سال انتشار: ۱۳۹۱

محل انتشار: اولین کنفرانس بین المللی نفت، گاز، پتروشیمی و نیروگاهی

تعداد صفحات: ۱۰

نویسنده(ها):

p Emrani – Department of Chemical Engineering, Tehran University, Tehran, Iran
S.M Hosseini – Department of Chemical Engineering, Arak University, Arak, Iran
A. Mohadjer – Department of Chemical Engineering, I.A.U., Mohajeran Branch
F Parvizian – Department of Chemical Engineering, Arak University, Arak, Iran

چکیده:

The production of oxygen from air to support fuel combustion is important for many industrial processes in order to produce heat in heat consumingdevices. Mixed-conducting ceramic membranes (BSCF) for air separation provide an effective approach and are studied in this paper. Both anempirical Artificial Neural Network (ANN) model and a mechanistic model were considered for the prediction of oxygen permeation and theirpredictions were compared to experimental measurements. For the development of the ANN models, data sets were collected from actualexperiments and used to train the network. Aftertraining, the models were tested by unseen data to evaluate their accuracy and trend stability. ackpropagation with various training algorithms such as Scaled Conjugate Gradient (SCG), Levenberg- Marquardt (LM) and Resilient Backpropagation (RP) methods was used. It was found that both the empirical and the theoretical based approaches are effective in modeling membrane performance andboth compare well with experimental measurements. An important advantage of the ANN approach, however, is that it does not require any theoreticalknowledge or human experience during its preparation.