Finite Element Simulation And Experimental Study Of Stress-magnetization Effect On Steel Member

Although widely applied in the field of steel structure engineering,the traditional steel-structure non-destructive testing technologies can only detect macroscopic defects rather than early structural damages.Since its coming out in the international non-destructive monitoring academics in 20 th century,metal magnetic memory technology(MMM)has provoked great repercussion,which is expected to play a huge potential role on early detection of the structure damages.Based on the magneto-mechanical effect and the magneto-elastic effect principles of the ferromagnetic materials,early detection is achieved by detecting change of the magnetic field leakage intensity in the stress or strain concentration area.In this paper,the finite element analysis and experimental study were used to study the variation law of magnetic memory signals of specimens under different loads.The main work is as follows:(1)The basic theory of magnetic memory detection of ferromagnetic materials was analyzed,and the feasibility of metal magnetic memory detection technology applied to the detection of steel structure is verified by Q345 B uniaxial tension test.The basic variation law of the normal component of the magnetic flux leakage signals changed irregularly in the initial stage,linearly in the elastic stage,nonlinearly in the plastic stage.The zero-point and the peak-peak phenomenon occur in the fracture stage.(2)Based on the experimental study,the finite element simulation of the two-dimensional and three-dimensional models was carried out by using ANSYS software.And the change of the magnetic field leakage intensity with the change of model size was numerically simulated.The results show that: the magnetic memory signal of two-dimensional finite element simulation has some differences in amplitudefrom the experimental results,but the data of both coincide with each other,so do the magnetic flux leakage signals of both three-dimensional model and two-dimensional model.The lift-off height and the defect width have a significant effect on the normal and tangential components of the magnetic flux leakage signal,while the effect of the defect depth and the defect angle is opposite.(3)The uniaxial tensile test and three-point bending test of notched Q345 B welded steel plates were carried out.In the uniaxial tension test,there is a brittle fracture of the defective welded specimen without entering plastic stage.Its normal component of the magnetic flux leakage signal Hp(y)occurs double-peak phenomenon around the weld,whereas the non-defective welded specimen cracks at the weld.Its normal component of the magnetic flux leakage signal is continuous without double peak phenomenon.In the three-point bending test.The positions of gradient peak value of the tension side are substantially same as the compression side.In the uniaxial tensile test and the three-point bending test,the area SH and Pearson correlation coefficient r can effectively characterize the change law of the magnetic flux leakage signal.(4)Based on the support vector machine(SVM),the steel structure was identified by the classification method.The results show that the support vector machine method can effectively classify the defect categories of welded specimens.The accuracy of each kernel function is different.The accuracy of Gaussian radial basis kernel function and linear kernel function are generally higher than polynomial kernel function.