Research Profile: Jose Andrade Researches Liquefaction

April 28, 2009

When an earthquake strikes, the ground shifts, and oftentimes, buildings crumble. But what if the buildings don’t crumble, but sink? In 1964, an earthquake in Niigata, Japan caused apartment buildings along the Shinano river bank to tilt and sink into the ground — even though the buildings had little structural damage.

This, scientists and engineers would later find out, was the result of liquefaction: the process under which soil, when excited, starts to behave like a liquid.

But how can we know when exactly this will happen? Enter José Andrade, assistant professor of civil and environmental engineering at the McCormick School of Engineering and Applied Science. Andrade and his group create mathematical and computational models that could potentially predict when liquefaction — and its relative, the landslide — will occur.

Andrade explains liquefaction like this: Imagine you’re walking on the beach, and when you step on the sand, the force of your weight gets transmitted into the grains through contact. At this point the grains are so close together that they are touching — and they are organized in what scientists call “load columns” — but there is also water in the empty spaces between the grains. That’s called interstitial water.

When an earthquake happens, it excites the water, which increases the water pressure. That means the water pushes on the grains, causing them to move the columns apart. Now, if you try to step on the same sand, your foot will sink. The sand can no longer transmit the load of your weight.

When Andrade and his students create models and simulations for how this will happen on a large scale, they can’t look at each individual grain of soil and determine how it will react under these conditions.

“Our model grabs a zone of the soil and performs an engineering approximation,” he says. “We have the ability to use the physics to tell whether a chunk of sand is undergoing liquefaction. We can then tell how much a building would be displaced if that happened, and we can tell when it will stop sinking.”

Such models can be used for other situations, too — for example, Andrade’s research group has modeled how the New Orleans levies failed during Hurricane Katrina.

“During the hurricane, the water level increased, which in turn increased the water pressure inside the levy,” he says. “That weakened the levy and caused a chunk to fall off. We have a simulation that shows how much pressure was built up, how this levy collapsed, and how we can use this concept of liquefaction to explain what happened.”

Creating these models isn’t easy — researchers must account for both the laws of physics and the material they are dealing with.

“My students spend years of their lives trying to solve specific problems that will allow us to simulate the physics that we want to simulate,” he says. “We constantly work on them to make them as realistic as possible.”

When the 1964 earthquake happened, scientists didn’t know how liquefaction worked. By creating models, researchers like Andrade can now account for the phenomenon and now can even perform a simulation in the lab and predict whether it will happen.

But field sites get trickier — researchers don’t always know what is in the soil beneath a building, and there might be other variables at work that researchers don’t account for.

“The real test is to predict the future and be able to tell whether this building will collapse if there was an earthquake,” Andrade says. “That is still ahead of us. More engineering is still needed to be able to predict that.”

Andrade is also performing similar research on models of landslides in order to tell where they will happen, when they will happen, and how much damage they could potentially do.

“If we understand the fundamentals, even at the lab scale, we will always be able to go to the larger scale,” he says. “We already have the tools that will soon allow us to jump to the real life situation.”

- Emily Ayshford