Login

Project

#81 Infrastructure Monitoring from an In-Service Light Rail Vehicle


Principal Investigator
Jacobo Bielak
Status
Completed
Start Date
Jan. 1, 2015
End Date
Dec. 31, 2015
Project Type
Research Advanced
Grant Program
MAP-21 TSET - Tier 1 (2012 - 2016)
Grant Cycle
2015 TSET UTC
Visibility
Public

Abstract

Globally, infrastructure is a vital asset for economic prosperity. However, condition assessment to ensure the continuing safe operation of these assets tends to be subjective and infrequent. We propose a low-cost, objective method to continuously monitor rail systems from the vibrations recorded in a passing train that could complement traditional inspection techniques. The concept of using passenger trains to probe infrastructure assets has been proposed previously, but most work has been limited to simulations, laboratory tests, and short-term experimentation. We have embarked on a long term monitoring project with the Port Authority of Allegheny County; through this deployment we have been testing and refining our techniques. We have instrumented one of Pittsburgh’s light rail vehicles and have been monitoring the rail system from the vibrations in the operational vehicle over the last year. Our objective is to provide accurate, rapid, nearly continuous assessments of the tracks, track structures and bridges along the line as cost-efficiently as possible. We already have seen promising results, as we have been able to detect changes in the recorded signals after construction activity on both the track and ballast. We are working on novel signal processing and machine learning approaches to automate change detection. Ultimately our goal is to build an infrastructure asset management technology which can serve not only the Port Authority of Allegheny County but infrastructure owners for a variety of transportation modes.      
Description
We propose to build on our long term infrastructure monitoring project by developingasemi-supervised change-detection algorithmfor automated damage detection. Over the last year, we have formed a successful partnership with the Port Authority of Allegheny County, and have been continuously collecting the vibration signals from one of their light rail vehicles. We will leverage this partnership and the success at identifyingcertain types of infrastructure change, to build a broadly applicable change-detection technology. 

Currently, the Port Authority maintains over 40km of track, much of it built in the early 20th century. They monitor their system by sending inspectors to walk the track at least every quarter, and hire a vendor to perform ultrasonic inspection of the track on a yearly basis. Between these infrequent inspections, they rely primarily on the train operators (the drivers) to report anomalies or issues. 

We have installed a computer, a GPS antenna, and accelerometers, (Figure 1) to continuously provide objective data on the state of the system from a moving train. This system complements traditional inspection techniques by providing additional information. For example, the Way Department may fix a section of track after operators report a banging noise. But prior to our system, they did not know the exact time when the banging starts, or whether the problem will recur after the repair. Our system has the potential for identifying issues earlier, and can provide data on the efficacy of the repair.

This exact situation occurred around a drain inside the Mount Washington Tunnel, on a section of right-of-way which is shared by both trains and buses. The rubber-wheeled buses occasionally would cause the drain cover to slide around, encroaching on the path of the flange of the train’s wheel. To solve the problem, the Port Authority filled in the drain with concrete (Figure 2). At times when the drain cover had moved, large peak accelerations were recorded as seen in Figure 3. These went away after the drain was removed on May 31st. In addition to the change from the repair, we can see the evolution of the damage over time.

While the detection of the drain provides some motivation for our monitoring, our ability to detect changes too subtle for the operators is what can make our system truly valuable. The Port Authority regularly tamps their track in order to correct the track geometry. We can look at a section of track before and after tamping as seen in Figure 4. Figure 4 shows six passes of the train between Boggs and Denise station.  On May 23rd the port authority tamped the tracks and replaced some cross ties. This work is not readily apparent in the raw signal, but by looking at the standard deviation of the same signals in Figure 5, we see additional bumps in the signal initially, which disappear after tamping. The standard deviation [1] is often used to assess ride quality (lower is better), so a change in this value indicates that the tamping was successful, as ride quality improves.   Leveraging our ability to detect the improvements due to tamping, we will now seek to identify when the track requires tamping so that the maintenance regime of the ballast can be optimized. 

Over the next year, we will focus on automated damage detection approaches. Under previous UTC funding, we developed novel semi-supervised machine learning algorithms using our laboratory scale data [2].The task is now to modify these approaches to work on the large database we have built (which is still growing) of data from the Port Authority. In addition, we will imbue the algorithms with the knowledge of the features we have found effective in tracking various types of damage. 

