Rail condition monitoring from vibroacoustic data
Streetcar turnouts are usually inspected by trained personnel every six months and receive a condition grade between A-D based on various criteria. The assessment is carried out according to the specifications of the respective network operator and is neither nationally nor internationally uniform. In order to make such inspections more efficient in the long term and to support employees with automatically generated measurement results, the basis for a procedure for assessing switch wear in the course of measurement runs is to be laid within the framework of STRAWE. The project includes the instrumentation of control vehicles in the Vienna streetcar network in order to create a vibro-acoustic data basis for different vehicles and to enable comparison with the currently operating track measurement car. The latter was equipped with a measurement microphone from the recording studio area, which was installed at an acoustically favorable position protected from contamination. In addition, triaxial vibration acceleration sensors are mounted on the axle bearings, on the bogie and on the car body.
Based on the vibroacoustic data from network-wide travels with the track measuring car, algorithms for the detection of curve screeches and ripples on grooved rails have already been implemented. For this purpose, regression models of the measured levels over the travel speed and the travelled curve radius were formed. In addition, a large number of features were determined, which were incorporated into the classification algorithm. For the automatic detection of turnouts, sound and vibration acceleration data at the axle bearings of the track measuring car were used in the preceding project ASSESS and an acoustic evaluation of the turnouts was implemented. In addition, a new algorithm for the laser light-section method is to be developed in the current project in order to assess switch wear as a combination of optical and vibro-acoustic conditions.
The STRAWE project is funded by BMK and FFG within the framework of the MdZ call.