Since the Baltimore & Ohio Railroad was first chartered in Maryland in 1827, trains have been a crucial part of the growth of the United States. Nearly 200 years later, locomotives powered by diesel fuel or electricity travel over 140,000 miles of track, hauling both passengers and freight across the country and contributing billions of dollars to the US economy through jobs, services, and taxes.
For those trains to run, the tracks and bridges have to be well maintained—inspected regularly and repaired where needed. The Federal Railroad Administration (FRA), an agency within the US Department of Transportation, typically requires one bridge inspection a year. Of course there is always the risk of damage between annual inspections, which is why Hai Huang, associate professor of engineering at Penn State Altoona, and Penn State graduate student Jonathan Ambrosino are working on a research project to bridge the gap between inspections with a real-time railroad bridge–monitoring and load-rating system.
Ambrosino came to Penn State Altoona as a Rail Transportation Engineering (RTE) major because he knew he wanted to work on railroad accident investigations. “Railroads carry a third of our nation’s economy. I’d like to improve the safety of the industry,” he explains. “The rulebooks and regulations are written in blood.” So when he was still a high school student he reached out to the National Transportation Safety Board (NTSB) looking for college recommendations and Penn State was on that list.
During Ambrosino’s four years at Altoona, Huang relates, “he took two of my classes and did extremely well. He is pretty strong in math, physics, and mechanics.” He also likes research: “He wants to know ‘why.’” With the benefit of the John P. Conrad Rail Transportation Engineering Research Scholarship for bridge inspection research, Ambrosino has been able to pursue that interest. In five years, he will complete his Ph.D. in civil engineering with a specialty in rail infrastructure engineering at University Park.
Their research project focuses on the integrity of railroad bridges. “We have about 67,000 railroad bridges in total. Around 40 percent are really old. A lot of them were designed and built a century ago,” Huang says. In addition, “some of them were overbuilt with a high factor of safety, so those are probably fine.” But what about the rest? How long will they hold? Have they developed any problems? To answer that question, Huang believes “there’s got to be something between those inspection intervals,” ideally 24/7 monitoring.
The monitoring of railroad bridges is a very specialized field, Huang notes; “railroad bridges and highway bridges are very different—with different loadings and different design specs.” Efforts are constantly being made to learn more about railroad bridges and how to improve their performance. The research project started “a couple of years ago when we were in a university transportation project with a big grant from the Consolidated Rail Infrastructure and Safety Improvements Program (USDOT) to come up with a system to do real-time railroad bridge monitoring.”
Jonathan Ambrosino, warming up from being in the field, with fellow graduate students (left to right) Saharnaz Nazarii, Yuliang Zhou, and Kun Zeng.
“Our plan is basically to do bridge sensoring,” Ambrosino explains. “We’re seeing what loads are applied to a certain bridge and seeing how the structure is doing overall.” He gives an example of a potential problem. “Maybe a section of the bridge is ‘acting strange,’” meaning the bridge members are deflecting (i.e., moving) more than allowed. “Deflection is expected for all structures. But AREMA standards say a bridge cannot deflect more than an inch for a 640-inch [53-ft 4-in] span.”
Ambrosino acknowledges that yearly bridge inspections, while well intended, are still to a certain extent subjective. “The problem is whenever an engineer goes out and makes a guess: One says this is fine, another says re-rate,” meaning permit a smaller load than previously allowed. “The problem is there are as many opinions as there are engineers!”
Of course, computers don’t have an opinion. In their research. Ambrosino continues, “We use AREMA and FRA standards as a baseline and then apply our own methodology. We run simulations and calibrate our model with the bridge.” And to make it more complicated, he says, each bridge is different. “There were, and are, many manufacturers for railroad bridges. There is no set standard across all railroads. Each site is also different and has its own challenges, which leads to different design philosophies.”
The railroad industry has the most to gain from better bridge monitoring. After all, it’s their trains and workers who are most at risk in case of an accident. “If you are going to do monitoring, you need the best sensors,” Huang says. “You also have to have very strong industry support. Not only class Is, but also short lines are very welcoming. We even got support from a sensor manufacturer who said they would give us the sensors ‘for free to test and then we can talk about a future plan.’”
Right now, Huang says, “we have four bridges across Pennsylvania that are being monitored 24/7.” But they’re looking for more. “We’re calling for volunteers from the rail industry. At some point there’s got to be someone to pay. It’s a zero initial cost for a railroad company—free installation—but there will be data fees.” That would be a small expense compared to an accident where it’s necessary to “replace the bridge, lose the locomotive and cars, maybe $10–20 million.” Adopting real-time monitoring means “we’ll have warnings if the structural capacity is not sufficient. It’s an extra layer to safety.”
—Therese Boyd, ’79