FATIGUE LIFE CALCULATION OF FLAWED STRUCTURES - BASED ON ARTIFICIAL NEURAL-NETWORK WITH SPECIAL LEARNING SET

Authors
Citation
Yl. Han et al., FATIGUE LIFE CALCULATION OF FLAWED STRUCTURES - BASED ON ARTIFICIAL NEURAL-NETWORK WITH SPECIAL LEARNING SET, International journal of pressure vessels and piping, 75(3), 1998, pp. 263-269
Citations number
11
Language
INGLESE
art.tipo
Article
Categorie Soggetti
Engineering, Mechanical",Engineering
ISSN journal
0308-0161
Volume
75
Issue
3
Year of publication
1998
Pages
263 - 269
Database
ISI
SICI code
0308-0161(1998)75:3<263:FLCOFS>2.0.ZU;2-8
Abstract
A method of calculating the fatigue life of flawed structures, by usin g an artificial neural network with the help of a special learning set , is discussed. The special learning set is composed of manmade exampl es and experimental ones including selected experimental fatigue data for two kinds of specimens. The neural network learns information abou t the fatigue life of flawed weldments from the special learning set f irst; then the trained neural network is employed to calculate the fat igue Life of flawed specimens. The calculated lives are compared with experimental lives; this shows that the neural network and special lea rning set can calculate the fatigue Life of flawed weldments quite sat isfactorily. The network is also trained by using the normal learning set and tested by using the same samples to check the effect of the sp ecial learning set. (C) 1998 Elsevier Science Ltd. All rights reserved .