A review of micro-scale focused ion beam milling and digital image correlation analysis for residual stress evaluation and error estimation

Abstract

Abstract In the past decade several versions of micro-scale residual stress analysis techniques have been developed based on Focused Ion Beam (FIB) milling at sample surface followed by Digital Image Correlation (DIC) for the determination of the resulting strain relief. Reliable and precise estimation of the error bounds on these measures is critical in determining the usefulness and accuracy of residual stress evaluation. Here we present an overview of the steps necessary for effective outlier removal, error propagation and estimation in order to provide reliable confidence limits for the stress value obtained. Error propagation analysis begins with DIC marker tracking errors that depend on imaging contrast and magnification, and can be improved with sub-pixel tracking and marker shift averaging. We demonstrate how the outliers and poorly tracked markers ought to be removed from the data set using correlation coefficient thresholding and/or correlation peak confidence intervals. Markers showing large displacements relative to their neighbours can also be identified as aberrant, and removed. By performing careful error propagation throughout the analysis chain we quantify the displacement and strain fields, and qualify them with the associated confidence intervals. These values, in combination with the elastic modulus confidence limits, are then used to provide the final confidence intervals for the determined residual stress values. The generic nature of the methodology presented ensures its suitability for all residual stress analysis techniques based on FIB milling and image correlation analysis. An example of implementation is presented for the micro-scale ring-core FIB–DIC approach.