System to enable rail infrastructure monitoring through the dynamic response of an operational train
This invention relates to a vehicle-based infrastructure diagnostic apparatus to assist in the maintenance of rail networks. This is a complete system—from the acquisition of vibration signals in a moving vehicle, to the presentation of the infrastructure changes to the maintenance workers. This apparatus relies on novel techniques to account for the variable speed of the train (making it useful even on non-dedicated rail vehicles) and novel techniques to normalize for various environmental conditions. This system could help maintenance departments identify and remediate anomalies more rapidly, as well as trace network deterioration over time. The apparatus consists of low-cost unobtrusive sensors, software for data analysis, and hardware that makes it easy to deploy widely even on existing rail networks.
1. A method for detecting changes in rail tracks comprising:
(a) gathering data generated by one or more sensors mounted on a railcar;
(b) applying signal processing and machine learning algorithms to said data to extract rail-specific information for one or more discrete sections of said rail track; and
(c) comparing said rail-specific information for one or more discrete sections of said rail track to historical data for said discrete sections of track to detect anomalies;
wherein said signal processing and machine learning algorithm uses a sparse constraint when solving an inverse problem to determine the profile of the track.
2. The method of claim 1 wherein said anomalies are deteriorations in said track condition or changes in said track condition due to maintenance activities.
3. The method of claim 2 wherein said anomalies include a change in the roughness of said discrete section of track over a period of time.
4. The method of claim 2 wherein said vibration sensors are oriented to measure vibrations in the vertical direction.
5. The method of claim 4 wherein said vibration sensors record vibrations as said railcar moves over said rail tracks.
6. The method of claim 2 wherein said anomalies include changes in track condition due to tamping activities.
7. The method of claim 1 wherein step (a) further comprises gathering data generated by one or more vibration sensors mounted on a railcar and a positioning system to determine the geographical position of said railcar.
8. The method of claim 7 wherein step (b) further comprises:
normalizing said data gathered from said vibration and positioning system for environmental and operational conditions; and
performing a change detection calculation to determine said anomalies in said rail tracks.
9. The method of claim 8 wherein said positioning system is a GPS receiver mounted on said railcar and further wherein said normalizing step includes correcting said data for position due to GPS errors.
10. The method of claim 8 wherein said change detection calculation comprises applying a change detection template modified for a specific type of change of interest.
11. The method of claim 10 wherein said change detection template is applied using a moving window representing a data collected from a pre-determined number of passes of a railcar over each discrete track section.
12. The method of claim 1 further comprising the step of storing said rail-specific information for discrete section of said track in a database.
13. The method of claim 1 wherein said rail-specific information is represented by the signal energy of a vibration signal generated by said one or more sensors.
14. The method of claim 1 further comprising deriving a transfer function for said railcar and a profile of track roughness that best approximates actual data gathered from said one or more sensors, given said sparse constraints.
15. The method of claim 1 wherein data generated from two or more railcars is fused together.
16. The method of claim 15 wherein said fusion of data uses a modified Kalman filter.