Abstract
This paper quantifies the reliability level of connections combining slip-critical bolts and longitudinal fillet welds using a probabilistic approach that integrates finite element (FE) modeling, meta-modeling, and Monte Carlo simulation. The proposed approach relies on experimental results to calibrate a FE model and quantify the uncertainties associated with various input parameters. The calibrated FE model is next integrated into a low-rank tensor approximation process to provide the variance-based sensitivity measures associated with the considered random variables. The influential variables, along with the calibrated FE model, are then used to evaluate the reliability by means of accelerated Kriging Monte Carlo simulation (AK-MCS). Finally, the proposed approach is applied to combination bolted-welded connections under direct tension to evaluate their reliability considering different faying surface conditions. The results of the conducted sensitivity analysis indicate that variability in weld geometry, bolt pretensioning forces, and friction coefficient of steel surface would significantly contribute to the variability of the load-carrying capacity of the investigated connections. In addition, the reliability of the combination connections with Class A and B faying surfaces is evaluated based on code provisions.
[Display omitted]
•Combination connections with slip-critical bolts and fillet welds are investigated•Experimental results are used to validate numerical models for reliability analysis•Polynomial chaos Kriging Monte Carlo simulation used to quantify the reliability•The adopted approach resulted in significant reduction in computational time•Combining slip-critical bolts and fillet welds results in highly reliable connection