This past year we have given a number of presentations to disseminate our findings; our work has piqued the interest of several groups. A potential relationship with such a group, Bosch, is particularly promising as they have an office in Pittsburgh and have offered us sensors and acquisition hardware. We hope to build a partnership with them over the next year and to take advantage of their expertise in automotive sensing. 

As we continue to work with the Port Authority and look for new partnership opportunities, we hope to build smarter and safer transit systems. In addition to our vehicle-based monitoring system, we have worked with the Port Authority to address other concerns in their system. In particular, we have connected them to academic experts who have offered solutions to their ongoing wheel-chipping problem. The Port Authority has welcomed these potential solutions and has implemented the recommendations on some of their railcars in an effort to determine if the chipping problem can be addressed. 

Our vehicle-based infrastructure monitoring technology has the potential to provide earlier indications of track and track structure anomalies, along with information about how the system ages. By examining our data over time, we can assess the efficacy of various maintenance interventions, to allow for smarter maintenance and enhanced safety in the future. 

[1] Westeon, P. F., Ling, C. S., Roberts, C., Goodman, C. J., Li, P., & Goodall, R. M. “Monitoring vertical track irregularity from in-service railway vehicles.” Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit,221(1), 75-88 2007.
[2] S. Chen, F. Cerda, P. Rizzo, J. Bielak, J. H. Garrett, and J. Kova?evi?, "Semi-supervised multiresolution classification using adaptive graph filtering with application to indirect bridge structural health monitoring," IEEE Trans. Signal Process., Volume 62, Issue 11. June 2014. (http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=6778068)
Timeline
This project began with simulations in 2010 and laboratory experiments 2011-2013. In September 2013, we instrumented a train and began testing our work in the field. In 2014, we have made significant progress on automating the transfer of data from the train into a database, and we have found features for detecting track and ballast changes when the location of the work is known. In 2015, we will broaden this work to have automated change detection, so that we can find changes without prior knowledge of their existence. In 2016 we plan to implement automated owner notification of infrastructure changes, with a particular focus on minimizing false positives. 
Strategic Description / RD&T

    
Deployment Plan
We have already deployed our system on a train, we will deploy a system on a second train, and work with partners like Bosch to instrument additional trains. Our goal now is to provide the Port Authority with the most useful information from the data we collect—by 2016, we hope to build an automated notification system. 
Expected Outcomes/Impacts
In 2015, we plan to be able to detect deterioration (or a proxy for deterioration), and know where on the transit system the deterioration occurs. The final step, automatically alerting the relevant authorities of this deterioration, will occur in 2016. Over the next calendar year, we will publish a paper about our novel instrumentation, and publish a paper about our ability to detect changes from our data.We will continue to incorporate improvements in the information we convey to the Port Authority, so they can directly benefit from our research. These advances will be useful also to other railways systems.
Expected Outputs

    
TRID


    

Individuals Involved

Email Name Affiliation Role Position
jbielak@cmu.edu Bielak, Jacobo CEE PI Faculty - Research/Systems
chensiheng1989@gmail.com Chen, Siheng ECE Other Student - PhD
jelenak@cmu.edu Kovacevic, Jelena ECE Co-PI Faculty - Research/Systems
noh@cmu.edu Noh, Hae Young CEE Co-PI Faculty - Research/Systems
zihao@cmu.edu Wang, Zihao CEE Other Student - PhD

Budget

Amount of UTC Funds Awarded
$122913.00
Total Project Budget (from all funding sources)
$122913.00

Documents

Type Name Uploaded
Final Report Infrastructure_Monitoring_from_an_In-Service_Light_Rail_VEhicle.pdf April 2, 2018, 5:16 a.m.
Publication Track-monitoring from the dynamic response of an operational train. Dec. 2, 2020, 9:40 a.m.
Publication Track monitoring from the dynamic response of a passing train: a sparse approach. Dec. 2, 2020, 9:41 a.m.
Publication A data fusion approach for track monitoring from multiple in-service trains. Dec. 2, 2020, 9:41 a.m.
Publication Dynamic responses, GPS positions and environmental conditions of two light rail vehicles in Pittsburgh. Dec. 8, 2020, 9:53 a.m.

Match Sources

No match sources!

Partners

No partners!