Summary

In support of the Transit Cooperative Research Program (TCRP) D7 Task Order 21, MxV Rail (formerly Transportation Technology Center, Inc.) conducted Phase 3 of the Performance-Based Track Geometry (PBTG^™) project. The following tasks were completed during Phase 3:

• Port Authority Trans-Hudson (PATH) track geometry comparison analysis for data between 2013 and 2018 and between 2013 and 2020 after PATH completed the commissioning of a newly acquired track geometry car.
• Artificial neural network (NN) model development for the PA5 car body accelerations, wheel forces, and evaluation of PBTG technology viability to assess ride quality for the PATH transit system.

PATH Track Geometry Data Comparison Analysis – 2013 vs. 2018 and 2013 vs. 2020

PATH track geometry was surveyed during Phase 2 in 2013. A track geometry comparison analysis was conducted for data collected between 2013 and 2018 as well as between 2013 and 2020 to analyze the extent of the changes over time. Critical areas in the 2013 data were identified by selecting locations with high ride quality values. These ride quality values were collected while the test crew was riding the PA5 car on the PATH network during Phase 2 of the project. The comparative analysis focused on these critical areas during each designated year.¹

For the 2013 versus 2018 geometry data comparison analysis, MxV Rail concluded that several areas of interest characterized by larger track geometry deviations in the data indicated differences between 2013 and 2018. The observed differences could result in a potentially different dynamic response from the PA5 passenger car operating over the track deviations during each respective year and, therefore, could change the ride quality output.

During the analysis of the 2013 and 2018 track geometry data, PATH was in the process of commissioning a newly acquired track geometry measurement system. Once PATH completed the commissioning, MxV Rail requested and received recent track geometry data collected in late 2020from the new track geometry system. MxV Rail then compared the 2020 data to the 2013 data in the track critical areas and determined the compared data segments had similar median and interquartile range (IQR) values but significant differences in the extreme values (maximums and minimums). Track locations with high geometry deviations were expected to generate adverse ride quality. Multiple areas with assumed adverse ride quality had a poor match between 2013 track geometry data and the data collected in 2020.

MxV Rail presented the analysis findings to the TCRP D-7 oversight panel and discussed the track geometry changes over a seven-year span (2013–2020). The track geometry variations between 2013 and 2020 were deemed acceptable, and MxV Rail proceeded with using the 2013 PA5 car ride quality (RQ) and geometry data to conduct the simulation work and evaluate the viability of the PBTG technology for transit systems.

___________________

¹ Charity Duran Ketchum, Abe Meddah, “Performance Based Track Geometry Phase 2,” Final Report for the Transit Cooperative Research Program (TCRP) Project D-07/Task 19, September 2014.

NN Model Development and RQ Assessment Using PBTG

The NN models were developed for the following PA5 car accelerations and wheel forces:

• Accelerations:
  • Driver cab maximum lateral and vertical accelerations
  • Car body trailing maximum lateral and vertical accelerations
• Wheel-rail forces:
  • Maximum lateral/vertical force ratio (L/V ratio)–left and right lead axle wheels of the lead truck
  • Minimum vertical force–left and right lead axle wheels of the lead truck

The NN model prediction confidence ranged from 60 to 70 percent for the car body lateral and vertical accelerations. For the maximum wheel L/V ratio, the prediction confidence was 73 percent for the left wheel and 60 percent for the right wheel. For the minimum vertical wheel force, the prediction confidence was 16 percent and 18 percent for left and right wheels, respectively.

Using track segments with more RQ issues, the NN model prediction technique still showed merit for predicted car body accelerations and L/V force ratios, but the model results do not meet MxV Rail’s standard for confidence, i.e., 80 percent or higher. The models, as they are, cannot be deployed with confidence for track inspection on the PATH system. Based on MxV Rail’s experience, the observed model performance is due to:

• Insufficient presence of degraded track geometry in PATH track segments and the corresponding adverse vehicle responses. The model needs additional degraded geometry data carrying predictive information to better learn how to recognize specific track and operating condition patterns that are likely to generate unwanted vehicle responses leading to RQ issues.
• More track distance and a wider range of track geometry and vehicle response data needed to develop models at an acceptable confidence level of performance.

The Phase 2 PA5 NUCARS^®2 model was updated so the track geometry parameters used for the simulation were integrated into the simulation results and synchronized with the dynamic response. The vehicle response outputs and track geometry measurements were aligned with their corresponding track locations. The data used for developing the acceleration models was derived mostly from revenue service test data collected in 2013 from the PA5 car on PATH track. Only a few events from the NUCARS runs were included to expand the training data. These events were generated at 10 mph lower than the allowable track speed.

For the wheel-rail forces, only NUCARS output was used as no instrumented wheelsets were deployed during the 2013 revenue service testing. The PA5 NUCARS model developed during Phase 2 was validated for the car body accelerations using only test results from on-track testing.

After analyzing common railroad RQ and safety standards and evaluating the viability of current PBTG for RQ assessment, it was determined that the current PBTG could not be adapted directly for transit systems to 1) predict frequency-weighted root-mean-square (rms) accelerations based on prescribed criteria and 2) generate maintenance reports recommending RQ-related remedial actions. The current PBTG system uses a segment-based approach and variable statistics, such as maximum, minimum, standard deviations, averages, and 5th and 95th percentiles from 0.1-mile-long segments of track geometry. Significant changes to the core PBTG technology are needed to be able to accomplish these critical tasks.

___________________

² NUCARS is a registered trademark of Transportation Technology Center, Inc., Pueblo, Colorado.