Evaluating the Structural Integrity of Cold-Formed Steel Profiles Under Combined Loading Conditions: An Advanced Analytical Approach to Load-Bearing Capacity Prediction

Cold-formed steel (CFS) beams are structural components created from thin sheets of steel that are shaped into beams by room-temperature cold-forming techniques. The refined cross-sections of these profiles cause buckling issues. Conducting experimental and numerical investigations of this behavior is challenging and time-consuming. In this paper, evaluating the structural integrity of CFS profiles under combined loading conditions, an advanced analytical approach to load-bearing capacity prediction (CFS-CLC-LBC), is proposed. First, input data are gathered from the bearing failure dataset. Then the collected information is pre-processed by the multiparticle Kalman filter (MKF) and it is used for data Normalization from the dataset. The pre-processed data are subsequently fed into a sparse spectra graph convolutional neural network (SSGCNN) to predict the LBC of CFS profiles. By then, the proposed CFS-CLC-LBC technique is implemented in the Python working platform and the execution is calculated with the present procedure. The proposed technique shows findings in all existing methods like support vector machines (SVM), artificial neural networks (ANNs) and feed-forward neural networks (FFNNs). The proposed CFS-CLC-LBC method has an accuracy improvement of 18.75%, 26.89% and 32.57%. The [Formula: see text] values are higher by 24.57% and 32.94%, and root mean square error (RMSE) is reduced by 18.43%, 25.64% and 31.40%, compared to the existing approach. The proposed method demonstrates significant improvements over existing approaches in accuracy, reliability and precision. It achieves better model fit with higher [Formula: see text] values and enhanced predictive performance. Additionally, it reduces error rates, showcasing its robustness and effectiveness.