The latest version of the widely used Sandia-developed shock wave physics computer code called CTH, which simulates high-speed impact and penetration phenomena involving a variety of materials, will soon be available to some 250 customers nationwide.
"This new version is really exciting because it offers a computational capability never before available in this type of code, an adaptive mesh refinement model [AMR]," says Paul Taylor, head of the CTH project in Computational Physics/ Simulations Frameworks Dept. 9232.
"AMR, developed for CTH by team member Dave Crawford [9232], gives the software the ability to increase resolution and accuracy in those regions of a simulation where it is needed and reduce resolution in those regions where it is not. For example, in the simulation of a projectile penetrating a target material, greater resolution can be achieved in the region surrounding the impact interface between the two materials where large distortions and high strain rates are occurring."
Interest in the soon-to-be-released version of the software is particularly high among customers like DOE and the Department of Defense (DoD), which use the software for studying weapons effects, armor/anti-armor interactions, warhead design, high-explosive initiation physics, and weapons safety issues. Major users include the national laboratories, the Army, Navy, and Air Force laboratories, and their subcontractors. At Sandia the code is used in national missile defense, hazardous material dispersal by explosive detonation, weapon components design, and reactive materials research.
For armor/anti-armor design — which is of interest to the DoD — the software allows users to determine which types of bullets or projectiles can best penetrate armor. It also provides information about how to design an improved penetration protection mechanism.
The medical community is also paying attention to Sandia’s CTH software. Paul currently has a small collaborative research effort underway with the University of New Mexico School of Medicine, which is interested in using the shock physics code to better understand brain injury caused by physical trauma, such as a person’s head hitting a car windshield. Using the magnetic resonance image (MRI) of an individual’s head to construct a CTH model, simulations can be performed showing how shock waves travel through the head and cause damage to the brain.
The software breaks down the penetration simulation into millions of grid-like "cells." As the modeled projectile (such as a copper ball impacting a steel plate) impacts and penetrates the target, progressively smaller blocks of cells are placed around the projectile, each showing in detail the deformation and breakup of the ball and target plate.
CTH with the AMR enhancement also offers the ability to analyze problems involving sophisticated materials with greater accuracy. With the addition of new material models, it can simulate a wider variety of materials, including metals, ceramics, plastics, composites, high explosives, rocket propellants, and gases (e.g., air).
Sandia developed the early precursor to CTH in the 1970s for one-dimensional problems, expanding it to simulate problems in two and three dimensions in the 1980s.
"The widespread popularity that CTH has today as the shock wave physics computer code of choice began in a competition with Los Alamos National Laboratory in the early 1990s," Paul recalls. "DoD wanted a code that could deal with problems such as armor/anti-armor design, weapons effects, and munitions design. The Los Alamos code was called MESA and ours was CTH. Both codes had comparable characteristics, but DoD selected ours."
The Labs began licensing the shock wave physics code in the early 1990s to DOE, DoD, their contractors, and some private US companies with interests in shock physics. An updated version of the software, which is export-controlled, is distributed to customers about every 18 months. Currently 259 licenses have been issued.
DoD, DOE, and their contractors receive licenses for a small distribution fee to use the software. Commercial companies can purchase licenses for $25,000. The updated software will be distributed on CDs at a cost of $400 for each noncommercial, licensed customer.
One of the important aspects of CTH development is validation of code predictions using actual physical testing. Data gathered in a variety of experiments are compared to CTH models.
"In cases where we are studying situations in which the materials are well- characterized, the code predictions and the actual experiments are very close," Paul says. "The fidelity of the simulations is very good. In fact, CTH is used in many programs to simulate events that are either too costly or dangerous to conduct in full-scale tests. Other researchers use the code to reduce the amount of experimental testing that would otherwise be required."
Paul offers CTH classes several times a year so that customers can fully understand how to use the software and have a full grasp of its capabilities. Users from all over the country come to Sandia to take the classes. Department team members help Paul with the classes, which are broken down into theory and lab segments held over a three-day period.
Paul says one of the most appealing aspects of CTH for users is that it can run on almost any computer platform. The code runs on most Unix and Linux-based workstations and personal computers running Windows NT or 2000.
For users who have access to parallel-architecture computers, CTH can run in parallel mode. This feature permits running large three-dimensional simulations using many processors or nodes to break the problem up into smaller pieces, each of which is solved in parallel with neighboring pieces of the problem.
CTH simulations, conducted in parallel mode on Sandia’s teraþops computer, tend to be the largest problems the code handles. Last December a CTH simulation was performed by Marlin Kipp (9232) on the teraþops computer for a problem containing more than 260 million cells, using 1,024 nodes (2,048 processors) and requiring over 60 hours of computer CPU time.
"CTH problems scale very well with the number of processors allocated for the job," says Paul. "The only limitation to the size of a problem that can be treated using CTH appears to be the availability of processors to complete the computing task."