Error loading MacroEngine script (file: )

Port Security

By Lesley Kriewald

Helping to increase the chance of detecting smuggled nuclear materials at U.S. ports and borders

With more than 300 sea and river ports, and more than 3,700 cargo and passenger terminals in the U.S. maritime system, security officials fear that terrorists may use the nation's ports and waterways to smuggle materials into the country for an attack.

Stopping a terrorist from smuggling nuclear material into the country inside a shipping container — 12 million of which enter the country annually through ports — is the goal of a collaborative effort funded by the U.S. Department of Homeland Security's Domestic Nuclear Detection Office (DNDO).

As part of the $7.5 million project, Gary Gaukler, an assistant professor in the Department of Industrial and Systems Engineering, leads a research team that includes Associate Professor Yu Ding, post-doctoral researcher Chenhua Li, and several undergraduate and graduate students.

Together, the team focuses its expertise in queuing networks and modeling and simulation on how best to use radiation detectors at U.S. borders and ports to help inspectors find — and recover — nuclear materials being smuggled into the country.

"What kinds of detectors can we rely on?" Gaukler says. "We need to know what kind of incidents a detector will see, and given the detectors you have, how probable is it that you will detect that material?"

Government agencies such as the DNDO are concerned that adversaries, or terrorists, might be able to acquire small amounts of nuclear materials — say, several kilograms of highly enriched uranium (HEU) or plutonium — and smuggle them into the United States. Once here, the small amounts can be assembled into a nuclear weapon.

"It's much easier to detect a fully assembled nuclear weapon than small amounts of nuclear material," Gaukler says, "because the more nuclear material you have, the brighter the signal it gives off to a radiation detector. So if you want to avoid detection, it may be much smarter to smuggle a weapon in piecemeal."

One way to increase the chances of finding smuggled material is to spend more time inspecting individual containers. But, Gaukler says, more time is a luxury that most border and port inspectors don't have.

"Consider the scale of the problem," he says. "One kilogram of HEU, a very dense material, would be about the size of a baseball. Imagine trying to find that in a standard 40-foot shipping container."

Gaukler also says that HEU is a low-brightness material, meaning that it doesn't give off many photons (gamma rays) or neutrons — which in turn means that radiation detectors that rely on counting these particles may miss the signal. Another problem that complicates detection efforts is nuisance alarms, which occur when background radiation sets off the detectors.

"It's hard to distinguish whether a particular reading is from smuggled nuclear material or from background radiation, either from the ground or the sky, or even from certain things such as ceramics and fertilizers that have a small amount of gamma readings," Gaukler says. "So there's the fundamental detection tradeoff: Either you set the radiation threshold on the detector high, which means you may miss some low-brightness materials, or you set it low and get a lot of false alarms."

At U.S. ports today, containers are first scanned through radiation portal monitors (RPMs). If no alarm registers, the container is loaded and sent on its way. But if an alarm does sound, the container undergoes a secondary inspection while remaining closed. If this secondary scan also detects radiation, inspectors open the container and examine it.

Intelligence information — such as who and where the shipment is coming from, where it is going and who loaded the container — is part of this escalation system. Software called the Automated Tracking System (ATS) tracks this information, which is known 24 hours before any shipment leaves its home port bound for the U.S. On the basis of this information, each container is issued a "risk score." Then, when the container arrives in the U.S., those risk scores are evaluated, and a percentage of the highest-risk containers are segregated and investigated.

Gaukler says the performance of the current ATS-based inspection system hinges on two things: accurate assessment of high- and low-risk containers and having sensitive passive detectors, such as RPMs.

"The problem is that we already know that RPMs are largely incapable of detecting small quantities of shielded low-brightness material, such as HEU," Gaukler says. "And we don't know how good the ATS risk score classification is because we've never had this type of nuclear material smuggled into the U.S. before. It's just never happened. So there is no historical data to test the performance of the ATS."

"There's the fundamental detection tradeoff: Either you set the radiation threshold on the detector high, which means you may miss some low-brightness materials, or you set it low and get a lot of false alarms."

So to increase the chances of discovering small amounts of hidden nuclear materials, Gaukler and his research team are looking at the inspection system at a port and modeling it as a queuing network. This bird's-eye view allows the team to evaluate any given configuration of detectors and containers and to determine the average detection probability versus how much time is spent on each container.

A key part is knowing what's supposed to be inside each container. Some materials, such as lead, help to hide the HEU by blocking the nuclear material's radiation signal. This effect is called shielding.

"If we know how much shielding is available in the container, we can make better decisions on whether to investigate further," Gaukler says. "For example, cotton T-shirts don't provide any shielding, but hiding the material inside the cylinder bore of an engine block provides a great deal of shielding that the RPM just doesn't pick up.

"So we argue that the differentiation ought to be done based on container contents, not necessarily based on ATS risk score."

Gaukler has suggested using radiography equipment to take a sequence of images of a container and its contents. Those images show the various densities of the materials in the container, and the images can be compiled to produce a 3-D model of what's inside the container. Then the researchers use Monte Carlo n-Particle software to simulate a radiation scan on that container.

"We believe that we have a better inspection policy for ports and border crossings. DNDO and DHS are interested in deploying radiography nodes in ports, probably within the next five years."

Each simulation is run twice, the first time simulating background radiation and the second time simulating a quantity of HEU in the container. This experiment simulates the amount of gamma and neutron particles that a passive detector would see coming out of that container. If the container has no nuclear material, the RPM should see very little radiation. If nuclear material is present, the RPM will detect more radiation, but the contents will shield the radiation signal.

"This approach produces two distributions, and the closer they are together, the harder it will be to detect that something is there," Gaukler says. "And the more dense the materials inside a container are, the harder it will be to detect the radiation passively by using RPMs, so in our proposed inspection policy, the more likely we will be to escalate that container to secondary inspection."

This radiography node replaces the "somewhat fuzzy" ATS node so that decisions are made on the basis of container contents.

"We can completely quantify that decision to escalate a container," Gaukler says.

The team has shown that their model performs better than the current inspection system in terms of time and probability of detection.

"We believe that we have a better inspection policy for ports and border crossings," Gaukler says. " In fact, DNDO and DHS are interested in deploying radiography nodes in ports, probably within the next five years."

Dr. Gary Gaukler
Dr. Gary Gaukler
Assistant Professor
Industrial & Systems Engineering