Influence of channel-stud shear connector parameters on predicting the flexural behaviour of steel-concrete composite space truss: an artificial neural network based approach

सार

The steel-concrete composite structures with channel-stud shear connectors facilitate the attachment of Mero nodes by connecting the channel section and stud component. This study involves choosing the shear connector parameters and variables to analyze the steel-concrete composite space truss structure. There are a total of 54 distinct models for the parameter variables. The calibrated analytical model incorporates these factors using SolidWorks for modelling and then performs transient structural analysis with ANSYS. The research outcomes have demonstrated the influence of variables on the model using load-deflection and load-slip measurements. Modifying the channel length and thickness has significantly decreased deflection and slip in the shear connectors. However, the changes in the width and height of the channel found to be less impact on the overall performance. Artificial Neural Network (ANN) has been trained with the input and output data using the Levenberg-Marquardt backpropagation technique. For network training, testing, and validation, 70% and 30% of the analytical model results have been used, respectively. The performance of the ANN model has been assessed by data analysis and statistical methods, including Mean Squared Error (MSE) and Regression (R). The deflection and horizontal slip values predicted by the ANN model were remarkably accurate, closely matching the analytical results with a negligible margin of error.