Planned Investigation

The Massachusetts Bay Transportation Authority (MBTA) and the Washington Metropolitan Area Transit Authority (WMATA) have participated in this project, including field testing on their tracks. The investigator established agreements with the participating transit systems to utilize geometry data gathered on their respective systems. A software routine was developed to scan the geometry track data and determine the location of curve points. Curve data were selected from the database using the curve point identification procedure developed in the previous software routine. Semi-automated curve designs were developed from the manual curve designs. The curve design model was tested.

The project has been successful in defining a continuous analytical function of the "desired curve" that can be made to best fit the measured track geometry curve and serve as a model for track alignment and crosslevel corrections. It has been shown that the desired track is not always a theoretical mathematical model but is rather more complex. The effort has developed a method whereby a digital foot-by-foot description of the regular safe accepted track configuration can be computer generated from automated track geometry measurement data. The process finds the exact position of curve points for the curve model. During the automated search for a solution, the half throws are minimized to within expressed limits. The process therefore provides valid track geometry exceptions as well as tamper input data for correcting the exceptions.

The project results include finding solutions for all 16 curves on a section of track. The tools developed for solving these curves prove that the basic model and concepts for automating the solutions are valid. Design engineers can do the selection and application of these tools manually, or they can be automatically applied in standard sequences as manual applications become routine. To date, two fully automated sequences have been found very useful and further extension of the automation is anticipated based on the elementary tools provided to date.

One challenge remaining to be solved is the synchronization of curve design data with each future geometry test. In theory, the model and the measurements must be located to the nearest foot. In practice, it has been shown that some curve point locations may shift with the seasons. Fortunately, curves in tunnels and complex trackwork do not move and can be used for automatic location detection.

Avenues of further investigation could include (a) making seasonal models of the track design to accommodate seasonal changes in the acceptable track configurations, (b) comparing the model curve points with the original design charts, and (c) proving in practice that the tamper data provided can be synchronized to eliminate the exceptions.

The method used in this project has been to model the existing track construction as the accepted design and to calculate the deviations from that design and restore the track to its accepted condition with automatic equipment. This is the process used to eliminate false exceptions in geometry reports and to generate valid input data for tamping and lining machinery. Fully implemented, the process can provide automatic location detection of track geometry measurements to the nearest foot. This method would improve the efficiency of rail transit track evaluation and of tamper maintenance.