Real Time Hunting Mitigation: Sensor Systems to Detect Hazardous Conditions in High Speed Rail Operations

With the advances in train speeds coupled with lighter weight cars, the problem of truck hunting is expected to grow. Even though modern truck designs are performing excellently in raising the critical speed where hunting occurs, no truck design can eliminate the tendency to hunt altogether. Also, as components wear, the critical speeds drop; often in unpredictable ways. In order to ensure safe high speed rail operations, a new technology is required to continuously monitor the performance of the train cars to alert the engineer if hunting is occurring so appropriate remedial action can be taken. In order to be effective, such a system must be able to differentiate between hunting and other vibration patterns which may not be a cause for alarm to limit the incidences of false alarms. Using advanced digital signal processing, International Electronic Machines Corporation (IEM) analyzed truck vibration data and identified the specific vibration signature of hunting. This signature clearly differentiates hunting from signals generated by such transient phenomena as switch points and rough track. This approach has been confirmed through three distinct methods: simulation, analysis of hunting data developed by the TTX Corporation, and the analysis of actual truck vibration data gathered by a prototype system developed by IEM used in six test runs on AMTRAK cars operating in the NE Corridor. The most significant research finding is a precise frequency spectra corresponding to the onset of hunting. This frequency spectra appears to be constant at all speeds and amplitudes. This means that the onset of hunting can be identified at speeds and amplitudes well below levels associated with safety or ride quality concerns. Once hunting has been identified, other signal signatures such as peak amplitudes of the root mean square (RMS) values of the lateral oscillations, slope of the RMS values, and duration of the oscillation pattern can be used to determine the significance and severity of the hunting event so that appropriate warnings can be generated and remedial action can be taken in a useful time frame. The second significant finding of the research program is the characteristic signature of a well-tuned truck to a lateral disturbance. This algorithm, based on a polynomial function representation of the RMS-peak-amplitudes of a transient response, proved to be very reliable given the limited amount of data available. This signature can be used to satisfy the fleet screening function called for in the original solicitation. Trucks which show a slower than normal recovery rate can be tagged for inspection. Furthermore, the same analytic approach which generated this signature can be used to identify additional signatures related to a broad assortment of truck defects including: worn wheels, flat spots, bearing defects, and worn dampers. Also, by interfacing a geographic database with the hunting detection system, it will be possible to classify track conditions based on their impact on ride quality and hunting. Additional research incorporating an expanded database from a variety of trucks in field service would enable us to more specifically identify such signatures. This technique has the potential to provide a valuable new ability to develop cost-effective fleet maintenance practices. In sum, the feasibility of using digital spectrum analysis techniques to determine the onset of hunting has been demonstrated The approach has been shown to be able to provide reliable information about the propensity of a specific truck to hunt by comparing the response of different trucks to known lateral disturbances at a constant speed. Finally, the robustness of the approach is such that it has a broad potential to identify a wide variety of truck defects along with the ability to classify track conditions relative to impacts on ride quality and truck performance. Additional research is required to acquire additional data from a broad variety of truck in service to further demonstrate the utility of the approach.

  • Record URL:
  • Corporate Authors:

    International Electronic Machines Corporation

    Department of Research and Development
    60 Fourth Avenue
    Albany, NY  United States  12202-1924

    Federal Railroad Administration

    1200 New Jersey Avenue, SE
    Washington, DC  United States  20590
  • Authors:
    • Mian, Zahid F
    • Arthur, D Richard
  • Publication Date: 1996-7


  • English

Media Info

  • Media Type: Digital/other
  • Features: Appendices; Figures; Photos; References;
  • Pagination: 139p

Subject/Index Terms

Filing Info

  • Accession Number: 01679508
  • Record Type: Publication
  • Files: TRIS, ATRI, USDOT
  • Created Date: Aug 13 2018 12:09PM