Automatic Extraction And Propagation Analysis Of Internal Crack In Workpieces Based On Industrial CT Images

According to statistics,more than 85% of the failure of various parts is caused by the fatigue fracture of materials,and the fatigue life analysis of workpieces has always been a hot research object.At present,the general method is using the finite element analysis technology to analyze the crack growth of the workpiece with cracks,and then calculate its growth life.It’s an important step that providing accurate shape and position data of the crack in the finite element analysis.In the conventional non-destructive testing,high-resolution Industrial CT is the only technique that can be used for the threedimensional imaging analysis of the internal fatigue crack of the workpiece,and then obtain the detailed shape and position information of the crack.However,how to use the extracted crack data to make it applicable to three-dimensional finite element analysis of crack growth,there is no good technical scheme.Based on the slice image of industrial CT,this thesis designs a method to automatically extract the crack defects from the 3D volume data,and proposes a method to obtain the non volume crack growth front edge based on the crack defect point cloud,and then imports the extracted crack front edge data into the crack growth finite element analysis for crack growth analysis,and calculates the crack growth life.The main research results of this thesis are as follows:(1)In the aspect of crack defect extraction,the traditional CV model is extended from two-dimensional to three-dimensional,so that it can directly segment the threedimensional volume data of industrial CT,and then a method based on the size of connected area is proposed to automatically determine the defect area.This method can automatically identify and locate the internal defects of the workpiece,and then automatically calculate the surface area,volume of the defects and the overall surface,volume ratio of the workpiece.That can provides the basis for defect grade judgment.(2)Aiming at the problem of interaction between crack defect data and finite element software,a parameterization method based on crack defect point cloud is proposed.Firstly,an ellipsoid is used to fit the crack defect point cloud data,and a smooth ellipsoid is obtained.Then,a space plane intersects the ellipsoid.Finally,the space smooth intersection line can be approximately replaced by the crack propagation front,so that the finite element analysis of crack growth will reduce the numerical error.(3)In view of the inconsistency between the crack defect data and the CAD model coordinate system,the automatic registration method of 3D reconstruction point cloud and CAD model is studied.The experiment shows that the combination of rough registration and fine registration can automatically register the crack parameters and CAD model to the same coordinate system,which is preparation for the crack parameters to be imported into the finite element analysis.(4)Finally,based on the registered crack parameter coordinates and CAD model,the crack parameters are imported into FRANC3 D to complete the crack growth analysis,which verifies the feasibility of this technical route.This technique has been successfully applied to the safety assessment of a key component.By using the high-resolution characteristics of industrial CT to provide accurate initial crack location and size parameters for the finite element analysis of crack growth,the crack growth life obtained by the finite element analysis is more accurate and credible.