Chem. J. Chinese Universities ›› 1997, Vol. 18 ›› Issue (6): 886.

• Articles • Previous Articles     Next Articles

Application of Artificial Neural Networks for Hydrocarbon Gas Mixture Analysis

LIN Yong-Jing, ZHU Er-Yi, LI Quan-Long, YANG Peng-Yuan   

  1. Department of Chemistry, The SEDC Research Laboratory of AnalyticalScience for Material and Life Chemistry, Xiamen University, Xiamen, 361005
  • Received:1996-07-03 Online:1997-06-24 Published:1997-06-24

Abstract: An array composed of sixtorganiceen metal oxide semiconductor gas sensors was constructed to analyze gas mixtures quantitatively. The responses of the sensor array to ethane, propane and propylene were treated by three-layer artificial neural networks (ANN)with the method of error back-propagation and partial least-squares (PLS)- The pattern recognition results indicated that the concentration predicted with ANNis better than that with PLS. The average prediction errors for ethane, propane and propylene were 5. 11%, 8.28%, 2. 64%, respectively, in the ANNprediction.

Key words: Artificial neural networks, PLS, Sensor array, Modeling

TrendMD: