Research in Elman Neural Network for AFR Model of Automotive Engine

被引:0
作者
Bu, YuHong [1 ]
机构
[1] Shanghai Univ, Dept Mech Engn, Shanghai, Peoples R China
来源
ADVANCED RESEARCH ON INDUSTRY, INFORMATION SYSTEMS AND MATERIAL ENGINEERING, PTS 1-7 | 2011年 / 204-210卷
关键词
Elman neural network; AFR Model; en-DYNA Simulation;
D O I
10.4028/www.scientific.net/AMR.204-210.755
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Air fuel ratio is a key index affecting power performance and fuel economy and exhaust emissions of the gasoline engine, whose accurate model is the foundation of accuracy air fuel ratio control. In the paper, at first, it has studied the Elman neural network (NN) simulation model of Air Fuel ratio physical model of automotive engine. Second, employing the S1-V8 in en-DYNA engine model as experimental device, the paper discussed the structure determination of Elman neural network; finally, it compared model identification performance between Elman and BP neural network. Experiment results show the generalization performance of neural network does not have a linear relationship to the neurons in hidden layer of Elman NN, and the air fuel ratio based on Elman neural network is better than the air fuel ratio model based on BP neural network. The average relative error of Elman NN air fuel ratio model is less than 0.5%, however, which of BP NN is more than 1%.
引用
收藏
页码:755 / 759
页数:5
相关论文
共 7 条
[1]  
Ge Zhexue, NEURAL NETWORK THEOR
[2]  
GUO AW, 2008, P 7 WORLD C INT CONT
[3]  
SHI WF, 2004, P 5 WORLD C INT CONT
[4]  
SONG JG, 2002, J CIVIL AVIATION U C, V20
[5]  
TAO T, 1997, ACTA AERONAUTICA AST, V18
[6]  
XIA GQ, 2009, 2009 WORLD C COMP SC
[7]  
YU Q, 2008, 2008 INT SEM FUT BIO