Gear fault diagnosis using electrical signals and its application to wind power systems

The objective of this paper is to investigate the detection of damage in electromechanical equipment using the electrical output signals. This can help designers in eliminating sensor redundancy as well as being a non-invasive method and a remote monitoring approach. As an example, a coupled gearbox and generator model of a wind turbine system is created. A crack in a gear tooth is introduced and modeled as a reduction in gear tooth stiffness during the meshing of the gears. The simulation results show that the angular acceleration at the side bands of mesh frequency increases according to damage severity, which is as expected. However, the frequency spectrum of electrical torque shows distinct peaks due to the gear tooth crack. It is also shown that cracks in different gears can be detected at distinct frequencies, which shows the possibility of identifying the source of cracks, providing an opportunity for diagnostics.