An Algorithm for Selecting RBF-ANN Centers of Species Mass in Engine

Author

Guo, Xiao-Ping ; Wang, Zhan-Jie

Author_Institution

Sch. of Energy & Power Eng., Dalian Univ. of Technol., Dalian, China

Volume

3

fYear

2009

fDate

March 31 2009-April 2 2009

Firstpage

40

Lastpage

44

Abstract

After analyzing the features of species mass in engine combustion process, considering the accuracy and calculating time together, based on a network for one specie, this paper proposes an algorithm for selecting RBF-ANN centers with sectional method to apply in engine combustion species mass: in the image of species mass fraction-temperature, mass fraction area is divided into N parts dynamically, temperature area is divided into 6 fixed parts, then selects a sample as RBF-ANN center in every overlapped area of vertical (mass fraction) area and horizontal (temperature) area. Through calculating 135 diesel combustion process, it shows that the accuracy of function approximation has been improved greatly with sectional method, compared with subtractive cluster method.

Keywords

diesel engines; function approximation; mechanical engineering computing; neural nets; radial basis function networks; RBF-ANN; diesel combustion process; engine combustion process; engine combustion species mass; function approximation; sectional method; Accuracy; Algorithm design and analysis; Combustion; Diesel engines; Function approximation; Image analysis; Temperature; Artificial neural network; Chemical species; Combustion; Function approximation; RBF;

fLanguage

English

Publisher

ieee

Conference_Titel

Computer Science and Information Engineering, 2009 WRI World Congress on

Conference_Location

Los Angeles, CA

Print_ISBN

978-0-7695-3507-4

Type

conf

DOI

10.1109/CSIE.2009.374

Filename

5170797