Evaluation of Compression Member Buckling and Post-Buckling Behavior Using Artificial Neural Network

There are various ways to investigate the buckling phenomenon in a steel compression member. Among these methods experimental formulas, semi-empirical formulas, closed form solutions, classic methods based on differential equations and also numerical methods based on finite element are mostly used. All of the mentioned methods have some complexities with their process. Thus, it is very difficult to perform a comprehensive parametric study on buckling phenomenon of a compression member using these methods. The simplicity of mathematical concepts used in ANN (artificial neural network) and its ability in modeling complex problems made ANN a popular facility. After learning ANN, it is able to introduce a function as a relationship between input variables and output parameters, which is load-displacement relationship in this article. Then it is possible to perform an extensive parametric study on the buckling behavior of the compression member. In this research, this network is also able to extract the critical loadslenderness relationship. Using this network it is possible to conduct an accurate study on the effects of various parameters at critical load, and also evaluate the sensitivity of critical load versus each variable that is used in ANN training.

Keywords:

Buckling , Post-Buckling , column , Neural Network , Sensitivity Analysis