Morad Shadfar - PhD. Student, school of railway engineering, Iran University of Science and Technology,
Habibollah Molatefi - Assistant Professor, school of railway engineering, Iran University of Science and Technology,
Asghar Nasr - Assistant Professor, school of railway engineering, Iran University of Science and Technology,

Introducing an economic operation of rail system is largely depends on railway track health condition. For that, rail track is needed to be monitored regularly. Inspection of arailway track is mainly done by two methods. The global parameters known as irregularity are measured and evaluated by measuring machines like EM120. The local parameters include rail welds, spalling and squats are based on visual inspection which is costly, limited and has human error factor. So a method which has advantages compared to visual inspection is always a great concern for modern rail networks. For that, axle-box acceleration method as an indirect method of measuring wheel-rail forces are introduced. This approach is based on the instantaneous response of wheel passing over a defect on rail. The acceleration signals contain effects of the both local and global defects, but need to be extracted by different methods. Besides that wheel structural vibration has a great influence on measured data especially for low quality contact conditions. This paper deals with mentioned challenge. Tehran metro is tested and results are discussed. The selected method to analysis data is FEM for wheelset and short-time Fourier transformation for measured data.

Wheel structural vibration, Rail surface defect, Axle-box acceleration, Fourier transform, Field test