NUMERICAL ANALYSIS OF SOCKET CONNECTION IN TWO PILE CAPS: RECOMMENDATIONS FOR MODELING AND PRECISION LEVEL

Non-linear finite element analysis (NLFEA) is an advanced computational technique that has proven effective in enhancing the accurate assessment of structural strength, leading to more efficient and reliable design outcomes. However, despite its potential, the implementation of NLFEA in the design of reinforced concrete (RC) socket connections has progressed at a relatively slow pace, mainly due to limited quantification of uncertainties associated with the modeling process and the absence of comprehensive sensitivity analyses of input parameters. This paper proposes a robust framework for applying NLFEA to RC socket connections using the Concrete Damaged Plasticity (CDP) model and investigates the influence of several parameters related to the constitutive model on the structural behavior of these connections. Initially, a set of modeling choices is presented and validated against 12 experimental test results from the literature. The tests include (i) pile caps cast monolithically with the column, (ii) pile caps connected to the column through socket connections, (iii) different types of socket connection interfaces, and (iv) various embedment depths of the socket-column connection. Following this validation, a parametric study was carried out by varying material parameters within the constitutive model to evaluate their impact on the structural behavior of the members. The outcomes demonstrated strong agreement between the numerical predictions and the experimental data, with an average predicted-to-experimental ratio of 1.04 for ultimate capacity and a coefficient of variation of only 9.21%. Furthermore, the model was able to accurately replicate failure mechanisms observed in physical tests, with a success rate of 91.67%. In summary, the proposed approach shows a high level of accuracy and reliability in predicting both the ultimate capacity and the failure mechanisms of RC pile caps with socket connections.