Analysis of A508 Class 2 and A533B pressure vessel steel fatigue tests in air

A large set of well-documented A508 Class 2 and A533B pressure vessel steel specimens, fatigue-tested in air at various temperatures up to 300 degrees C, were analyzed on a consistent basis starting from the basic experimental data. These data were from a variety of sources worldwide, and were analyzed using an advanced computer code, FATDAC. The reduction of experimental (a,N) data to (da/dN,ΔK) data utilized an optimization technique that minimizes scatter arising from numerical differentiation. All points exhibiting threshold behavior or not meeting elastic validity criteria were deleted; thus, the resulting model applies to Region II crack growth only. A model including a load ratio (R) factor was fitted to a randomly selected four-fifths sample of the specimens, using non-linear least squares techniques. The data were weighted to give more weight to points representing greater crack extension, and to compensate for the lack of experimental data in certain ranges in ΔK and R-Ratio. The resulting model shows a flatter slope than does the current ASME Section XI Code model for air data and subsurface flaws. The remaining one-fifth sample and other sets of data not used for fitting were compared with the model to verify its accuracy, with excellent agreement. Appropriate statistical bounds were constructed on the data and on predicted mean values using the model. Changes to the Section XI reference line are recommended