Research On Defect Iterative Inversion Method Based On Magnetic Detection

At present,a lot of oil and gas pipelines in service in china have been used for decades,which have substantial security risks,and may cause major accidents such as explosions,burns and injuries.Therefore,pipeline defect detection and inversion of contours on the pipeline wall have significant meaning for regular security maintenance and predicting the pipeline leak.Research and analysis about pipeline inspection technology and inversion methods for defect signal is done,main tasks are as follows:Firstly,the background and significance of the thesis are introduced,then the two major problems are put forward,namely forward problem of the magnetic field and inversion of the magnetic field signal,which can be solved by defect inversion technique,and detailed introduction and analysis on it is done.Theoretic analysis of magnetic field is done based on Maxwell equations and research and analysis in magnetic field which has defect is done,combined with pipeline the magnetic inspection technology.Based on the principle of the optimal magnetic circuit,modeling and research of magnetic circuit and design of size parameters of various component Theoretic analysis of magnetic field is done based on Maxwell equations and research and analysis in magnetic field which has defect is done,combined with pipeline the magnetic inspection technology.Based on the principle of the optimal magnetic circuit,modeling and research of magnetic circuit and design of size parameters of various component models are prepared for the solution to forward problem of magnetic field.Second,research about solution method of forward problem in magnetic field are done,after comparing analytical method with numerical method,eventually determine to adopt finite element method for representative of numerical method to solve magnetic field distribution at defects,and use software of finite element analysis ANSYS to build magnetic detection model and do three dimensional simulation,detailed research has been done about the influence on magnetic field signal which defects model size,liftoff other factors have,find the characteristics of distribution of the axial,radial and circumferential components of magnetic field signal,and solving the forward model can provide data support for subsequent defect inversion studies.Then,research on magnetic signal inversion,namely inversion method of pipeline defects,establishing forward model of iterative algorithm,and determining the inversion algorithm were done.Using conjugate gradient iterative algorithm based on finite element model,defect model inversion was done and initial value of inversion parameter was selected.By analyzing various inversion results,we found that when defect model depth is larger,the correct inversion results can't be gotten.Then analysis of the reasons for this phenomenon is done and improved algorithm is put forward.Finally,to overcome phenomenon that Conjugate gradient algorithm falls into local extreme and leads to defects inversion failure,the improved algorithm is put forward,which combines original algorithm with simulated annealing algorithm with global convergence,and regard simulated annealing algorithm as main subject,when iteration model reaches a minimum,adopt conjugate gradient algorithm to realize fast convergence,finally getting the global optimal solution.The inversion of 5 different depths of defect models are done by for inversion through the inversion results,we can draw the conclusion that improved algorithm performs better.