Experimental Methods for Determination of Precision and Estimation of Accuracy in XRD Residual Stress Measurement

Abstract

AbstractIt is important to establish precision and accuracy of an X-ray diffraction (XRD) residual stress measurement procedure in order to compare capabilities of instrumentation and techniques, as well as to provide confidence limits for experimental data. There is no broadly acceptable method for establishment of precision and accuracy. This paper describes a proposed approach and one which is used at The Pennsylvania State University.One impediment to the measurement of precision and accuracy is that no standard specimen with a known residual stress level is available. Proposed standard specimens have been abandoned for various reasons, including the concern for stability of the original stress condition and the existence of stress gradients, i.e., stress inhomogeneity, in the specimens. However, there is one type of specimen which has been accepted by ASTM as an alignment and zero residual stress confirmation standard. That type of specimen is a powdered sample of metal or ceramic which provides XRD peaks in the vicinity of the Bragg angle in which residual stress measurements are to be performed.Some researchers tend to report detector counting statistics as the uncertainty of stress measurement but such statistical scatter accounts for only one part of uncertainty in precision and accuracy. The total uncertainty is best determined directly through repeated residual stress measurements performed by removal and readdressing the test specimen or through a repetition of measurements under predictably changing conditions. This paper describes results from the use of powder specimens to establish the repeatability of measurements with a portable instrument after removal and readdressing of the specimens. Also, results showing the uncertainty of the measured stress change in specimens subjected to known loads are discussed.