Prediction of axle fatigue life based on field measurements

To facilitate the adoption of a condition-based maintenance approach for railway axles, more knowledge regarding operational loading is needed. In the present work, statistical distributions on axle stresses for revenue vehicles have been derived. To this end, raw strain spectra have been gathered during field measurements using an instrumented telemetry mounted on a powered axle running within the Swedish railway network. Strain spectra are transformed into bending stress spectra which are used to estimate the statistical distributions of axle stresses for different track sections. Both the derived stress spectra and the estimated statistical distributions are used as input to fatigue life analyses. In these analyses, W\uf6hler (stress–cycle) curves estimated for varying axle surface conditions (which can be related to different axle maintenance conditions) are used to predict axle lives. The proposed method allows to rapidly post-process data obtained during field tests, to quantify indications on the health status of track and of the wheelset from these, and to estimate resulting fatigue life. This would aid in asset management by enhanced status characterisation, improved inspection and maintenance planning, and enhanced possibilities to follow-up any non-conformities