Deixis Online

Sometimes, just picking the right model is hard. Thermonuclear weapons
operate under extreme conditions that aren’t easy to understand. As Doebling notes, weapons models must simulate a broad range of environments, from room temperatures and pressures
to conditions in the center of the sun. A theory may work in one regime but prove less reliable in another.

Knobs

There are several ways to get a handle on how aging and reconditioned warheads will perform. One is to learn more about how they age. NNSA does that by dismantling weapons and examining the pieces as part of its stockpile stewardship program.

Deciphering the physical evidence is harder than it sounds. Radiation may have turned a polymer a different color, but its other properties may remain unchanged. A metal may look new but behave differently.

Those changes are critical only if they affect the bomb’s performance. To predict whether they will, model builders need to understand how those materials behave when exposed to the extreme heat, temperature and radiation of a nuclear explosion.

Researchers finessed the problem by including what they call “ad–hoc parameters” in the mathematical models. These parameters accommodated both the unknowns of aging materials and unknowns in the models themselves.

The ad–hoc parameters, however, were based on data accumulated during underground tests that ended more than 15 years ago. Researchers adjusted their materials models until the model outputs matched the test outputs.

Modelers call ad–hoc parameters “knobs.” “From the weapon design perspective, they are necessary to get the job done in design engineering,” Doebling says. “Now we’re going back
and trying to understand the physics of those knobs.”

Verification

Verification and validation, Doebling says, is about ensuring the calculations yield the right answers for the right reasons, which means reducing and eventually eliminating dependence
on knobs. “If we want our models to predict unknown behaviors, they cannot be right just because we tuned the model to our experimental data,” he adds. “They have to be right
because the physics embedded in the models represents how the system truly behaves.”

The Los Alamos V&V team approaches that issue in two ways. First, it attempts to verify that its models solve equations properly. This takes both finely honed math and programming skills and judgment.

The team starts by looking for programming errors and re-checking the math.

Equally important, verification ensures that the model provides enough precision without sacrificing run speed.

Validation

The Los Alamos team’s second approach is to validate a model’s underlying physics. In other words, it ensures the model reflects reality.

New testing facilities make this possible. NNSA now operates several sites that can produce conditions of extreme temperature, pressure and radiation.

The experiments are costly and time-consuming to set up and run. So instead of testing everything, the V&V team identifies areas where precision counts most.

“These are incredibly complex projects, but verification and validation has to be tightly linked to testing,” Doebling says. “Decisions about the safety, security and reliability of nuclear weapons have pretty high consequences. For them to act as a deterrent, we have to know that after aging and modifications, they will still operate as designed.”