Damage Detection of Bridge Cables in Strong Noise Based on Improved Matching Pursuit Method
In order to reduce the influence of noise on magnetostrictive guided waves and realize the recognition of bridge cable defects in strong noise,an improved matching pursuit method was proposed and was used for damage detection of bridge cables in strong noise.Firstly,an asymmetric Gabor function was constructed as the atomic dictionary to match the guided wave signal in the strong noise background,and the adjacent residual ratio was used as the iterative termination condition.Then,the bridge cable defects were localized by the peak moments of the defect echo envelope.Finally,the localization errors for different noise reduction methods were compared.The results show that the maximum error is only 0.28％when the peak moment of the defect echo envelope is used to locate the bridge cable defects.When the signal-to-noise ratio of the defect echo is higher than-5 dB,the EMSE is almost zero and the ESNR increases to more than 7.56 dB by using the proposed method.The noise reduction accuracy and defect localization accuracy of the proposed method are better than wavelet threshold denoising,wavelet-EMD denoising and traditional matching pursuit denoising.When the signal-to-noise ratio of the guide wave signal is-5 dB and below,the superiority of the proposed method is more significant.Even if the signal-to-noise ratio of the guide wave signal reaches-10 dB,the defect localization error can still reach within 1.5％.It is concluded that the method proposed can effectively reduce the strong noise of guided waves and improve the defect detection accuracy.
万方数据
维普
