Entropy-Matching-Based Multiple Squeezing Velocity Synchrosqueezing Frequency Modulation Transform and Its Application to Bearing Fault Diagnosis
Traditional time-frequency analysis methods perform poorly in dealing with multi-component non-stationary signals,espe-cially facing noise interference and rapid changes in time-frequency characteristics,thus the clarity and accuracy of time-frequency repre-sentations are seriously affected.Based on the theory of multiple squeezing velocity synchrosqueezing chirplet transform(MSVSCT)de-rived from chirplet transform(CT),the time-frequency analysis algorithm based on entropy-matching-based multiple squeezing velocity synchronous chirplet transform(E-MSVSCT)was proposed.Multiple squeezing velocity synchrosqueezing chirplet transform was investiga-ted,and the E-MSVSCT was proposed by optimizing the rotational parameters using Renyi entropy.Finally,the effectiveness of the pro-posed method was verified by using the simulated signals and laboratory bearing fault data.The experimental results show that E-MSVSCT is able to accurately identify the time-varying fault characteristic frequency and their multiplicative frequencies of bearings,it has the low-est Renyi entropy value compared to other methods,and exhibits higher time-frequency resolution and noise robustness.
NETL
NSTL
万方数据
