Weibull Model Development for Fatigue Crack Growth

This effort is an attempt to "Weibullize" fatigue crack growth rate data as a function of stress intensity range (ΔΚ). It is not a classical Weibull application in that it does not address a population but rather treats rate of crack growth as a fraction of the highest rate of crack growth rate in the data set (usually the final data point of the set) as the "failed fraction of a population." All da/dN values are converted to a percentage or fraction of the 100% data value, processed into a linear relationship and plotted against the natural log of the corresponding ΔΚ value. The crack growth rate data (da/dN vs. Δ Κ), when converted to a Weibull type plot, can be fitted with a linear curve with R2 values in the 0.99 range and characteristic Weibull parameters β and η can be successfully determined. For the data analyzed, the mean value determined for β is 2.86 with a standard deviation of 0.36. However, data from sets of apparently similar specimens produced more closely packed results. The data sets that could not be fit with the Weibull distribution (4 of 32 sets) appeared to have been corrupted, possibly by sorting the data in the spread sheet without including all of the columns. R2 values of > 0.99 were obtained for the linear curve fit of 15 of 28 data sets. R2 values > 0.95 were obtained for an additional 7 while the remaining 6 were > 0.64