Neural networks simulation of thin-film nanocomposite membrane for enhanced water treatment

This paper simulated of flux and salt rejection the nanocomposite membrane process by back-propagation neural network with Levenberg–Marquardt training algorithm. Network with one hidden layer was optimized among several types of networks. ANN model and experimental data were compared .The results demonstrate that there are less error (MSE = 1.69E−5) and high relationships (R2 = 0.9996) between the experimental data and the predicted face value. As well as that sensitivity analyses to make known that the input nanoparticle is the most sensitive parameter on the output flux and rejection. As a consequence, the aim ANN model can be used to simulate and optimize the nanocomposite membrane process

Keywords:

Nano composite membrane , neural networks , Simulation , Levenberg–Marquardt algorithm , Flux , Rejection