RR Big Data
Photos by Evan Krape December 19, 2016
Conference addresses big data in railroad maintenance planning
More than 150 representatives of the railway and academic communities turned out for the 2016 Big Data in Railroad Maintenance Planning Conference at the University of Delaware on Dec. 15-16.
The third annual conference was chaired by Allan Zarembski, professor of civil and environmental engineering, and presented by the UD Railroad Engineering and Safety Program, the UD Big Data Center, and UD Professional Engineering Outreach.
“One of our goals is to come up with usable and practical results from the mountain of data that’s now available to railroad engineers and others charged with keeping our railroads and transit systems safe and efficient,” said Zarembski, an internationally recognized expert in railway track engineering, derailment analysis, wheel-rail interaction, and rail problems and maintenance including rail inspection and grinding.
Conference presentations covered a range of topics, including applications and case studies, big data analysis techniques, maintenance strategies, risk management, automated inspection technologies and asset management.
The focus was on making more effective use of the growing volume of inspection and other data now available to railway engineering managers to help them make their rail systems safer, more efficient and more productive.
“Specifically, we wanted to bring to bear the new emerging science of data analytics to address and analyze this high volume of data, which can be as much as several terabytes of data per day for a single major railroad, to convert this data to useable information,” Zarembski said.
Railway engineering at UD
Tripp Shenton, chair of UD’s Department of Civil and Environmental Engineering, said that the department has witnessed tremendous interest in this area, and he provided some highlights of UD’s railroad engineering program, which was established in 2012:
• Four research-based master’s students are completing or have completed their degrees, and four students are currently working on doctoral degrees.
• Close to 200 undergraduate and graduate students have taken UD’s railway engineering courses, all of which are electives.
• Zarembski is teaching not only four courses at UD but also summer classes at Technion-Israel Institute of Technology, as well as a number of online courses that reach students across the world.
• UD is a partner, along with Virginia Tech, on a Tier 1 University Transportation Center (UTC) program recently awarded to the University of Nevada at Las Vegas for “Improving Rail Transportation Infrastructure Sustainability and Durability” (see section at the end of this article for more details).
Lisa Stabler, president of the Transportation Technology Center Inc. (TTCI), a wholly owned subsidiary of the Association of American Railroads, delivered the keynote address, “Big Data in the Rail Industry.”
Headquartered in Pueblo, Colorado, TTCI conducts research and testing on its extensive network of track facilities and in its state-of-the-art laboratory facilities.
With degrees in both mathematics and engineering, Stabler is an expert in quality management, quality engineering and reliability engineering.
After providing a brief overview of TTCI, Stabler used statistics about ice cream consumption and drowning to make a point about turning data into useful information.
“As ice cream consumption increases so does your risk of death by drowning,” she said. “But our understanding of the world gives us the insight to know that it doesn’t make sense to say you can reduce your risk of drowning by avoiding ice cream.”
Most of us can figure out that the data on ice cream and drowning appear to be correlated only because we eat more ice cream and swim more when the weather is warm — there is no real cause and effect at work.
But what happens when the data reflect more complex phenomena?
Stabler shared several concerns that arise in working with big data: understanding the quality of data is critical; big data doesn’t necessarily mean that the final results use all of the data; big data can have factors that are confounded or correlated over time, or auto-correlated; and, finally, many analyses will find statistical significance due entirely to the large sample size.
She said that it’s important both to understand statistical analysis and to have practical knowledge of the issues being examined in order to avoid spurious associations.
“You can have a mountain of data, but you need analysis techniques to turn it into usable information,” she added. “Huge computers don’t change basic statistics – your experiments still have to be well designed.”
She finished with a real-world example in which the meticulous use and analysis of big data resulted in an important rule change aimed at improving safety of railway operations in North America.
“We used our knowledge of statistics as well as the expertise of our engineers at TTCI to reduce risk for the industry,” she said.
About the railway UTC
The consortium funded by the University Transportation Center program recently awarded to the University of Nevada at Las Vegas will develop advanced approaches for managing big data that results from high-tech inspection systems to improve the performance and maintenance of critical railway components and infrastructure, explore new materials and technologies for maintaining and re-conditioning rail surfaces and monitoring key rails components, and provide guidelines for the more rigorous demands of high-speed rail infrastructure by bringing together global knowledge with the geological and topological information of the railway location.
Among UD’s research activities will be the integration of a new generation of inspection technology with track degradation analysis and maintenance planning and the use of data analytics to develop improved planning and forecasting tools for railroad infrastructure.
Additional activities will include education and workforce development, such as developing and enhancing accredited degree-granting programs, and technology transfer to develop partnerships across sectors and move research into practice.
The $1.4 million grant is one of 35 five-year grants awarded to lead consortia under the UTC program, and the only rail focused grant.
The U.S. Department of Transportation received a total of 212 applications, the largest number submitted in the history of the UTC Program. UTC consortia are selected to advance research and education programs that address critical transportation challenges facing the nation.