Development of a probability distribution model for the SCFs in tubular X-connections retrofitted with FRP

Abstract

In this study, results of performing 324 FE analyses on 108 circular hollow section (CHS) X-connections with fiber reinforced polymer (FRP) have been used to derive a probability distribution model for SCFs in the retrofitted joints under axial, IPB, and OPB loads. As far as the authors are aware, no study has been conducted on the probability distribution of SCFs in X-connections with FRP. However, the probability distribution models are essential for the reliability-based analysis and design of offshore structures. Therefore, based on the retrofitted X-connection FE models which were verified with several experimental tests, a set of SCF sample databases was generated. In these SCF sample databases, different FRP properties (layer number and material) and connection geometry (β, τ, and γ) are included. 10 various probability density functions (PDFs) were fitted to the relative frequency histograms of the SCF samples. For obtaining the parameters of each distribution, the maximum likelihood method was used. Assessing the goodness-of-fit based on the Chi-squared and Kolmogorov-Smirnov tests shown that the Log-normal distribution is the best one for the maximum SCF values in CHS X-connections retrofitted with FRP subjected to axial load. In addition, the Log-logistic and Generalized Extreme Value distributions were set as the governing probability function for the retrofitted X-connections under IPB and OPB loads, respectively. In the last step, five fully defined PDFs were established for the maximum SCFs in CHS X-joints retrofitted with FRP under three various types of loadings.