CNN-LSTM Hybrid Model for Predicting Strength and Durability of Steel-Concrete Composite Structures

Steel-concrete composite structures are widely used in industrial and civil infrastructure due to their better structural performance. Composite constructions let engineers maximise the benefits of each material, especially concrete and steel. These materials must be integrated efficiently to optimise structural performance, safety, and durability. A fibre beam-column model of steel-concrete composite constructions incorporates round-end section analysis. An analytically-based fibre concretisation program enhances the analysis. To assure data reliability, preprocessing processes use Z-score normalisation. Also, the study uses AE technology to explore FRP/steel-concrete composite beam damage modes and critical failure circumstances. The SCCS method uses the CCSO-LSTM model and CNN feature extraction. Optimised deep LSTM training detects network steel and improves steel performance. The suggested SCCS method is efficient, with CNN feature extraction boosting performance. Comparisons with existing models reveal the suggested model's greater accuracy. Deep learning with optimal training achieves the greatest accuracy of 95.67%. Advanced computational models are used to analyse and improve steel-concrete composite constructions in this work. Fibre-based section analysis, AE technology, and deep learning algorithms improve structural assessment and failure prediction. To optimise composite structure performance, computational intelligence and structural engineering must be combined.