Rotation Capacity Prediction of Open Web Steel Beams Using Artificial Neural Networks

Artificial neural networks (ANN) are artificial intelligence technologies used in various fields of science and engineering to represent fascinating human behaviours. Engineers frequently deal with incomplete and noisy data, which is one of the areas where neural network (NN) shine. Aim of this study is to use an ANN approach to determine the rotation capacity of open web steel beams. Using a single point load applied at the span's centre, a theoretical, experimental, and analytical study was conducted. Following the results of experimental and analytical comparisons, the ABAQUS software tool was used to assess a total of 88 nonlinear finite element models. Local slenderness ratios of several finite element models differentiate them. Different elements comprising geometrical and mechanical features of open web steel beams were delivered as input to NN models, including flange and web slenderness, depth and breadth of section, load span and angle section. Suggested formulation's accuracy is confirmed by arithmetical regression created using analytical nonlinear finite element modelling and behaviour of the open web steel beam derived analytically was tested experimentally. Based on research and statistical analysis, the current study found that ANN has a great potential for forecasting the rotation capacity of open web steel beams.