Design And Research Of Non-destructive Testing System Based On Digital X-ray Imaging Technology

In view of the current testing status of casting defects in the 239 factory,this paper designed a non-destructive testing system for X-ray digital imaging technology to realize multi-angle and all-round real-time detection of the cylinder workpiece under test,and gradually upgrade to replace the second and third.At the premise of ensuring the quality of product inspection,the Nine Factory is still using the radiographic film photography inspection technology to improve the detection efficiency,reduce the detection cost,and achieve green environmental protection testing.This paper establishes the overall scheme of NDT system based on digital X-ray imaging technology,and designs the complete mechanical part of the testing system.The forward and inverse kinematics models of the testing manipulator are established and solved by using D-H matrix theory.The simulation model of the testing manipulator is established and simulated by using ADAMS virtual prototype simulation software,and the testing manipulator is obtained.The displacement curves,angular velocity curves and angular acceleration curves of the end effector in the X,Y and Z directions are smooth,small in value and smooth in overall motion.There is no interference phenomenon and no "dead point" phenomenon,which verifies the correctness of the kinematics model.In this paper,the PLC control principle is used to design the hardware system of the detection device.The overall design scheme of the control system is given,and the linkage and synchronization control between the axes,the single-axis speed and acceleration/deceleration control,and the high-precision position control are provided.A variety of working modes,fault detection and other functions;through the PLC controller,servo motor selection and I/O address allocation design,completed the main power unit wiring diagram and PLC electrical design diagram of the detection device,for the experimental system accurate Sex and stability provide protection.The cylindrical defect image detected in the experiment has noise and the defect looks fuzzy and difficult to distinguish.This paper uses a new experimental method to process and identify the image defect.After the experiment,the adaptive threshold denoising algorithm is applied to the defect image.The noise is denoised,and the image quality is enhanced by histogram equalization.The maximum entropy segmentation method is used to achieve fast calculation and combined with binarizedmorphology processing to remove the influence of pseudo-defects on the image.Finally,according to 8-chain code tracking The method extracts the defect feature contour and combines the defect geometry with BP neural network algorithm to identify and classify the defect.Using the new experimental method for the experimental test and data analysis of the cylinder casting defect image,the leakage rate of the cylinder casting defect is 12.5%,the correctness of the cylinder casting defect type identification is 92.5%,and the false positive rate is 5%.The effectiveness of the new experimental method in this paper.