Title :
Gas/Liquid Two-Phase Flow Regime Recognition by Combining the Features of Shannon´s Entropy with the Improved Elman Network
Author :
Liu, Jing ; Han, Jun ; Dong, Feng
Author_Institution :
Dept. of Educ. Technol., Capital Normal Univ., Beijing, China
Abstract :
This paper investigates a new method for gas/liquid two-phase flow recognition by combining the features of Shannon´s entropy and the recurrent neural networks. The information of the method that provided by cross-sectional measured resistance Information (CSMRI) is the measured data in horizontal pipe. The feature vector of Shannon´s entropy that can express the essential information of gas/liquid two-phase flow is constructed and the improved Elman neural network is used to recognize the gas/liquid two-phase flow regime. The Elman neural networks with the context units can memorize the past input and express the change of flow process. The obtained results indicate that the method is suitable to estimate the flow regime and the higher recognizing rate of the flow regime is obtained.
Keywords :
computational fluid dynamics; pipe flow; recurrent neural nets; two-phase flow; Elman neural network; Shannon entropy; cross-sectional measured resistance Information; feature vector; gas-liquid two-phase flow regime recognition; recurrent neural network; Conductivity measurement; Electrical resistance measurement; Electronic mail; Entropy; Feature extraction; Fluid flow; Fluid flow measurement; Neural networks; Recurrent neural networks; Voltage; Elman network; Gas/liquid two-phase flow; Shannon´s entropy; recurrent neural networks; regime recognition;
Conference_Titel :
Natural Computation, 2009. ICNC '09. Fifth International Conference on
Conference_Location :
Tianjin
Print_ISBN :
978-0-7695-3736-8
DOI :
10.1109/ICNC.2009.720
