Keynote Speech
AAR’s Strategic Research Initiatives and Workforce Development
Gonzales, Kari
The central point for AAR’s railroad research is with MxV Rail. During the keynote address, conference attendees will learn about the architecture of the AAR’s Strategic Research initiative program and see program highlights in the areas of rolling stock, infrastructure, and operations. In addition to the industry research programs, Ms. Gonzales will highlight emerging technologies being designed and tested in partnership with MxV Rail. Ms. Gonzales will also provide insight into how MxV Rail is working to attract talent to railroading and discussing the railroad workforce of the future.
Technical Session Presentations
Molzon, Michael
Improving Trackside Inspection of Rolling Stock Using a Track Crawler Robot: A Step Toward Better Diagnosis and Prognosis of Railcars
Molzon, Michael and Ahmadian, Mehdi
A mobile robotic platform, equipped with digital cameras, thermal imaging, and similar systems is presented for augmenting trackside inspection of railcars for early detection of any undercarriage deficiencies. The inspections also provide a time history of the condition of railcars for prognosis and maintenance scheduling purposes. The presentation provides the details of the mobile platform, called Track Crawler Robot (TCR), which is capable of traversing over and in between the rails to capture visual and thermal videos of the undercarriage in places such as railyards or sidings. The ability to scan railcars safely will enable detecting and documenting their condition for improved maintenance diagnostics and preventing maintenance. The results of preliminary tests with off-the-shelf cameras capable of 120 frames-per-second (FPS) recording are also included. The tests indicate that the close proximity between the cameras and the undercarriage components pose a challenge in ensuring that clear and visible images can be captured, especially when moving at higher speeds (say, more than 5 mph) and in low lighting. Traveling on ballasts and ties also poses challenges in terms of the vibrations caused by the rough terrain. The image stabilization feature that is present in most modern cameras is particularly helpful with reducing some of the blurring effects from the vibrations.
Ahmadian, Mehdi
Theory Guided Data Science and Railway Informatics
Attoh-Okine, Nii
Railway engineers and researchers in the past couple of years have made enormous use and implementing machine learning tools in addressing various maintenance issues. These include both geometry and railway defects applications. Nevertheless, prior to the application of these machine learning tools and techniques, there exist a well-grounded physical model in the railway engineering applications based on long term field and controlled laboratory studies. The next logical stage of the analyses, should be: a) how the railway researchers will make use of existing well-established physical models in combination with the machine learning tools; b) how to address issues like data overfitting, considering the physical models; c) how to select hyperparameters in machine learning applications considering physical models; d) how do you use physical models to address covariate shift in railway data; e) how to combine both quantitative and qualitative information in addressing some predictive analytics in the railway engineering domain. Finally, cases studies will be presented demonstrating some applications of the theory guided data science in railway track engineering and the future of theory guided data science in railway applications.
Aragona, Ivan
Product Manager/Automated Maintenance Advisor/Virtual Track Walk, ENSCO Inc.
Originally from Brazil, living in the United States since 2008, Ivan Aragona is a Railroad Researcher with Master’s degree in Railroad Engineering. He is an active member of the AREMA Track Measurement and Assessment Systems (Committee 2) with research experience on wheel performance, ties, fasteners, rail wear, track geometry and rail neutral temperature. Part of the ENSCO team since 2019, Ivan is a Product Manager leading two of the ENSCO Solutions: Automated Maintenance Advisor and Virtual Track Walk helping the railroads across the World to alleviate maintenance planning challenges through centralized data management and reporting tools that integrate all information and provide intelligent decision making.
Using Comprehensive Tools to Relate Vehicle Design, Operations, and Infrastructure
Craft, Michael
As the railroad industry has matured over the course of many decades, railroads in North America have typically focused personnel on specialized areas: Track, Rolling Stock, and Operations. Railroads have been able to adequately manage how these disciplines interact with one another using well-defined operating rules and procedures, even when only a small fraction of staff had expertise bridging these disciplines. As these disciplines become more specialized (usually in pursuit of efficiency improvements), it has become vitally important for the industry to use comprehensive tools to understand how state-of-the-art mechanical and operational systems may impact infrastructure or vice versa.
