Quantitative representation of microstructural contributions to fatigue crack growth.

摘要

The problem of quantifying the effects of microstructural contributions to the fatigue crack growth process was dealt with using two approaches. The first explored the correlation that exists between observed microstructural events and fatigue crack growth data, while the second involved the representation of material variables in terms of an uncertainty in life estimation.;Both approaches necessitated the development of data acquisition and data reduction software. The data acquisition software made use of the compliance technique to remotely monitor crack length and to pause the test at regular intervals. The data reduction software took the test data and reduced it for presentation using any one of a number of different formats. In addition to this, data were able to be fit using three types of linear least squares fit, a second order polynomial fit, and the four parameter Weibull function. The Weibull function's advantages and ranges of applicability were investigated thoroughly and reported.;The LFM was modified to provide greater observational flexibility through enhanced specimen positioning capability and instrumented to allow crack length to be monitored remotely using the compliance technique. Crack length intervals of 50 microns were resolved, although the system should be capable of detecting intervals as small as 10 microns. The compliance calibration for the nonstandard LFM specimen was determined experimentally and using finite element analysis. The latter was then used to obtain a stress intensity calibration.;The influence of grain boundaries on the fatigue crack growth rate was resolved and documented. Increased crack length resolution would make it possible to study the effects of smaller microstructural features using this technique.;An alternative fatigue data presentation format was developed that requires much less data reduction and allows for improved life estimation convenience as compared with currently used procedures. The format was used to demonstrate the manner in which uncertainties caused by the test system and microstructural contributions could be translated into life estimation uncertainties.