Extracting Mixed-Mode Fracture Parameters Using Two Vision-based Methods: Comparison of Combined Fields Method with Legacy Approach

Background

Over-deterministic least-squares methods of extracting SIFs from measured full-field quantities in conjunction with asymptotic fields has been the mainstay of experimental fracture mechanics. The vision-based methods of Digital Image Correlation (DIC) to determine displacements and Digital Gradient Sensing (DGS) to determine stress gradients have played an important role in this regard.

Objectives

In DIC and DGS, two or more orthogonal fields are measured simultaneously. Yet, while extracting SIFs, often only one of the components is picked based on intuition/legacy. This could result in erroneous SIF values under mixed-mode conditions.

Methods

Robustness of SIF extraction by utilizing all components in tandem is demonstrated over a wide range of pure- and mixed-mode conditions. An edge-notched semi-circular specimen geometry is used to create different mode-mixities. The data from DIC and DGS are processed using both the combined fields and legacy approaches. The accuracy and robustness of the former relative to the latter is demonstrated for (a) different number of higher order terms in the asymptotic series, (b) crack tip location uncertainty, and (c) different regions of data extraction.

Results

An order of magnitude reduction in standard deviation and root-mean-squared error in mixed and pure mode SIFs are seen for DIC and the combined fields method. Marginal improvements are seen when the crack tip position or the region of interest are varied in DGS.

Conclusions

Robustness of extracting mixed-mode SIFs accurately by employing all measured fields concurrently in an over-deterministic least-squares approach is superior to using a single component based on intuition/legacy.