Study On Nondestructive Testing Of Ultrasonic Tomography Of Metal Plate

With the development of society,metal sheets have been widely used in various engineering fields.While using the metal sheets,the defect of the material itself,or the defect brought about by the usage of metal-sheets-made equipment,might pose a safety issue,and even cause catastrophic results.Thus flaw detection is a crucial part and is of great importance to production safety.Lamb wave is known for its fast and highly efficient quality in the field of nondestructive testing,and is very suitable for the nondestructive testing of large-area laminal structures.Computerized tomography has already been widely used in medical area for a while.The research work of this paper is to combine ultra sonic nondestructive testing techniques and computerized tomography,to detect and recognize the surface defects of aluminum sheets,and finally to conduct segmentation and smooth processing to the ultrasonic tomography obtained by relative digital image processing techniques,and to conduct quantitative evaluation on the result of image processing,using information entropy and regional consistency parameters.The main content of this research paper is as follows:By referring widely to research papers in the world,I acknowledged lamb waves' s transmission qualities in isotropic and anisotropic materials.Since the object of this researchaluminum sheet is isotropic material,we only discuss the dispersion equations and dispersion curves of lamb waves in isotropic materials in this research paper.This paper introduced computerized tomography techniques and relative basic theories in image segmentation techniques in detail,including fundamentals in projection and back projection,central slice theorem,fan beam convolution filter back projection algorithm,pulse coupled neural network,genetic algorithm and so on,which provides theoretical basis for the extraction of the travel time of ultrasonic excitation signal,ultrasonic tomography and its image processing of aluminum sheets in finite element simulation.This paper uses ABAQUS as the finite element simulation tool to build aluminum shee t models with defects in different sizes and locations,and simulates the transmission of lamb waves in aluminum sheet,extracts the travel time by loading the sine signal of five-periods hanning-window.Uses MATLAB to write fan beam convolution filter back projection algorithm code,substitutes the travel time data into the algorithm to conduct reconstruction imaging to the aluminum sheet model.Fine imaging result was achieved,as the defect information is clearly visible in the sheets.Finally this paper uses three different image processing algorithms-global threshold segmentation,PCNN segmentation,GA improved PCNN segmentation to conduct image processing to the aluminum sheet image,and uses two quantitative evaluation standards-information entropy and regional consistency parameters to conduct analysis and evaluation to the result of image processing.The result shows that the method implemented in this research paper can detect the location,shape etc.of defects in metal sheets efficiently,and proves that GA improved PCNN segmentation works well as an image segmentation algorithm.Therefore,this paper lays groundwork for its implementation in engineering practice.