Study On The Method Of Casing Deformation Estimation Based On Pulse Eddy Current Technology

In recent years,with the rapid development of China’s social economy,the demand for oil and gas energy is increasing day by day.The effective evaluation of oil and gas well acquisition system by means of nondestructive testing has become an important means to ensure the safety of oil and gas acquisition operation.Pulsed eddy current testing(PECT)method has attracted much attention in oil and gas well structural defect detection because of its advantages of non-ferromagnetic interference and fast detection speed.In this paper,combined with related scientific research projects,for the actual needs of oil and gas well engineering,the detection instrument is placed in the tubing,to detect the casing defects,and the research on the detection method of casing defects through the tubing is carried out.This paper focuses on the following three aspects: the problem that it is difficult to remove the various noise sources in the downhole complex environment,the real-time problem of oil and gas well engineering on-line detection,and further improving the estimation accuracy of casing deformation.The specific work is as follows:1.Propose an adaptive denoising method of pulse eddy current data for double-layer tub based on noise model is proposed.Firstly,four kinds of noise,such as electromagnetic noise,jitter noise,temperature noise and tubing eccentricity noise are analyzed,and the depth weight is introduced to establish the mixed noise model.Secondly,the overall deep learning network model based on the noise model parameters is designed,and the eddy current data of different well sections are trained to obtain the noise fitting model.Thirdly,the noise fitting model of the whole double-layer tub is obtained.Lastly,the noise part of double-layer tub is removed adaptively according to the calculated noise fitting model.2.Propose a quantitative method based on improved ocean predator algorithm optimized least squares support vector machine regression(LSSVR)to detect the deformation degree of oil casing structure.In the process of updating prey vector,the similarity value is introduced to optimize the penalty parameters and kernel parameters of LSSVR,and the inversion relationship between pulsed eddy current signal and deformation degree of oil casing is established,which improves the accuracy of detection speed of line real-time detection.3.Propose a quantitative estimation method for the deformation degree of oil casing based on convolution capsule network.Firstly,several convolutions are designed to extract the characteristics of eddy current signals from different probes.Then the output layer based on capsule network is designed to construct the constraint function based on module length to estimate the minimum arm value quantitatively.