Acoustic Detection of Crack in Aircraft Rivet Based on Entropy of Instantaneous Frequency

In the ultrasonic test of rivets, flaw signal is often confused with the interface signal, which increases the difficulty of recognition by conventional methods. Aiming at these shortcomings, a novel method of rivet test based on entropy of instantaneous frequency (IF) was proposed. Algorithm of multi-correlation of time series was applied in the preprocessing of the ultrasonic signals, and IF and its entropy could be realized through the calculation of empirical mode decomposition (EMD). The problem of false intrinsic mode functions created by EMD was overcome by this method. The position of flaws in the rivet could be got through the entropy curve because of the difference between the IF of flaw echo (including interface echo) and the IF of noise in the system. Experiments show that comparing with other time-frequency methods, this method has higher accuracy in the ultrasonic test of rivets, which can meet the requirements of aircraft out-field test and has a high application value.