Time–frequency data fusion technique with application to vibration signal analysis

To overcome the inherent deficiencies of conventional time–frequency analysis (TFA) methods, i.e., different TFA methods or the same TFA method with different control parameters will present different results for the same target signal, a novel scheme named as the time–frequency data fusion (TFDF) is developed in this study by extending the idea of data fusion technique. By combining the results produced by two or more different TFA methods, the TFDF technique can present a more accurate time–frequency presentation for the target signal than what can be achieved by any individual TFA method. Therefore, the TFDF has potential to render a significantly improved time–frequency representation and greatly facilitates extracting time–frequency features of target signals. This will promote the applications of TFA in engineering practices and make TFA methods more acceptable to field engineers. The effectiveness of the TFDF technique is validated by three numerical case studies and the analysis of a rubbing-impact signal collected from a rotor test rig.

► We a time–frequency data fusion technique to overcome the diversity problem with the TFA methods.
► Two data fusion engines are suggested for the TFDF technique.
► Simulation and experiment signals are used to validate the effectiveness of the TFDF technique.