This paper presents the application of a novel vibration based technique for detecting fatigue cracks in structures. The
method utilizes the empirical mode decomposition method (EMD) to establish an effective energy-based damage index.
To investigate the feasibility of the method, fatigue cracks of different sizes were introduced in an aluminum beam
subjected to a cyclic load under three point bending configuration. The vibration signals corresponding to the healthy and
the damaged states of the beam were acquired via piezoceramic sensors. The signals were then processed by the
proposed methodology to obtain the damage indices. In addition, for the sake of comparison, the natural frequencies of
the healthy and damaged states of the beam were obtained through the Fast Fourier Transform (FFT). The results of this
study concluded in two major observations. Firstly, the method was highly successful in not only predicting the presence
of the fatigue crack, but also in quantifying its progression. Secondly, the proposed energy-based damage index was
proved to be superior over the frequency-based method in terms of sensitivity to the damage detection and quantification.
Moreover, this technique could be regarded as an efficient non-destructive tool, since it is simple, cost effective, and
does not rely on analytical modelling of the structure.