Nondestructive Testing Of Steel Pipe Based On Eddy Current Array

Due to the advantages of steel pipe, it is widely used in oil and gas transportation and other industries. It has a positive impact on the national economy development. Because the tube usually works in harsh conditions for a long time, it may cause the defects like holes, cracks, wear and so on, which would cause significant economic losses. So the way to avoid economic losses or casualties is to detect the defects timely and accurately. The purpose of detecting defects is to quantify the defect and determine whether it needs to replace. Detection of steel pipe using eddy current array is realized, and quantify the defects after detection. In this thesis, the mainly work is completed as following:Firstly, the basic theory of eddy current testing is analysed. Using equivalent circuit model to analyze the factors of the impedance sensor. By the electromagnetic penetration equation, theory of skin effect is analysed which effects eddy current testing. Classification and comparison of eddy current array sensor is done. This thesis introduces the principle of eddy current array, and influencing factors of the induced voltage of detection coils. The comparison and analysis about the different incentives is introduced.The block diagram of the detection system is used to design the structure of the excitation coil, the detection array and skeleton. The hardware module and software module of the detecting system are separately presented. In hardware module, the drive module, detection module, and data acquisition and transmi ssion module is analysed and designed. In the software module, the FPGA, DSP and USB function is analysed and designed.Using the detection system to test steel bars to prove the system has the ability to detect the cracks. During the detection, it used hall array and coil array to detect the cracks. After detect steel bars, using the system to realize detection of steel pipe. Also two detecting methods, hall array and coil array, are used. Different detection results are compared to judge their performance. It is found that the trend of defects influence on detect results. The E-type magnetic is used to optimize the coil detection. The differential mode of detection is presented in this detection. The results are compared and estimate the defect parameters preliminary.By the principle and characteristics of median filtering and wavelet theory, the advantages of wavelet transform and median denoising are analysed. And using the wavelet theory and the median filter on different types of defect detection sig nal to denoise. Using soft thresholding method and Sym8 wavelet when denoising. After denoising, using radial basis function neural network to realize the depth of defect quantification processing. The estimation of the defect parameters is realized by using the number of elements which have the defect signal.