Non-destructive Testing Of Pipeline Defects

Based on the investigating actuality and the correlated literature of the tesing oil well tubing's detection, the paper breaks through the tradition which used single transducer, and presents the uniting testing, i.e. vortex flow test and magnetic leakage flux test, after careful survey and theoretical analysis. vortex flow test adapts to detect the portrait crack and the defection within the oil well tubing's, magnetic leakage flux testing adapts to detect landscape crack ,the combingning test improve the precision and reduce the leak of detection. The test principle and each feature of the transducers are researched in this paper. The hall elements which are used in the magnetic leakage transducer and two coils , one is used to drive coil the other is used to receive , which are used as vortex flow transducer, they constitute multi-access combination inspecting. The two methods are magnetic and need to magnetize the test pieces. Permanence magnetzing technology is used in this paper. Probe is not moved and the oil well tubing goes ahead spirally in order that the probe can scan the whole tested surface in the oil tube to avoid leak of detection. In order to reduce the external interference and noise and judge the data accurately, the output signal of the transducer is preprocessed and the uninterested scattered signals and insignificant isolated points in the data are eliminated. Wavelet multi scale edge detection technique was used to eliminate noise., improve the Signal-to-Noise. The functional model, structural model and mathematical model of the data fusion are given, use two BP neural network to fuse the data , The eddy signal is processed by wavelet transform, gained the local maxima of a wavelet transform modulus and its place, use them as eigenvalue to input the fist BP data fusion network. Distill the eigenvalue of the magnetic leakage to the second BP data fusion network to judge the length, width, depth. To design threshold value for two BP neural network's export , and judge ponderance and the oil well tubing can be accepted or rejected. Finally, the computer-aided program of the defect signal process is designed in the MATLAB. The computational analysis is implemented on the base of the experimental data. It is easy and effective to detect the defect information. The whole system is simulated in the laboratory. It is easy to install. It works stably and there is not leak of detection.