On the recognition of anomalies in gears by means of the discrete wavelet transform

Static transmission errors and engine torque fluctuations are two major issues that characterize the dynamic behaviour of gears. Torque fluctuations are typical, as an example, in internal combustion engine. Static transmission errors, such as tooth profile errors, tooth “spalling” defects, or wheel assembly eccentricity, are present in every actual gear pair. This paper introduces a technique that uses a wavelet-based signal processing procedure, applied to the signal coming from two meshing gears. Discrete wavelet transform (DWT) is used to decompose the meshing signal, and the wavelet coefficients variations are analyzed to describe the dynamic behaviour of gears. Some examples of application of the proposed technique are described adopting, as input data, the meshing signals acquired by two high resolution incremental encoders on a gear pair test rig. The ability of DWT in detecting either gear defects or torque fluctuation is shown. The analysis demonstrates that, by means of wavelet transform, it is possible to identify and evaluate the presence of specific anomalies in the meshing signal. The analysis also suggests the possibility of adopting such DWT based technique in an on-line control context. Finally, in the field of gear rattle, the technique can have also a practical relevancy in the definition of new rattle indices useful to the NVH departments of automotive industries.