Simulation-based 4R fatigue assessment of high-frequency mechanical impact treated welded joints

High-frequency mechanical impact (HFMI) treatment improves the fatigue strength of welded joints by compressive residual stresses, improved local geometry, and strain hardening, which can be considered in fatigue predictions using the 4R method. This study investigates the applicability of a simulation-based fatigue assessment model to HFMI-treated longitudinal stiffener joints. The finite element (FE) simulation of fabrication and cyclic loading integrates computational weld mechanics (CWM), material prediction in the heat-affected zone (HAZ), HFMI treatment simulation, and cyclic material response analysis under mechanical loading. The fatigue assessments were carried out using the multiparametric fatigue analysis, the 4R method, which employs FE-simulated material local stress responses. The simulation results were experimentally validated through geometry and residual stress measurements, hardness characterizations, and fatigue tests, showing a reasonably good agreement with the experimental findings. The simulation-based fatigue life values showed good agreement with experimental results at the low stress ratio ( R = 0.1), with simulation-based predictions being slightly conservative. At a higher stress ratio ( R = 0.5), the simulation-based model, with simplifications, was not able to capture all fatigue characteristic factors of the specimen, predicting non-conservative capacity for the HFMI-treated joint compared to the experimental data. Nevertheless, simulation models are an effective tool for estimating the local fatigue-effective stresses and strains for HFMI-treated joints in validated and applicable ranges of load conditions, and offer a cost-effective alternative to experimental specimen fabrication and fatigue testing. • FE model to predict fabrication-induced conditions of HFMI-treated welded joints. • Master curve fatigue prediction by mean stress correction of simulation history. • Experimentally validated numerical fabrication and fatigue results. • Experimental and simulated fatigue data in good agreement at moderate mean stress. • Simulation model was unable to capture all fatigue factors at high mean stress.