Positive Train Control System Management
Efemuai, Martins
Positive Train Control is a communication-based train control system designed to prevent accident occurrence and enforce safety measures for train movements. PTC is designed to mitigate mainly the following:
-
Prevention of train derailments due to over-speeding
-
Prevention of Train-to-Train collisions
-
Prevention of unauthorized entry into a Work Block (foreman’s limit with a form B)
-
Prevent train movements through misaligned switches.
The United States railway safety improvement act of 2008 (RSIA) provided a framework to ensure that train movements are made possible using the newly deployed PTC technology, and that fully deployed PTC enabled trains would have increased reliability and availability. When PTC system failures occur, they can lead to penalties and cost overburdens resulting from prolonged train delays. To optimize the level of service, PTC System management tools collect big data from IIoT devices integrated with train-to-wayside communication devices and train management control systems. However, experience has shown that most railroads adopt a reactive maintenance and repair-only-on-failure (ROOF) strategy. With further insight on collected system data, an evidence-based asset management approach can be adopted in resolving PTC dysfunctionality for locomotive and wayside infrastructure. Currently, this subject is yet to be fully researched upon, and there is a need to explore this subject. This need will be discussed in this presentation.
Can Autonomous Track Geometry Inspection Replace Traditional Manual Track Inspection? An Industry Update
Kerchof, Brad
Class 1 railroads are currently expanding their autonomous track geometry testing hoping that this testing will eventually replace some of the FRA-mandated manual inspections that have been the industry’s standard for decades. The keys to this track inspection initiative include development of new autonomous testing technologies and data processing capabilities; establishing a measurement standard that allows the railroad and FRA to determine if rail safety is, in fact, improved by this new method; and a waiver from FRA allowing railroads to reduce the frequency of manual inspections as their autonomous testing capability progressed. This presentation looks at the technologies and processes that railroads have employed, as well as the political obstacles they have faced, as they have sought to prove to FRA that a fundamental change in the way track is inspected can enhance safety.
Rail Infrastructure Diagnostics and Prognostics using Big Data
Kleman, Daniel
Topics will include:
How BNSF utilizes advanced inspection and detection tools.
Creating machine learning for image processing.
Analyzing data to predict defects.
Building component life cycle models.
Development of a Framework for Determination of Access Charges on a Shared High-Speed Rail (HSR) Corridor using VISSIM Simulation
Kaseko, Mohamed and Boyapati, Komal Sree Teja
Shared High-Speed Rail (HSR) networks are networks where two or more railway operators use the same railway network infrastructure for train operations. It is common in such cases for operators to be required to pay an access fee or, charge, for use of the network. The access charge is designed to cover for increased capital and operating costs, as well as the impact of additional traffic congestion and incidents on the network. The objective of this study was to develop a framework for the determination of access charges based on an analysis of train operations, including normal operations and incidents, on a shared HSR system using VISSIM traffic simulation software.
Delivering the Nation’s First 200 mph+ High-Speed Rail System, The California High-Speed Rail
Lor, Rikkito
The California High-Speed Rail Program’s overall mission is to connect California’s diverse communities by fundamentally transforming how people move around the state. The high-speed rail system, when completed, will extend to Sacramento and San Diego, totaling more than 800 miles with up to 24 stations at speeds of over 200 mph. Because of the magnitude of the program, the Authority has elected to deliver the Railway System in two phases and with multiple contracts. This presentation will discuss how the Authority intends to procure and deliver the Rail System beginning with the planning phase, final design and construction of the rail infrastructure, including integration testing, and operations.
A Data Driven Approach to Predicting Rail Transverse Profile Shape
.jpeg)
Palese, Joe
Maintaining the wheel/rail interface is critical for safe and efficient operations of railways. This interface is often controlled by maintaining the shape of the wheel and the rail through wheel truing and the transverse shape of the rail through grinding/milling. This controls the contact stress distribution and subsequent rate of degradation and failure, while controlling vehicle steering. Understanding how the rail changes shape over passing wheel loads can greatly enhance the planning of these important maintenance operations. Railroads use machine vision technology to measure and monitor the in situ shape of the rail. With the advent of more frequent track inspections using autonomous inspection technology, sufficient data now exists to evaluate the evolution of rail profile shape with passing wheels. This research focused on developing a model to predict rail transverse profile based on past shape changing trends, considering only the frequently measured rail profile data. By accurately predicting the future shape of the rail profile, railroads can compare this transverse shape to a desired shape, often referred to as a template, and forecast appropriate rail grinding cycles and plans.
