Ramrods Key to Extraordinary Record
IT seems that every scientific discipline is divided into many specialties. A physicist, for example, may specialize in such areas as nuclear reactions, lasers, computation, cosmology, or quantum mechanics. Bucking the growing trend toward specialization are Livermore ramrods, each of whom is a scientist, engineer, technician, diagnostician, and high-explosives, safety, and hazardous materials expert—all combined into one individual.
As described in the article entitled Ramrods Shepherd Hydrodynamic Tests, ramrods ensure hydrodynamic (or hydro) tests conducted at the Laboratory’s
Site 300 perform as designed. A hydrodynamic test uses nonnuclear materials to analyze the first phase of a nuclear weapon, the implosion of the primary stage. A fully integrated hydrotest can cost more than $1 million and require more than a year of planning, parts production, and assembly. Because the test involves blowing up the carefully instrumented experimental package, we have only one opportunity to get things right. In more than 55 years of hydrotests at Livermore, totaling several thousand tests, a handful of individual experiments have failed, but we have always obtained the data we needed from the test series. That amazing record is a tribute to the ramrods, Site 300 bunker crews, and design physicists.
Ramrods, who are unique to Livermore in the National Nuclear Security Administration (NNSA), warn the design physicist when an aspect of a proposed experiment doesn’t make sense or can be done more efficiently. Ramrods are often asked to come up with a clever solution to a seemingly intractable problem. In solving a problem, they are just as likely to use a socket wrench as a sophisticated software program.
Before Livermore physicists, led by Seymour Sack, developed the first two-dimensional computer codes in the 1960s, scores of hydrotests were conducted in trial-and-error fashion prior to each nuclear detonation at the Nevada Test Site. Over the years, as our computer simulation power has increased, we have required fewer hydrotests for stockpile stewardship. At the same time, the amount of data obtained from each test has increased significantly.
Although we anticipate a continued decline in the number of hydrotests, we will always need them because they enhance our knowledge of how nuclear weapons operate and how materials behave under extreme conditions. Two types of data from hydrotests are fed into two- and three-dimensional supercomputer simulation codes. The first type concerns the properties of strongly shocked materials. The second type shows the integrated behavior of components comprising a certain primary design. These data are compared to those from other hydrotests and our theoretical models and are incorporated into our simulation codes.
The Laboratory’s simulation capabilities have continued to grow as new generations of NNSA’s Advanced Simulation and Computing supercomputers are installed. During last year’s effort to design a reliable replacement warhead, 1,370 years (yes, years!) of computer runs were executed in less than six months to develop a prototype primary design for a hydrotest. This feat was possible only because of the miracle of tens of thousands of computers running in parallel. The hydrotest went off perfectly, generating data that were so outstanding we dispensed with the originally scheduled second test.
In 2006, we also conducted groundbreaking simulations of the growth of tantalum grains. We can’t yet watch such an event occur experimentally. Neither can we figure out the dynamics theoretically. But we can watch grains grow on a computer, atom by atom.
This year, a team from the Defense and Nuclear Technologies Directorate’s B Program accomplished a major breakthrough in understanding how a nuclear weapon functions. This discovery, which was not anticipated theoretically, was observed in ultrahigh resolution on our BlueGene/L machine, the most powerful supercomputer in the world.
It’s doubtful such advanced simulations would exist had it not been for Edward Teller. Teller understood that numerical simulation of physical problems would be invaluable to scientific discovery. Even before the Laboratory opened in September 1952, Teller and Ernest O. Lawrence placed an order for the first Univac computer.
I believe Teller would be pleased with our latest high-resolution simulations. I know he would be proud of our ramrods, who combine scientific acumen, experience, and practical know-how to advance national security.