Thermal Conductivity Calculation at Moderate Pressure for Polyatomic Gases Using a Neural Network Approach
Bouzidi Abdelkader1, Khelfi Djillali2, Rebhi Fayçal3
1Dr. Bouzidi Abdelkader, Birine Nuclear Research Center  Oussera W. Djelfa, Algeria.
2Mr. Khelfi Djillali, Birine Nuclear Research Center  Oussera W. Djelfa, Algeria.
3R.Mr. Rebhi Fayçal, Birine Nuclear Research Center Oussera W. Djelfa, Algeria.
Manuscript received on January 22, 2013. | Revised Manuscript received on February 07, 2013. | Manuscript published on February 28, 2013. | PP: 404-409 | Volume-2 Issue-3, February 2013.  | Retrieval Number: C1069022313 /2013©BEIESP

Open Access | Ethics and Policies | Cite
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (

Abstract: The main aim of the present work was the development of a new method based on neural network to accurately evaluate thermal conductivity of pure polyatomic gases included both polar and no polar gases at atmospheric pressure and over wide range of temperature. Two multilayer feed forward neural networks have been trained using five and four physicochemical properties for polar and no polar gases respectively; molecular weight (M), boiling point (Tb), critical temperature (Tc), critical pressure (Pc) and dipole moment, for polar gases, combined with absolute temperature (T) as their inputs . The thermal conductivity and the properties for each individual gas were compiled on different temperatures, ranging from their boiling points to 1100 K. The maximum absolute error in thermal conductivity, predicted by the artificial neural networks ANNs, was less than 4%.
Keywords: Conductivity, Polar gases, No polar gases, Artificial neural networks.