Research On Magnetic Optic Image Nondestructive Testing Technology For Defects Of Conductive Material

With the rapid development of the industrial and manufacturing in China,metal alloy and other material like this are more and more used in exceptional high temperature and pressure situation,such as the magnalium and titanium alloy which is very light and hard.By the way,in the process of aircraft manufacturing,many of big components are produced by the molding so that it can ensure that there no seam in the aircraft,and these seams may influence the flight performance and damage the aircraft.Thus,it is very formidable and challenging to the non-destructive testing technology for repairing the equipment and do the equipment maintenance.As a kind of high efficiency and adaptable method,eddy current testing(ECT),is widely researched by the scholar in structure defects detection in large scale equipment electric material.The ECT method is now in the qualitative and invisible developing stage.It cannot meet the requirements of fast visible quantitative detections to the large equipment because of the insufficient in quantitative and visible detection aspects.So,this dissertation proposed a method based on magnetic optic effecting based visible eddy current test(VECT)technique,which is supported by the National Natural Science Foundation of China(NSFC)project,and the dissertation focuses on the research of quantified pulsed eddy current testing technology,the mechanism of detected magnetic optic image and crack identification in magnetic optic image(MOI).And it built the corresponding mathematical model based on the PECT and MOI theory,and it completed the fast quantified visible crack detecting.The more detail research contents and innovation points in the dissertation are as below:(1)It deeply analyzes the coupling mechanism between the eddy current signal and defects information.Furthermore,the signal distribution character and the relationship between signal feature and defect geometry are studied in order to optimize the eddy current excitation parameters.In order to improve the crack information extraction algorithm,the influences of conductivity and permeability of material to the eddy current signal are researched.It is the basic theory support to visual eddy current test.(2)It builds the eddy current test mathematical model based on the research of coupling theory between material and eddy current signal.By quantitative calculation on the material of the eddy current field,the influence of the eddy current field to the magnetic distribution is obtained.It reduces the dependence to the experimental data set by set a calculating model to separate the crack geometrical parameter and electromagnetic parameters.This can provide the theoretical foundation for the visual eddy current test in signal characteristic extracting.(3)Through the hierarchical division eddy current skin depth,it proposes a material geometric parameters evaluation model of eddy current test.By using this model,it can calculate the excitation frequency response to the eddy current skin depth.The results prove that it can accurately control the eddy current skin depth,so that it supplies the technical support in eddy current excitation quantitative loading.By the way,using this model,it is helpful to reduce the quantity of reference specimen when doing the research,and it improves the eddy current test technology in two aspects of theory and engineering.(4)By studying the differences between the crack and some influences like magnetic domain of MOI in shape and change rule,it designs an algorithm to identify the crack based on the connection theory.Furthermore,In order to extract the crack image,a selfadaption image separation algorithm is proposed based on the basic magnetic optic imaging theory.The experiment results show that the crack image and the influence image can be separate very well and it is helpful to identify the crack and locate the crack in detecting image.(5)It proposes the intelligent image separation algorithm.By analyzing the information coupling rule between the pixels in the magnetic optic picture,based on which it designs the MOI customized artificial neural network calculation method,including improved artificial neural network image separation algorithm,PCNN based image separation algorithm and the slicing image based PCA fusion image separation algorithm.The results show that by using the algorithm,it can extract the crack information and locate the crack accurate.These methods hold the heavy pertinence to the magnetic optic image,and it can widely be used in different magnetic optic image detection system.Furthermore,it has a strong practical application value and research significance.