Application of an optimal wavelet transformation for rail-fastening system identification in different preloads

• Optimal wavelet transform for the modal identification of a rail clip is proposed.
• Appropriate time–frequency resolutions were achieved to estimate modal parameters.
• Effects of the fastening torque on the rail-clip modal parameters are investigated.
• The preload decreases the clip damping in contrary to its natural frequency.
• Performance of the newly defined conditional entropy is experimentally validated.

A new methodology for the modal parameter identification of a rail fastener system is proposed based on the optimal wavelet transformation. The method namely the conditional entropy criterion is proposed and the Morlet wavelet transformation is optimally tuned for the case study. It is shown that one may reach to desired frequency resolution to separate the mode shapes and simultaneously arrive at the necessary time resolution to mitigate the end effects and properly estimate the connection damping. An experimental set-up has been provided and different tests have been arranged in different fastening conditions. A parametric study on the influence of the fastening torque on the modal parameters is provided. It is proved that this new wavelet-based strategy can accurately estimate the modal frequencies and damping of such a nonlinear case.