A digital servohydraulic test system has been utilized to apply stress spectra, which represented the railhead stress cycles caused by passage of a train, on laboratory compact tension specimens. The system was designed so that tasks that required immediate action were built into the machine controller firmware. Tasks that did not require immediate action were generally performed in the control software on a personal computer. This system was utilized to measure fatigue life of railroad rails under real-life stress spectra that represented the reconstructed railhead stresses for various types of trains. A stress intensity mapping procedure was utilized to simulate the railhead “detail fracture” in laboratory compact tension specimens.
This test system was used to measure fatigue crack growth rate under five different stress spectra. Two of the stress spectra were ordered according to decreasing maximum stress (DMS). The fatigue crack growth rate and crack closure behavior were different for the real sequence ordered (RSO) and decreasing maximum stress ordered (DMS) spectra. Constant amplitude fatigue crack growth rate was measured at three different stress ratios to develop a fatigue crack growth rate model for the material. The fatigue crack growth rate could be modeled with either a modified Walker equation that accounted for stress ratio effects or with an effective ΔK model that accounted for crack closure effects. The constant amplitude test data were used to predict the spectrum load crack growth rates using a modified Palmgren-Miner's rule. The predictions based upon the modified Walker equation were nonconservative. Predictions based upon constant amplitude test data, when corrected for crack closure effects, compared well with the experimental results.