Prioritizing probabilities

“Civil engineers have done a fantastic job in the last few centuries. We can build anything,” says Paolo Bocchini, assistant professor of civil and environmental engineering. “Yet we keep having significant systemic failures.”

In 2011, for instance, the six boiling water reactors at Japan’s Fukushima Daiichi Nuclear Power Plant reactors survived an earthquake and a tsunami, but subsequent power failures caused several of the reactors to overheat and release radiation.

From buildings and reactor vessels to roads and bridges to the electrical grid and emergency response workers, engineers and disaster planners have traditionally looked individually at each component of infrastructure. In reality, says Bocchini, they comprise an interdependent whole. “The recovery of one infrastructure system usually depends on another.”

Bocchini was one of the first engineers to create computer models that factor the resilience of larger infrastructure systems into disaster planning. Today he leads a multidisciplinary, nationwide team with a $2.2 million National Science Foundation grant to develop a platform—called Probabilistic Resilience Assessment of Interdependent Systems, or PRAISys—“that will provide decision makers with a prediction of what will happen” after a disaster and “prioritize strategies that will bring back functionality most quickly,” he says.

The PRAISys team includes engineers from several disciplines, as well as experts in economics and the social sciences, from Lehigh, Florida Atlantic University and Georgia State.

“Resilience brings to the table the idea that we shouldn’t just look at the extreme event itself,” Bocchini says. The long-term effects are important, and the losses are usually much larger than the initial impact.

Bocchini’s work zooms out from traditional engineering thinking, giving a higher priority to the overall recovery of a system than to the integrity of a system’s individual components.

“The problems we are trying to address make sense when you think about communities and functionality,” he says. “We don’t care so much if there is a crack in a column. We want to know if the bridge will be open.” Ultimately, planners want to know that traffic can reach its destination.

With limited resources and decaying infrastructure, says Bocchini, community leaders need to make decisions with the greater good in mind. After Hurricane Sandy, for example, New York’s public transit system spent an extraordinary sum to drain flooded subway tracks quickly, only to find there were no riders because many offices and attractions in Manhattan were still closed. “It was a loss that could have been avoided with more coordination,” Bocchini says.

As the federal government and many states push to develop better disaster plans, PRAISys researchers are using a probabilistic approach to assess how a community’s systems interact and to give decision makers better options to reduce casualties, long-term socioeconomic losses, and environmental impacts.

“We never try to give a single answer,” Bocchini says. “We provide a set of trade-offs” that officials can use to make better decisions.

It’s usually not politically useful for leaders to shore up systems that are mostly invisible to their constituents, so infrastructure issues are not high on the public’s wish list. Bocchini believes the engineering profession has a role to play in educating citizens.

“We have to make the case with credible numbers that infrastructure resilience is valuable,” he says. One goal of PRAISys is to provide those numbers by quantifying how improvements could save lives and dollars.

“I can see this as the first seed of a life’s work,” he says. “If at the end of my career I have completed such a platform, I will call it a success.”

Bocchini applies the same probabilistic analysis to predicting how Ebola virus spreads, in collaboration with Javier Buceta, associate professor of chemical and biomolecular engineering. Ebola is carried by bats, and migration patterns are affected by complex factors, including temperatures and other weather patterns, he says. Knowing the probabilities of how, and in what direction, an outbreak will spread can allow officials to rapidly direct doctors and supplies.

Bocchini’s research group is also working with Bethlehem city government on its entry for the Rockefeller Foundation’s “100 Resilient Cities” project. And he has developed software that uses guided ultrasonic waves to test pipelines and rail tracks.

In 2009, when Bocchini began working with the probabilistic assessment of complex systems with Dan Frangopol, the Fazlur R. Khan Professor of Structural Engineering and Architecture, only a few research groups in the world were working in this area. Today, infrastructure resilience is a hot topic in government and industry circles.

“Everyone is going this way now,” Bocchini says. “We’re just trying to run a little faster.”

 

Story by Robert W. Fisher ’79