Identification of the most suitable probability models for local joint flexibility in T/Y-connections stiffened with collar or doubler plates

The reliability-based design of steel structures requires accurate probability density functions (PDFs) of the relevant random variables. Also, the local joint flexibility factor (fLJF), as an inherent feature of a connection, is one of the factors affecting the dynamic responses and global static of a steel structure. This study proposes two theoretical probabilistic models for predicting the fLJF of T/Y-connections retrofitted with either collar plates or doubler plates. The present research work used the results of 320 finite element analyses on the retrofitted T/Y-joints under in-plane bending. The data generated from these simulations provided a robust foundation for probabilistic modeling. Twenty PDFs were fitted to the resulting histograms. Among the tested models, the Kumaraswamy distribution emerged as the most accurate for describing the fLJF behavior of T/Y-connections retrofitted with doubler plates. Also, for the joints with collar plates, the Generalized Gamma (4P) distribution is the best model. Literature survey showed that no study was conducted on the probabilistic analyze of fLJF in any type of reinforced joints. Hence, the present findings lead to introducing a novel perspective for assessing and enhancing the structural resilience in engineering practices.