Paris Fatigue Life Modeling of Pressure Vessel Service Simulation Tests

Abstract

Results from four groups of full-scale pressure vessel service simulation tests are described and analyzed using Paris fatigue life modeling. The objective is to determine how the vessel and initial crack configurations and applied and residual stresses control the as-tested fatigue life of the vessel. The tube inner radii are in the 40–80 mm range; wall thickness varies from 6 to 80 mm; materials are ASTM A723 pressure vessel steel and IN718 nickel-base alloy; applied internal pressure varies from 90 to 700 MPa. The Paris constant, C, and exponent, m, that describe the fatigue crack propagation rate versus stress intensity factor range for the various vessel materials, were measured as part of the investigation. Extensive, previously published fatigue life results from baseline A723 pressure vessels with well characterized autofrettage residual stresses and C and m values are used to demonstrate that a Paris fatigue life model gives a good description of the measured life. The same model is then used to determine the variables with predominant control over life in three types of pressure vessel for which less information and tests results are available. A design life for pressure vessels is calculated for a specified very low probability of fatigue failure using the log(N)-normal distribution statistics often used for fatigue of structures. The results of the work showed: (i) X-ray diffraction measurements of through-wall autofrettage residual stresses are in excellent agreement with prior neutron diffraction measurements from a baseline autofrettaged A723 pressure vessel; these verified autofrettage residual stresses then provide critical input to the baseline Paris life modeling; (ii) comparison of the various full-scale fatigue test results with results from the Paris fatigue life model shows close agreement when autofrettage residual stresses are incorporated into models; (iii) model results for A723 steel vessels with yield strength reduced from the initial 1400 MPa value and degree of autofrettage increased from the initial 40% value indicates a significantly improved resistance to brittle failure with no loss of fatigue life; (iv] comparison of model fatigue life results for IN718 nickel-base alloy vessels with their full-scale test results is improved when near-bore residual stresses measured by X-ray diffraction are included in the model calculations.