Column Buckling Mode Classification Using a Back Propagation Neural Network

by Paul C. Xirouchakis, George Mason Univ, United States,
Eugene M. Norris, George Mason Univ, United States,



Document Type: Proceeding Paper

Part of: Electronic Computation

Abstract:

A back propagation neural network (Rumelhart, Hinton and Williams, 1986) is used to analyze and classify the buckling mode shapes of structural columns. The input patterns consist of vectors of digitized nondimensional deflection points of the buckling mode shapes of columns. The output consists of the mode number and the associated effective column length ratio as predicted by the network. The results of this study are of interest in the area of numerical finite element post-processing and results interpretation of structural mechanics applications. In particular, the objective of this paper is to investigate the column buckling mode and associated compressive buckling load classification potential of a back propagation neural network.

Subject Headings: Neural networks | Buckling | Structural analysis | Failure analysis | Finite element method | Numerical analysis | Network analysis

Services: Buy this book/Buy this article

 

Return to search