Implementation of high strength steels in welded structural designs in the automotive, defence and construction industries is constantly increasing. Prolonged usage of such structures requires deep understanding of welded joint fatigue as well as a reliable and feasible life estimation methods development. Conventional fatigue testing methods often require costly, expensive in maintenance, high loading capacity equipment. They are also commonly restricted to specific specimen geometry and are time consuming due to the limit of a single specimen per test setup. This work presents high cycle fatigue (HCF) testing of high strength MARS 600 welded steel using a quick, simple and efficient resonance fatigue testing (RFT) method. The specimen is a simple cantilever fillet welded to a base plate using austenitic stainless steel 307L consumable electrode. Electrodynamic shaker is used for harmonic base excitation at a constant operating frequency. Several specimens welded to the common base are tested simultaneously, allowing completion of a high number of cycles and statistics in a relatively short time period. A hybrid, practical research approach combining experimental, finite element analysis (FEA), numerical and analytical calculations is presented. Fracture mechanics approach for fatigue life assessment is implemented. Crack growth calculation is based on the Paris - Erdogan law. Reduction in structural integrity due to crack propagation causes a reduction in natural frequency and transmissibility. The change in gain is evaluated via the open crack FEA model and integrated into the crack propagation algorithm. Resonance search, track and dwell module (RSTD) for maintaining constant gain throughout the test is not required. Fatigue life Wohler (SN) curve is constructed. Standard weld fatigue data is often provided for direct loading (tensile stress) and for different stress ratio (R) values. Corrections for mean stress and loading application are required. Current fully reversed (R = -1), indirect loading (bending stress) test results may be readily applied for random vibration fatigue analyses post processing. As expected, actual fatigue life results are higher compared to standard design curves, implying correctness of the manufacturing welding process of examined specimens. The presented procedure is of interest for research as well as for industrial welding processes testing, optimization and qualification.
Read full abstract