The elimination of cross-terms based on the fusion of time-frequency features and its application in machine fault diagnosis

Time-frequency representation, as a commonly method for non-stationary signals processing, includes linear time-frequency representation and bilinear time-frequency representation. Though low in time-frequency concentration, there are no cross-terms in linear time-frequency representation; in contrast, bilinear time-frequency representation, high in resolution, suffer seriously from the cross-terms, which shall lead to the ambiguous representation of a signal in the time-frequency domain. A fusion strategy for elimination of cross-terms based on linear time-frequency feature and bilinear time-frequency feature was proposed, which not only reduces the cross-terms but also preserves the superb time-frequency concentration of the bilinear time-frequency analysis. Application in bearing fault diagnosis shows that this strategy is effective in exacting the time-frequency features in vibration signal.