Magneto-optical Imaging Based Defect Detection Using Low Frequency Orthogonal Excitation

The structural damage detection and health status assessment for the carbon steel are fundamental to quality control and safe operation.Due to high strength,good plasticity,impact resistance and reliable performance,the carbon steel is widely used in key components of many fields,such as petroleum,chemical,electric power,etc.magneto-optic imaging(MOI)is an intuitive and visual non-destructive testing(DNT)technology,which has been widely studied and applied in defect detection.However,current research on MOI is mainly for DC,power frequency or eddy current excitation.Limited by material hysteresis and imaging technique,MOI has a single spectrum which can only recognize simple shape defects.However,in practical applications,current detection technologies can't accurately and effectively recognize a defect because of its complicated shape.Therefore,under the support of the National Natural Science Foundation of China,based on low frequency and orthogonal excitation,this paper studies the MOI complex shape defect detection from the MOI detection theory,modeling,detection platform and recognition method.The main research and novelties are as follows,(1)Based on ring current view,solenoid model(SOL)of magnetic flux leakage(MFL)is constructed.The magnetic dipole model(MDM)and the molecular current model are deeply studied.This paper analyzes the advantages and disadvantages of those two models and then constructs the SOL based on loop current.Theoretically,this model overcomes the problem that the MDM is highly dependent on the setting of magnetic charge distribution,and analyzes the influence mechanism of the magnetoresistance,test specimen and the interval ring current on the ring current model,and solves simulation problem between small defects and large test specimen,and verifies the validity and accuracy of the solenoid model of MFL,and points out that the calculation speed of the SOL is better than finite element method(FEM).In order to realize the dynamic modeling of the MFL under alternating excitation,Jiles-Atherton(JA)hysteresis loop is introduced to construct the dynamic solenoid model of MFL of defects under alternating excitation.The dynamic solenoid model of MFL provides a reliable theoretical model for complex shape defect detection.(2)Based on the interaction forces of the solenoid,a SOL for complex shape defects is proposed,and the distribution law of the MFL of complex shape defects in orthogonal magnetic field is studied.For the unknown defect direction and shape of a test specimen,under unidirectional excitation,the MFL characteristics of simple defects and complex shape defects are analyzed.It is pointed out that unidirectional excitation can't effectively excite the MFL of simple defects and complex shape defects at different angles.In order to solve the above problems,the orthogonal excitation is introduced and the MFL characteristics of complex shape defects are analyzed.Based on the interaction forces between solenoids,a solenoid model of MFL for complex shape defection is constructed.The theoretical interpretation of the inflection point distortion and midline distortion characteristics of MFL is introduced.The proposed model is compared with the MDM and FEM,which verifies the validity of the solenoid model of MFL for complex shape defects.(3)With the low frequency and orthogonal excitation as the core,the MOI detection platform is designed,and its excitation frequency is optimized.In order to meet the requirements of effective magnetization of the test specimen,rich spectral information,high-quality MOI acquisition,etc.,the key components,such as yoke,excitation source,light source,magneto-optic sensors and image collectors of MOI detection platform,are designed.By theoretical analysis and simulation,the single-period MFL intensity information of defect MFL sequence under different orthogonal excitation frequencies are compared.And the optimal excitation frequency is selected to realize the high-precision MOI sequence acquisition of complex shape defects,which provides research platform for high-precision identification of complex shape defects.(4)Based on the time-frequency domain characteristics of MOI sequences,the pixie-level decision approach for complex shape defect detection is proposed.Combining with the characteristics of the MFL of complex shape defects,the MOI is divided into different feature regions to analyze the time-frequency domain characteristics of different regions in the MOI sequence.For the image distortion and contrast imbalance caused by the multiplicative noise and distortion,a multi-scale homomorphic filter frequency domain enhancement algorithm is proposed.As a homomorphic filtering parameter,Contrast Limited Adaptive Histogram Equalization(CLAHE)acts on the amplitude frequency domain to enhance multi-scale frequency domain from magnitude spectrum normalization and distortion area contrast equalization.The result shows that the proposed approach for frequency domain enhancement can effectively segment the distortion and accurately recognize complex shape defects.The experiments show that the proposed model can effectively detect more complex defects of ferromagnetic components.From the theoretical model,simulation and experimental results,it is proved that the MOI detection platform,based on low frequency orthogonal excitation,has wide applicability,strong practical value and guiding significance for structural damage of ferromagnetic components.