A new method to eliminate negative frequency interference based on wavelet transformation and grey correlation theory

As for the common low-frequency roll vibration signals in the fields of mechanical vibration and fault diagnosis, the negative frequency interference is an important factor to influence the precision of spectrum analysis and correction. In order to eliminate the negative frequency interference, a new method for the analysis and correction of extremely low frequency signal based on wavelet transformation and grey correlation theory is presented. First, the original signal is constructed as a standard reference sequence. Second, the original signal is transformed via one-dimensional wavelet transformation to calculate the grey correction degree between transformed signal components and the standard reference sequence, by which the grey correction sequence can be obtained. The value of the grey correction degree of negative frequency interference components is small, so the negative frequency interference components can be picked out as “useless ingredients”. At last, the remaining signal components are reconstructed to restore the original signal in which the negative frequency interference has been eliminated. Simulation experiments and verification examples have shown the effectiveness of the proposed method.