Research On Non-destructive Testing Of New Weakly Conductive Materials Using Eddy Current Magneto-optical Imaging

With the progress of China’s industrial level and the demand for high-quality development,high-strength lightweight materials such as titanium alloys,carbon fiber reinforced polymer（CFRP）and other new weakly conductive materials are being used in large quantities in aerospace manufacturing and other fields.Their damage and reliability assessment techniques are directly related to product quality and usage safety.The eddy current magneto-optical（ECMO）imaging non-destructive testing technology is widely used in the defect detection of aircraft and its components due to its advantages of noncontact,high efficiency,high resolution,unaffected by the surface state and intuition.However,the eddy current magneto-optical inspection system has not been able to detect defects in the above new weakly conductive materials,mainly due to the fact that the conductivity of these materials is 1-4 orders of magnitude lower than that of the traditional inspection materials,such as aluminum and steel,and the secondary magnetic field generated by the induced eddy currents is weak,which makes it difficult to extract signals.In addition,most of the identification of eddy current magneto-optical imaging defects still rely on manual identification,the lack of automated intelligent detection system,low detection efficiency.In order to efficiently detect defects in titanium alloy and CFRP parts with high resolution and non-contact,this thesis carries out a systematic and in-depth research on eddy current magneto-optical imaging defect detection technology,clarifies the spatial distribution law of the magnetic field on the material surface,and proposes a regional phase imaging detection method,which significantly improves the sensitivity of defect detection in weakly conductive materials.The system has realized the detection of crack defects on the surface of titanium alloys,as well as cracks,delaminations and impact defects in CFRP;and proposed an intelligent data processing algorithm,which realizes the automatic identification of defects in a highly efficient way.The specific research contents and main innovations are as follows:1.The sensitive region of the magnetic field of eddy current excitation space is clarified,which lays a theoretical foundation for high sensitivity detection.Through model derivation and finite element simulation,the distribution law of the magnetic field on the surface of the measured part under eddy current excitation is deeply analyzed,and it is clear that there exists a region with linear phase change,slow amplitude change,and sensitivity to defects on the outside of hollow excitation coils,and accordingly,a highsensitivity regional phase imaging detection method is proposed.2.Proposed a high-sensitivity magneto-optical imaging detection method for weakly conductive materials magneto-optical imaging detection laid the data foundation.In response to the low sensitivity of magnetic field detection in traditional camera imaging（usually at the mT or sub-mT level）,a combination of differential detection of balanced photodetectors and phase-locked amplification has been used to increase the magnetic field sensitivity to 110 nT/（Hz）^1/2（@300kHz）,realizing the magnetic field detection at a spatial resolution of 25 μm at the μT level,which not only lays a foundation for the detection of magneto-optical imaging of weakly conductive materials,but also reduces the conventional eddy current magneto-optical imaging excitation current from amperes（A）to tens of milliamperes（mA）,which reduces the excitation device and provides technical support for the miniaturization and portability of the eddy current magnetooptical imaging system.3.A magneto-optical imaging data processing method based on generative adversarial network has been developed,realizing efficient and intelligent recognition of defects.Aiming at the complexity of the electromagnetic coupling physical model of magneto-optical imaging,the poor quality of the data generated by a single data model,and the low efficiency of the computational complexity of a single physical model,a hybrid generative model based on the simulation of the physical model and the conversion of the data model is proposed,and a large number of data samples close to the real defects are obtained.With the intelligent processing algorithm developed on this basis,the defect recognition accuracy reaches 95%.On the basis of the theoretical analysis,a scanning eddy current magneto-optical imaging defect detection system was developed,and validation experiments were carried out using natural cracks on the surface of titanium alloys and cracks,delaminations,and impact defects of CFRP.The results show that the proposed method can not only successfully detect defects on titanium alloys,but also effectively detect anisotropic CFRP defects,with high sensitivity and good applicability,and has the value of popularization and application.