A comparison of strain calculation using digital image correlation and finite element software

Published

September 4, 2016

Author(s)

Dilip K. Banerjee, Mark Iadicola

Abstract

Digital image correlation (DIC) data are being extensively used for many forming applications including constitutive law calibration, benchmark calibration, and for comparisons with finite element analysis (FEA) simulated results. The most challenging comparisons are often in the area of strain localizations just prior to material failure. This is because limit strains produce an inhomogeneous strain field prior to imminent failure. There is not a consensus on what constitutes "raw" data in DIC measurement and how estimates of errors and uncertainties in "raw" measurands (e.g., shape or displacement) affect the comparison of DIC to FEA. Smoothing related to DIC analysis parameters used for matching and the effective gauge length used for strain calculation can affect the limit of spatial resolution and appropriateness of comparison to FEA. For example, strain measurement uncertainty increases when the virtual gauge length (over which strain is ascertained) decreases. While qualitative comparisons can be misleading, quantitative comparisons are difficult because of insufficient information about the type of strain output from FEA software. This is because FEA software often do not specify how exactly strains are computed for a given type of element. Additionally, each software computes strain differently for a given type of element i.e. shell etc. In order to understand this, three bechmark problems are constructed and results obtained from three commercial FEA software are compared with manual computation performed using properties of element shape functions and known displacements at nodes. Finally, strains computed from DIC displacements from a Marciniak [1] forming limit test (just prior to failure) are compared to those from three commercial FEA software. Quantitative differences in calculated strains are assessed to determine if the scale of variations seen between FEA and DIC calculated strains constitute real behavior or just calculation differences.

Proceedings Title

The 10th International Conference and Workshop on Numerical Simulation of 3D Sheet Metal Forming
Processes

Conference Dates

September 4-9, 2016

Conference Location

Bristol, UK

Conference Title

Numisheet 2016

Pub Type

Conferences

Download Paper

Local Download

Citation

Banerjee, D. and Iadicola, M. (2016), A comparison of strain calculation using digital image correlation and finite element software, The 10th International Conference and Workshop on Numerical Simulation of 3D Sheet Metal Forming Processes, Bristol, UK, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=921168 (Accessed April 19, 2024)