desalination
نویسندگان
1 UNESCO Chair on Water Reuse, School of Chemical Engineering, College of Engineering University of Tehran, Tehran, Iran
2 UNESCO Chair on Water Reuse, School of Chemical Engineering, College of Engineering University of Tehran, Tehran, Iran
3 School of Metallurgy and Materials Engineering, College of Engineering, University of Tehran, Tehran, Iran
4 School of Electrical and Computer Engineering, College of Engineering, University of Tehran, Tehran, Iran
5 Department of Biotechnology, Iranian Research Organization for Science and Technology (IROST), Tehran, Iran
6 UNESCO Chair on Water Reuse, School of Chemical Engineering, College of Engineering University of Tehran, Tehran, Iran
doi
10.22090/jwent.2026.2078566.2016چکیده
Zinc oxide (ZnO) nanoparticles are extensively incorporated into thin-film nanocomposite (TFN) membranes to improve separation efficiency, permeability, and antifouling performance in reverse osmosis (RO) desalination systems. In this study, a machine learning (ML) framework was developed to systematically model and predict the performance of ZnO-based TFN membranes under diverse fabrication parameters and operating conditions. A multilayer neural network architecture was constructed, and its reliability was verified using a repeated 5-fold cross-validation strategy to ensure robust generalization. Comprehensive hyperparameter tuning identified an optimal configuration of 13-7-7-9-2 neurons, which achieved a high coefficient of determination (R2 = 97.83%) and a remarkably low mean squared error (MSE = 0.00128). To enhance interpretability, SHAP analysis was applied to quantify feature importance, revealing that solvent type and applied pressure exert the most significant influence on water flux. The resulting ML model demonstrates strong predictive capability and provides actionable insights to guide the rational design, optimization, and performance forecasting of ZnO-enhanced TFN membranes for advanced water purification and desalination applications.
