|
چکیده
|
The buckling of steel columns is a critical concern in structural engineering design and analysis. Accurate prediction of buckling behavior is necessary for ensuring the integrity and safety of steel structures. Buckling phenomena in steel columns present a challenging and intricate issue within the realm of structural engineering. In the past few years, diverse artificial intelligence (AI) techniques have been employed to address complex problems in structural engineering. Artificial neural networks (ANNs) encompass a category within the field of AI that can learn complex patterns and relationships from datasets. This article endeavors to predict the buckling load in steel columns, addressing it as a complex problem in structural engineering. By training an ANN on a dataset that includes information about the parameters affecting buckling, such as column dimensions, material properties, and load conditions, it is possible to develop a predictive model. In this research, the behavior of steel columns under various loading conditions using finite element (FE) is simulated, a large amount of data for training ANNs have been generated, and multiple ANNs are trained using various architectures and training algorithms. The performance of trained ANNs is evaluated using statistical parameters such as mean squared error (MSE) and coefficient of correlation (R2). The results show that ANNs are well-suited for predicting complex and nonlinear problems such as buckling load in steel columns. The paper also discusses the importance of proper training and validation procedures and the challenges associated with extrapolation beyond the trained data range.
|