A Novel Class of Strain Measures for Digital Image Correlation

Journal Article · Tue May 05 00:00:00 EDT 2015 · Strain

DOI:https://doi.org/10.1111/str.12138· OSTI ID:1401391

Lehoucq, R. B. ^[1]; Reu, P. L. ^[2]; Turner, D. Z. ^[3]

---

1. Computational Mathematics Sandia National Laboratories Albuquerque NM USA
2. Sandia National Lab 01535, PO Box 5800, MS‐1139 Albuquerque NM 87185 USA
3. Multiscale Science Sandia National Laboratories PO Box 5800, Albuquerque NM 87185 USA, Civil Engineering Stellenbosch University Private Bag X1, Stellenbosch Western Cape 7602 South Africa

Abstract We propose a novel class strain measures for use with digital image correlation (DIC). Whereas the traditional notion of compatibility (strain as the derivative of the displacement field) is problematic when the displacement field varies substantially because of either measurement noise or material irregularity, the proposed measure remains robust, well defined and invariant under rigid body motion. Moreover, when the displacement field is smooth, the classical and proposed strain measures are approximations of each other. We demonstrate, via several numerical examples, the potential of this new strain measure for problems with steep gradients. We also show how the non‐local strain provides an intrinsic mechanism for filtering high‐frequency content from the strain profile and so has a high signal to noise ratio. This is a convenient feature considering image noise and its impact on strain calculations.

• Sponsoring Organization: USDOE
• OSTI ID: 1401391
• Journal Information: Strain, Journal Name: Strain Vol. 51 Journal Issue: 4; ISSN 0039-2103
• Publisher: Wiley-Blackwell
• Country of Publication: United Kingdom
• Language: English