A Hybrid Denoising Method for electromagnetic acoustic detection

The electromagnetic acoustic transducer (EMAT) consisting of racetrack coils presents directionality and hence waves propagating in side-lobe directions experience significant energy attenuation, resulting vulnerability to noise interference. To address the challenge of weak signals denoising, a novel denoising method is proposed based on a combination of the Butterworth bandpass filtering, an improved continuous wavelet transform (CWT) incorporating high order statistical (HOS) and block threshold (BT), and Wiener filtering. The proposed method is verified by means of simulations and experiments. In the simulations, a periodic permanent magnet electromagnetic acoustic transducer (PPM-EMAT) model was established to illustrate the directivity of PPM-EMAT and generate mimic shear horizontal (SH) waves to demonstrate the effectiveness of the proposed denoising method. In the experiments, the actual receiving signals from different transmitting angles were extracted based on the fabricated PPM-EMAT. Experimental results showed that the proposed method can significantly improve SNRs of the signals received at both the main-lobe direction and side-lobe direction while maintaining the signal characteristics compared with other denoising methods, especially presenting SNRs increase from 3.67 dB to 13.74 dB within a 60° beam angle of the radiation pattern. The proposed denoising method will provide a foundation for high-resolution imaging and weak signal denoising below 20 dB based on PPM-EMAT.