Yapay Sinir Aglari ile Eksenel Yüklü Kolonlarin Burkulma Analizi

Yıl 2006, Cilt: 2 Sayı: 2, 36 - 45, 01.08.2006

S. Kaya

Öz

Son yillarda, insaat mühendisligindeki bilgisayarli hesaplamalarda yapay zeka uygulamalari ilk sirayi
almistir. Bu uygulamalar genellikle uzman sistemleri içermektedir. Bu makalede yapay sinir aglarina
deginilmis ve bir uygulama yapilmistir. Eksenel yüklü kolonlar tasiyabilecekleri burkulma yükleri dikkate
alinarak tasarimlanirlar. Bu çalismada çesitli mesnet kosullari iç in eksenel yüklü kolonlarin burkulma yükünü
veren çok katmanli bir ag yapisi egitilmistir. Geriye yayilma egitim algoritmasi kullanilan çalismada dairesel,
kare, dikdörtgen ve I kesitli kolonlar incelenmistir. Dört farkli mesnet durumu iç in egitilen ag, veriler
karistirilarak dört farkli sinir kosulu için test edilmistir. Elde edilen sonuçlarin yeter duyarlilikta oldugu
görülmüstür. Mantiksal programlama tekniginin bu alandaki uygulama potansiyeli vurgulanmistir.

Anahtar Kelimeler

Burkulma, Ag yapisi egitilmesi, Yapay sinir aglari, Elastik kolonlar

The Buckling Analysis of Axially Loaded Columns with Artificial Neural Networks

Öz

Computation on Civil Engineering has concentrated primarily on artificial intelligence applications in the
past few years. These applications generally involve expert systems. This article deals Neural Networks and
applications were presented. Axially loaded columns are designed according to the their buckling load capacity. In
this study, a multi-layer artificial neural network is trained to give critical load for axially loaded columns and
various support conditions. Backpropagation training algorithms are used considering the circular, square,
rectangular, and I cross-sections. The artificial neural network, with is trained for circular , rectangular ,square and I
sections for four support conditions, is tested for the four support conditions. The results found using trained neural
networks are sufficiently close to the theoretical solution. It is emphasized that logical programming has application
potential in this area.

Anahtar Kelimeler

Buckling, Training of network topology, Arti_cial neural networks, Elastic columns