A New Deep Learning Accelerated Blast Loading Effect Analysis of FRP Retrofitted Concrete Slab

Abstract

Sudden rapid release of energy in blast loading results in complex material behaviour in the surrounding. The study on these destructive events and its serious impact is only increasing day by day due to its geopolitical importance. Traditionally concrete structures have been used for many blast-resistant structures. It is well known that retrofitting reinforced concrete (RC) slabs with exterior fiber-reinforced polymer (FRP) improves blast resistance, reduces displacements and cracking, and helps to collect debris. Even with modern sophisticated computational tools there are still issues with accurate and efficient prediction of the results through numerical simulation. In this work we propose a machine learning assisted blast loading analysis model than can predict maximum displacement of the FRP retrofitted reinforced concrete (RC) slab subjected to blast loading from the standard input parameters including the features of the blast loading, reinforced concrete slab as well as FRP retrofitting. The dataset is collected from the literature which includes several experimental as well as numerical data. A deep learning model is trained with these data and its performance is tested for prediction of the maximum displacement. The model is seen to perform satisfactorily within acceptable engineering errors. Another advantage is that after training the prediction is extremely fast and can save a lot of time. Also, if the training sample includes variations of cases across several possible scenarios the machine learning can be quite robust. 2520COMPDYN 2023 9 th ECCOMAS Thematic Conference on Computational Methods in Structural Dynamics and Earthquake Engineering M. Papadrakakis, M. Fragiadakis (eds.