Research And Implementation Of Signal Processing Methods Of Magnetostrictive Guided Wave Based On Phase Space Reconstruction

The magnetostrictive guided wave technique receives many attentions in recent years because of its non-contact and long detection distance features. However, as its energy conversion efficiency is low, the signal to noise ration(SNR) of the test signal is low too. Sources of the noise are complex, making traditional linear filtering method hard to eliminate the noise. To solve problems above, this paper focuses on the phase space reconstruction technique to study the signal processing method of the magnetostrictive guided wave on whose feature based.Firstly, chaotic characteristics, including the Lyapunove exponent, the correlation dimension and the Kolmogorov entropy, are proposed to detect the defect based on the phase space reconstruction technique. Methods to choice the embedding dimension and the time delay are proposed based on features of the magnetostrictive guided wave signal. Simulation results showed that the defect will make the magnetostrictive guided wave system more chaos and it also suggested that the Lyapunov exponent is more suitable to detect the defect. Experiment results suggested that even in a noisy environment, the Lyapunov exponent can also detect the defect.Secondly, the singular value decomposition is proposed to reveal the phase feature of the magnetostrictive guided wave signal, thus detect and locate the defect. This method is also based on the phase space reconstruction technique. The Hankel matrix is improved based on the phase feature of the defect signal and the improved Hankel matrix is sensitive to the central symmetric signal. The singular decreasing index(SDI) is proposed to detect and locate the defect even the magnetostrictive guided wave signal has low SNR based on the new matrix form. Simulation and experiment results suggested that the SDI spectrum can offer good detection and location results even in a noisy environment.Finally, the algorithms are discussed and program was developed using the MATLAB. The phase space reconstruction signal processing software was also developed by Visual Studio.