Early Diagnosis of Track Stability through Non-contact LiDAR Measurements and Unsupervised Machine Learning Algorithms
Radmehr, Ahmad and Ahmadian, Mehdi
An approach for diagnosing early stages of track stability in revenue service using Doppler Lidar measurements is presented. The Lidar system is installed onboard a track geometry car for non-contact measurement of the lateral and vertical track movements. An unsupervised machine learning technique is developed to identify potentially unstable track segments using an automated data analysis process in Python. A visualization platform is also created to show the analyzed segments on the Google Map to accommodate any visual inspection or track maintenance. The study indicates promising results for early diagnosis of track movement that could develop, in time, into track instability.
Condition Based Maintenance of Primary Electrical Systems
Rini, Guy
Condition Based Maintenance (CBM) of primary electrical systems for diesel powered vehicles (truck, marine and rail) can be achieved by combining onboard circuit analysis with statistical analysis. These two fundamentally sound sciences quantify the state of health (SOH) for batteries, starters, alternators and associated electrical cables. The use of the vehicle communication network allows for downloading of data, scheduling of maintenance and integration with fleet management systems. Current maintenance practices rely on test tools to analysis each component independently. This new generation of CBM analyzes the components as a system. It provides a triage of the electrical system prior to maintenance.
Development of Adjustable Perturbation Slab Test Track for Track geometry Measurement System Evaluation
Tajaddini, Ali
FRA contracted with the Transportation Technology Center, Inc. (TTCI), to design and build a High-Speed Adjustable Perturbation Slab Track on the Railroad Test Track (RTT), located at the Federal Railroad Administration’s Transportation Technology Center (TTC) in Pueblo, CO. This 500-ft test track has capabilities of precisely adjustable track geometry variations that can be used to create different types of track geometry anomalies at different wavelengths and amplitudes. FRA is using this track to test the accuracy of its track geometry measurement systems (TGMS). Track geometry of this test track is adjustable, so that a maximum vertical perturbation of 1.75 inches and lateral adjustment 1.5 inches can be installed on each rail.
In addition of checking the accuracy of the TGMS’s this track will be used to assist FRA and others to create validated vehicle models for simulations. Accurate track geometry data is very important for finding track defects, performing vehicle/track interaction simulations, and have accurate data for data analytics and track performance prediction.
Tian, Yijun
Aragona, Ivan
Automated Data Processing for Predictive Analytics and State of Good Repair/Railroad Capital Planning
Tian, Yiqun and Aragona, Ivan
Predicting the health of track plays a significant role for effectively managing track maintenance and capital planning. Identifying potentially problematic areas long before they become a critical hazard would have far-reaching consequences for the improvement of railway safety, reliability, performance and costs. A real-world end-to-end case study will be presented including track inspection vehicle datasets such as track geometry, and machine vision system data whose conditions are assessed via trained automated algorithms to optimize development of a proactive capital plan in an effort to establish and maintain a state of good repair.
.jpg)
Do We See the Infrastructure Forest for the Trees?
Woody, Sean
This presentation will discuss some of the big-picture issues related to monitoring railroad infrastructure. Topics will include thoughts about how to define the infrastructure forest’s shape and health and what can happen when monitoring and analysis focus too heavily on a particular group of trees or even the wrong trees. The talk will give examples of how FRA’s Track Research Division tries to address these issues and to successfully transfer research results to the industry.
3D Printing Rail Surface for On-Site Repair
Wang, Zhiyong
This presentation outlines 3D printing technologies for on-site repair of worn/damaged rail track surface. Laser powder deposition, hard facing, arc welding, Thermit welding, and modified resistance seam welding, etc. were explored in the past years, lab test results on hardness, tensile/compressive/residual stresses, microstructures, and distribution of micropores and microcracks are discussed in the presentation.