Research On Imaging Technology Of Ultrasonic TOFD Weld Defect Detection

Major safety accidents in the national economy construction have a strong effect on the development,While material flaws in the operation components are significant factors that cannot be neglected.Time of flight diffraction(TOFD,Time Of Flight Diffraction)method is a precise nondestructive testing technique that is sensitive to acute flaws.Besides in traffic and transportation,this method is widely used in spaceflight,navigation,petrochemical industry and national defense.However,due to the effect of the imaging characteristics,detection system and the acoustic characteristics,the ultrasonic TOFD images have serious noise and low resolution,which causes the feature information of the flaws difficult to be obtained.Thus,how to detect the images with low resolution and how to reduce the noise are key issues in ultrasonic TOFD detection.This paper first introduces the theory of ultrasonic inspection and imaging,establishes the theoretical model of simulated ultrasonic TOFD imaging,studies the imaging law and characteristics through the process of simulated ultrasonic TOFD inspection and imaging,so as to find the method to improve the resolution level of ultrasonic TOFD imaging,at the same time,analyzes the factors and sources of image noise,studies the method to reduce noise,and finally establishes ultrasonic through the research TOFD imaging processing experimental system.Besides,analyzes the factors that produce the clutter of TOFD detection image,establishes the noise model that is close to the noise characteristics of TOFD detection image,and carries out simulation experiments.The characteristics of detection image clutter are compared with the experimental results of various filters.It is determined that the low-pass filter is more suitable.The experimental results show that the effect of clutter suppression is more obvious by using the low-pass filter.According to the basic theory of wavelet transform,the characteristics of several main wavelet de-noising methods are analyzed and compared,and the soft threshold application method of high order gradual and gentle zero setting is studied.At the same time,based on the general threshold,an adaptive threshold selection method is proposed which changes with the number of wavelet decomposition layers.The experimental results show that the signal-to-noise ratio of the image is increased from 19.11 db to 31.84 db by the improved wavelet denoising method.According to the principle of ultrasonic TOFD detection,the experimental system of weld ultrasonic TOFD detection is designed and completed.The TOFD detection image is processed in two steps: firstly,the detection image is translated and superimposed,and then the improved wavelet method is used for noise reduction.The experimental results show that the hybrid processing method can not only greatly reduce the image noise,but also effectively improve the detection image resolution level.