Feature Extraction Of Barkhausen Signal And Its Application

Barkhausen-based nondestructive testing(NDT)technology has received extensive attention in recent years,especially in the application of the Barkhausen signal to the performance of materials and the detection and application of the damage situation has made gratifying research results.Commonly used eigenvalues for Barkhausen signal analysis include root mean square,envelope peak,envelope half-width and so on.However,due to the randomicity of the Barkhausen signal,these signal characteristics in different time detection results are greater differences resulting in greater error,therefore,the new Barkhausen signal to improve the detection accuracy of the features Method is very necessary.In this thesis,the mechanism of Barkhausen signal is studied theoretically,the reason of magnetic domain and magnetic domain wall is explained,the reversible and irreversible displacement process of magnetic domain wall is deduced,and the mechanism of Barkhausen signal is analyzed microscopically.The influence factors of Barkhausen signal are analyzed and the experimental results are given.Then,the overall framework of the testing system based on the Barkhausen effect is introduced.The design process of each module of the system is described in detail.,Including the design of excitation module,the design of signal detection module,the design of signal conditioning circuit,the selection of acquisition card and the design of stress loading platform.Secondly,two new Barkhausen signal feature extraction methods are proposed: based on probability distribution,Bertrand-Barthesen signal is extracted by Bachhausen signal.At last,the stress detection and material classification of Barkhausen signal are realized by the Barkhausen signal test system.At the same time,Different characteristics of the test results were compared.The results show that the new Eigenvalue Hilbert marginal spectral energy and the Barkhausen signal amplitude distribution parameter are more accurate than the conventional eigenvalue detection method.The recognition rate of the new feature of Barkhausen signal is higher than Routine feature detection,and the required training sample size is not large.