Custom Computational Software and Systems to help Drive your Technology needs
|
|
|
Army Research Labs XCT Ballistic Image Filter Program
Development and Preliminary Application of Image Filters for Quantitative XCT Damage Analyses
Jeff Wheeler @ University of California - Santa Cruz
Caonabo Delgado @ Rutgers University
Herb Miller @ University of Maryland - Baltimore County
Abstract
X-ray computed tomography (XCT), is an accurate and powerful way to image and evaluate damage,
including damage, including cracking in materials due to ballistic impact. However, a clear and
consisten algorithm must be applied to XCT image data to determine damage on a quantitative basis.
A methodology was developed using a number of software applications to graphically basis. A
methodology was developed using a number of software applications to graphically interrogate the
physical amount of ballistic meso-damage(>250 micro meters) in impacted titanium alloy (Ti-6Al-4V)
disk samples. The disk samples were sectioned from an imported titanium alloy plate.
Full Paper
Advances in Image Filtering Techniques for Quantitative X-Ray Computed Tomography Ballistic
Damage Analyses
Jeffrey R. Wheeler @ Santa Cruz Institute for Particle Physics, University of California - Santa Cruz
Herbert T. Miller @ University of Maryland - Baltimore County
Abstract
An analytical technique has been developed using LabView to quantify the volumetric ballistic damage
in impacted material samples. This technique is a significant improvement over a previous, mor labor
intensive, method which used a variety of software applications to analyze and quantify the extent
and morphology of 3D ballistic impact damage. This new software technique processes the digitized
X-Ray computed tomography (XCT) images and characterizes the amount of damage in a substantially more
efficient manner than the previous method. The application of this technique greatly enhances our
capability to quantify the volumetric meso-scale damage details of impacted targets.
Full Paper
A Framework for the Analyses and Visualization of X-Ray Computed Tomography Image Data Using a
Compute Cluster
Jeff Wheeler @ ORISE Contractor at US Army Research Laboratory, Weapons & Materials Research Division
William H. Green @ US Army Research Laboratory, Weapons & Materials Research Division
Michael Schuresko @ Baskin School of Engineering, University of California - Santa Cruz
Michael Patrick Lowery @ Computational Mathematics Department, University of California
Abstract
A fast consistent data analyses and visualization software package has been created to assist the
defense community in analyzing post penetration volumetric ballistic meso-scale damage. A library
was written using open source and cross platform software application program interfaces (APIs),
OpenGL and LAM-MPI, to allow high performance visualization and analyses of X-ray computed tomography
(XCT) image data using a cluster of computers. The cluster installation is discussed. A statistical
summary of individual images is performed to determine traits in order to characterize and database
common features among data sets to identify similarities and record them. The aim of this project is
to streamline the production of 2d and 3d visualizations of XCT image data in a small and easy to use
analytical package.
Full Paper
Ballistic Damage Visualization & Quantification in Monolithich Ti-6Al-4V with X-Ray Computed Tomography
J.M. Wells @ JMW Associates
W.H. Green @ US Army Research Laboratory, Weapons & Materials Research Division
N.L. Rupert @ US Army Research Laboratory, Weapons & Materials Research Division
J.M. Winter, Jr. @ ORISE Contractor at WMRD
J.R. Wheeler @ ORISE Contractor at WMRD
S.J. Cimpoeru @ DSTO Melbourne
A.V. Zibarov @ GDT Software Group
Abstract
This paper addresses the investigation of the in situ volumetric internal ballistic damage of monolithic Ti-6Al-4V
alloy V50 targets using X-ray Computed Tomography, XCT. The interrogation was conducted with individual XCT scans
of ~ 0.5 mm thickness to reveal 3D volumetric damage larger than the minimum feature resolution limits (> 0.25 mm).
Image processing and visualization techniques utilized include contiguous 2D scan sequencing, reconstructed 3D
solid objects and virtual sectioning, and recently developed colorized point cloud imaging which more clearly
allows visualization of the volumetric damage network by making the matrix itself transparent. Additionally, new
methods to quantify the volumetric axi-symmetric damage fraction are introduced. Observed damage features include
multiple meso-scale cracks and voids. Of particular interest is the distinct and interesting meso-scale cracking
morphology described herein as 'Spiral orbital cracking' located beyond the penetration cavity in the exit half
of the sample.
Full Paper
The Analysis and Classification of Images - Generated by X-Ray Computed Tomography
Presented At: 41th Annual Armament Systems: Guns and Missile Systems Conference & Exhibition - "Enhancing
Our Capability for Tomorrow"
Michael Lowery @ University of California - Santa Cruz
Jeff Wheeler @ University of Kentucky
Abstract
An X-ray computed tomography (XCT) analyses and visualization software
package has been created to assist the defense community in analyzing post
penetration volumetric ballistic meso-scale impact damage for
characterization and classification of traits in novel materials. This
application not only quantifies depth of penetration, fragmentation, and
cracking in materials but also attempts to identify damage propagation
throughout the material. In order to accomplish this task, routines were
developed to calculate the physical amount of meso-damage. Based on a
central axis of penetration the discrete data in each XCT slice is
determined analytically by using annular ring, wedge, and square segments.
Within each of these a percent damage (black pixels versus total pixels)
is assigned. These segments are of varying dimensions based on several
criteria involving the pixel widths of each as well as different theta
increments for the wedge segments. Different methodologies are used to
determine these dimensions and each describes the information within the
XCT data uniquely and consistently. Several thresholds are used to create
the sets of binarized XCT image data used as the input of this program.
The results are placed into a database built on characteristics found
during analyses. The XCT data takes time to process. To help alleviate
this issue the software was adapted for a high performance computing (HPC)
environment. In order to streamline the processing of XCT image data, a
stand-alone compute cluster was created to analyze and database the
results. Various manufacturers were considered who had commercial
solutions for the cluster computing needs, including Penguin Computing and
Dell. It was determined that the best solution, however, was not to buy
an off the shelf compute cluster, but to build a compute cluster from off
the shelf computers. The LAM-MPI library is used for the parallel
programming. This paper will discuss the details of the aforementioned
processes as well as the conversion from the serial to parallel computing
environment. It will also detail the results of several data sets.
Full Presentation
(Upcomming Publication) The Evolution and Application of Asymmetrical Image Filters for Quantitative XCT Analyses
N.L. Rupert @ KLNG Enterprises
J.M. Wells @ JMW Associates
W. Bruchey @ Service Engineering Corporation
J.R. Wheeler @ Marshall University
Abstract
The successful use of nondestructive x-ray computed tomography (XCT) techniques to
characterize and visualize ballistic impact damage has been demonstrated. Their
high resolution two-dimensional (2-D) tomographic slices and three-dimensional (3-D)
virtual solid object reconstructions represent a telling, spatially accurate, and
qualitatively informative rendering of the internal damage present. However, the
damage is not presented in a format that allows the XCT damage reconstructions to
be directly compared with hydrocode modeling of the same ballistic event. An
initial attempt to bridge the difference between the two diagnostic techniques was
made by converting the XCT data into damage fractions as a function of depth and
radius (from the impact center), assuming an axial-symmetrical damage profile. The
next step addressed the automation of the conversion process. In this paper, the
authors address the conversion of XCT impact damage data into true 3-D damage
fractions as a function of depth, radius, and rotational angle. This data
conversion more closely describes the asymmetrical damage observed in the 2-D
tomographic slices and 3-D virtual solid reconstructions, and the type of
information required for future validation of full 3-D hydrocode damage modeling.
Full Paper Soon
|
|
