Stress Detection And Evaluation Of Ferromagnetic Materials Based On Magnetic Barkhausen Effect

During service,ferromagnetic materials can be subjected to varying degrees of stress damage due to the harsh environment,resulting in reduced performance and even catastrophic accidents.Therefore,timely stress damage detection of ferromagnetic materials is an important tool to ensure production safety.As a new non-destructive testing technique,the magnetic noise detection method uses the sensitivity of magnetic noise to changes in stress to determine the magnitude of stress on the material by detecting changes in the magnetic noise signal to assess and detect early degradation of the properties of ferromagnetic materials and stress damage.Based on the existing conditions in the laboratory,a test system based on the magnetic Barkhausen effect is developed to detect the stresses in the material and to evaluate the test results.Firstly,the mechanism of the magnetic noise signal is discussed based on the magnetic domain theory and the intensity of magnetization process of ferromagnetic materials;the main influencing factors of the magnetic noise signal are analyzed,and the excitation effect is optimized by changing the excitation source signal parameters to ensure that the material is fully magnetized and does not reach saturation too quickly.Secondly,a mathematical model was established to illustrate the relationship between stress and excitation parameters,and the signal characteristics(root mean square,mean value,standard deviation)and the envelope characteristics(half-height width,peak-to-peak value,peak)of the magnetic Barkhausen noise were extracted respectively;the time-frequency analysis of the magnetic Barkhausen noise signal was carried out by Hilbert yellow transform,and the time-frequency characteristics including the marginal spectral peak1,marginal spectral peak2 and marginal spectral energy,and the above nine eigenvalues were analyzed to initially screen the eigenvalues that provide a good characteristic description of the stress situation.Finally,as too many eigenvalues can lead to data redundancy,which affects the accuracy of stress assessment by the magnetic Barkhausen noise technique.Therefore,this paper proposes a method of principal component analysis and eigenvalue parameter correlation to select the eigenvalues,and the screened root mean square,half-height width and marginal spectral peak2 are used as principal components to re-build the multiple linear regression model and compare it with the regression model built from all the eigenvalues.The results show that the regression model established by the principal components can accurately assess material stress and is more sensitive than the conventional model,where the root mean square can be more excellent in surface features the MBN signal for stress detection.