Publications

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

It is well known that the derivative-based classical approach to strain is problematic when the displacement field is irregular, noisy, or discontinuous. Difficulties arise wherever the displacements are not differentiable. We present an alternative, nonlocal approach to calculating strain from digital image correlation (DIC) data that is well-defined and robust, even for the pathological cases that undermine the classical strain measure. This integral formulation for strain has no spatial derivatives and when the displacement field is smooth, the nonlocal strain and the classical strain are identical. We submit that this approach to computing strains from displacements will greatly improve the fidelity and efficacy of DIC for new application spaces previously untenable in the classical framework.

There has been a lot of interest in the matching error for two-dimensional digital image correlation (2D-DIC), including the matching bias and variance; however, there are a number of other sources of error that must also be considered. These include temperature drift of the camera, out-of-plane sample motion, lack of perpendicularity, under-matched subset shape functions, and filtering of the results during the strain calculation. This talk will use experimental evidence to demonstrate some of the ignored error sources and compile a complete “notional” error budget for a typical 2D measurement.

Three-dimensional deformation of rupture discs subjected to gas-dynamic shock loading was measured using a stereomicroscope digital image correlation (DIC) system. One-dimensional blast waves generated with a small-diameter, explosively driven shock tube were used for studying the fluid-structure interactions that exist when incident onto relatively low-strength rupture discs. Prior experiments have shown that subjecting the 0. 64-cm-diameter, stainless steel rupture discs to shock waves of varying strength results in a range of responses from no rupture to shear at the outer weld diameter. In this work, the outer surface of the rupture discs were prepared for DIC using 100–150 _m-sized speckles and illuminated with a Xenon flashlamp. Two synchronized Shimadzu HPV-2 cameras coupled to an Olympus microscope captured stereoimage sequences of rupture disc behavior at speeds of 1 MHz. Image correlation performed on the stereo-images resulted in spatially resolved surface deformation. The experimental facility, specifics of the DIC diagnostic technique, and the temporal deformation and velocity of the surface of a rupturing disc are presented.

Abstract not provided.

DIC is a non-linear low-pass spatial filtering operation; whether we consider the effect of the subset and shape function, the strain window used in the strain calculation, of other post-processing of the results, each decision will impact the spatial resolution, of the measurement. More fundamentally, the speckle size limits, the spatial resolution by dictating the smallest possible subset. After this decision the processing settings are controlled by the allowable noise level balanced by possible bias errors created by the data filtering. This article describes a process to determine optimum DIC software settings to determine if the peak displacements or strains are being found.

Abstract not provided.

Abstract not provided.

Abstract not provided.

The work presented in this report concerns the response and failure of thin 2024- T3 aluminum alloy circular plates to a blast load produced by the detonation of a nearby spherical charge. The plates were fully clamped around the circumference and the explosive charge was located centrally with respect to the plate. The principal objective was to conduct a numerical model validation study by comparing the results of predictions to experimental measurements of plate deformation and failure for charges with masses in the vicinity of the threshold between no tearing and tearing of the plates. Stereo digital image correlation data was acquired for all tests to measure the deflection and strains in the plates. The size of the virtual strain gage in the measurements, however, was relatively large, so the strain measurements have to be interpreted accordingly as lower bounds of the actual strains in the plate and of the severity of the strain gradients. A fully coupled interaction model between the blast and the deflection of the structure was considered. The results of the validation exercise indicated that the model predicted the deflection of the plates reasonably accurately as well as the distribution of strain on the plate. The estimation of the threshold charge based on a critical value of equivalent plastic strain measured in a bulge test, however, was not accurate. This in spite of efforts to determine the failure strain of the aluminum sheet under biaxial stress conditions. Further work is needed to be able to predict plate tearing with some degree of confidence. Given the current technology, at least one test under the actual blast conditions where the plate tears is needed to calibrate the value of equivalent plastic strain when failure occurs in the numerical model. Once that has been determined, the question of the explosive mass value at the threshold could be addressed with more confidence.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

This report evaluates several interpolants implemented in the Digital Image Correlation Engine (DICe), an image correlation software package developed by Sandia. By interpolants we refer to the basis functions used to represent discrete pixel intensity data as a continuous signal. Interpolation is used to determine intensity values in an image at non - pixel locations. It is also used, in some cases, to evaluate the x and y gradients of the image intensities. Intensity gradients subsequently guide the optimization process. The goal of this report is to inform analysts as to the characteristics of each interpolant and provide guidance towards the best interpolant for a given dataset. This work also serves as an initial verification of each of the interpolants implemented.