Research on non-destructive testing of stress in ferromagnetic components based on metal magnetic memory and the Barkhausen effect

In this paper, the relationship between the effective field of metal magnetic memory (MMM) emission and the structural stress of the ferromagnetic components is studied. The relationship between the voltage of the received magnetic Barkhausen noise (MBN) and the excitation source is also studied. Moreover, the characteristics of the emission spectrum of the MMM and MBN detection signals, and the influences of the lift-off distance on the stress detection accuracy of the two detection methods are studied. The advantages and disadvantages of the two magnetic non-destructive testing (NDT) techniques are analyzed separately, and their functions are combined to enhance the advantages and weaken the disadvantages. Modern signal analysis methods such as wavelet packet transform and signal feature extraction are introduced to process MMM signals, and a classifier for stress grading is established in combination with machine learning methods. On this basis, graded detection of the temperature stress of the continuous welded rail (CWR) was carried out. The detection accuracies of the 30-50 MPa, 50-70 MPa and 70-90 MPa stress grade ranges of CWR reach 71.43%, 82.76% and 75.00% respectively. The rapid grading detection of the temperature stress of the CWR and the rapid judgment of the structural stability of the CWR are realized based on the MBN-MMM detection technique.