ABSTRACT
Recently hydrogen gas introduced as valuable gas which could be recovered from various streams and membrane science is one of the applicable and economically attractive way for the separation of hydrogen. In this present contribution, artificial neural network modeling was used to model the gas permeability of hydrogen at various operational condition and different composition of feed. Also, multi-layer perceptron (MLP) used as optimization algorithm for improving the accuracy of developed ANN model. Feed pressure, gas composition and weight percent of nano filler was used as input parameter and hydrogen selectivity is the output of the model. Modeling results show that ANN modeling shows best performance among the all models reported in the open literature. Moreover, statistical parameters resulted from ANN modeling compared with other artificial intelligence-based models. Coefficient of determination, root-mean-square error, and average absolute relative deviation of ANN modeling is 0.9984, 0.01184, and 4.85, respectively, which shows the accuracy and robustness of the developed model for the prediction of hydrogen selectivity.
Additional information
Notes on contributors
Jingwen Zhang
Jingwen Zhang is now a professor in basic Department of Chongqing Creation Vocational College in china.