Experimental Verification of Nonlinear Analysis of Planar Truss under Static Loading by Neuro-modelers

Please enable javascript in your browser.

Pashoutani, Sepehr | 2013

499 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 45789 (09)
University: Sharif University of Technology
Department: Civil Engineering
Advisor(s): Joghataie, Abdolreza
Abstract:
In the recent years, neural networks have been used to solve many problems in different fields, specifically in engineering problems. Neural networks have different applications in material science and structures including modeling stress-strain relationship of composite materials and static and dynamic analysis of structures. Recently, a new method called neuro-finite element has been developed for nonlinear static analysis of structures manufactured by one-dimensional elements. In this method, by investigating the load-deflection diagram of each element, it is possible to analyze and produce a load-deflection diagram for the whole structure using neural networks. This experimental study tries to verify the mentioned method and specify its points of strength and weakness by conducting experiments on planar trussesand their members with highly nonlinear load-deflection behavior. Then neural networks will be trained well to model behavior of membersin order to obtain load-deflection diagram of corresponding truss
Keywords:
Neural Network ; Static Analysis ; Experimental Verification ; Experimental Evaluation ; Neuro-Finite Elements Method ; Nonlinear Analysis

No TOC