Reliability Analysis Of Ultrasonic Nondestructive Testing Of Stainless Steel Defects

Ultrasonic testing is one of the most widely used detection methods in all non-destructive testing means.It is an indispensable testing method especially in the industrial fields such as aviation,shipbuilding and nuclear industries.With the continuous improvement of the requirements for inspection efficiency and detection reliability in industrial production,ultrasonic inspection is also required to be more rapid,and the description of defects needs to be more comprehensive and accurate.The task of non-destructive testing is to enture that the defects in materials or parts are detected before use or during periodic maintenance on the premise of satisfying certain detection reliability.Defects are detected before they reach a certain size through nondestructive testing,so that they do not grow beyond critical size due to stress corrosion or fatigue during use,which is essential to ensure the safety and structural integrity of the devices or equipment.However,the signal of the NDT depends on various factors,including the geometry of the defect,material properties,noise,human factors,etc.,so whether the defect can be detected,accompanied by uncertainty.In order to solve this problem,the predecessors proposed and the concept of probability of detection?POD?,which is used to describe the reliability of NDT.In this paper,the ultrasonic nondestructive testing is utilized to detect the stress corrosion cracks in the austenitic stainless steel welds.The POD is regarded as the index to evaluate the reliability of ultrasonic nondestructive testing.The simulation and physical detection methods are used to carry out the study from following two aspects:?1?The simulation data of 2D and 3D simulation models with welded parts are constructed by ComWAVE software based on finite element method?FEM?.The influence of the tilt angle of the defect and the displacement of the probe on the POD result is analyzed by two-dimensional data.The results show that the defect which is tilted in the negative direction is more reliable in the range of 10°tilt,when the depth of the defect is 0.4671mm or more,the ultrasonic test can detect the presence of the defect with a probability of 90%and 95%confidence;in the range of 2 mm for displacement,the a_90/95 is smaller while probe moving in the negative direction,0.3341mm.Based on the 3D data,the effect of defect length on detection reliability is further considered on the basis of 1D-POD,and the two-parameter POD curve of ultrasonic nondestructive testing reliability under different defect depths and lengths is studied.?2?42 stainless steel flat specimens with different depth and length defects are tested using UI-25 ultrasonic flaw detector,and corresponding 3D simulations are carried out according to experimental parameters.Based on the experimental and simulation results,the reliability analysis is conducted using the proposed 2D-POD model.The curve shows that the dependence of the POD on these two parameters and it's consistent with the general detectability of the defect:even if the defect is deep,it is still difficult to be detected when its length is extremely small.According to the probability contour of 50%and 90%POD,it can be known that the value of depth and length corresponding to such a defect can be detected by the ultrasonic non-destructive testing method with 50%and 90%reliability.When defect depth is less than 3mm,the trend of curve is unstable and the confidence interval become wider.The reason for this is that the mesh size is not small enough in 3D simulation,which leads to the poor accuracy of calculation,and the total amount of data is less,resulting in unreliable data in the small defect range.These issues will be improved in the future to obtain more accurate POD curves.