In an effort to increase the number of wheels with measured profile parameters and reduce the number of condemnable wheels in service, machine vision-based wayside inspection systems are being developed to “virtually” gage all wheels of passing trains. In 2003, Transportation Technology Center, Inc. (TTCI), a wholly owned subsidiary of the Association of American Railroads (AAR), evaluated a pair of these wheel profile monitoring systems from two different vendors. Wheel-related expenses (inspection, maintenance, and replacement) make up about 37 percent of annual car maintenance costs. A significant portion of these expenses is directly related to maintenance actions associated with worn wheels. The primary indicators of worn wheels are wheel profile parameters that reach condemnable limits imposed by industry maintenance standards. These parameters include flange thickness, flange height, rim thickness, and tread hollow (hollow-worn wheels). To monitor profile parameters, inspectors attempt to visually check each wheel on inbound and outbound trains. They also measure wheel profile parameter values with steel gages on about 5 percent of the wheels annually. Each system TTCI evaluated used a different method to measure wheel profiles and determine the four primary parameters of interest. One system used lasers to highlight the wheel profile, and the other used high intensity strobes to take a picture of the wheel. Both systems used video frame capture technology and proprietary algorithms to analyze the data and calculate profile parameters. Both systems were installed at wayside locations at the Federal Railroad Administration’s Transportation Technology Center (TTC), Pueblo, Colorado. The systems were set up and evaluated over a period of several months. For each system evaluation, a test consist was assembled and run by the system at various speeds and lighting conditions. The profiles for test wheels were measured with a MiniProf® profilometer, and the four primary profile parameters were determined for each wheel prior to testing. Both systems were used during the tests to measure the wheel profiles and associated profile parameters. Through subsequent analysis, the system-derived parameters were compared to MiniProf parameter values for each test wheel to determine the tested system measurement accuracy. Both systems were found to be capable of delivering measurement accuracies of greater than 90 percent for three of the four parameters.
Review on Machine Vision based Insulator Inspection Systems for Power Distribution System