Use of vibration signature in structural health monitoring (composite/internal damages)

Abstract

This dissertation develops and analyses the use of vibration signature as a method to detect damage in composite structures. A glass fibre pultrusion square hollow section member was used in this analysis. The member was subjected to a forced oscillation produced by a shaker. The response of the system was gathered by an accelerometer at several data acquisition points along the length of the member. The force and acceleration data obtained was then post processed to create fast Fourier transformations and frequency response functions. This process was completed for the undamaged section and several damaged scenarios. The frequency response functions of each of these scenarios at each node were then cross-referenced with the relating undamaged data using a finite difference approach. This finite difference approach produced damage index values for each node at every damage level tested. This provided an indication of the position of any potential damages. The damage indexes at each level for a known damage were then used to create an arbitrary scale that could be used to determine the extent of damage sustained. This method was completed with inherent issues pertinent to interpolation, a small data sample frequency and initial delamination of the composite. However, despite these issues the vibration damage detection technique did correctly identify damages and damage levels when a node was directly damaged. Detection of damages between data acquisition nodes was less successful. Through further code optimisation and an increase in computing power this method may provide results accurate enough to be used in engineering endeavours. It is intended that this method be used to detect damages in structures used in low safety factor situations such as in the aerospace industry. It may be used to alert maintenance to particular points of interest and alert operators of anomalous damages before catastrophic failure occurs.