Research On Nondestructive Testing Of Metal Defect Based On Pulsed Current Potential Drop Technique

Metal equipment is widely used in aerospace,national defense,energy and other fields,and it is essential to ensure the safety of this equipment.These metal structures may suffer general corrosion,cracks,pitting and other defects during service period,especially in the harsh,airtight and uninterrupted conditions such as oil and gas pipelines,chemical containers,etc.Pulsed current potential drop technique(PCPD)is a safe,valid and nonintrusive testing method to monitor all kinds of defects.Compared with other nondestructive technique,PCPD has the advatages of high signal precision and high anti-interference capability over the past 20 years of development,which has been used in the field of industry.This thesis will do some research on the PCPD key technologies of solving the depth of cracks and recognizing the type of defect,which based on the National Natural Science Fund project: on-line monitoring technology of metal pipeline corrosion based on Alternating Current Field Signature Method.The brief structure of the research and novel approaches are as follows.The electric field distribution and parameter setting in were studied in this paper.Finite element simulation software COMSOL has been used to build metal plate model.Based on the simulation result,the effects of the rise time of and pulse width were analyzed,and the relation between the distance of excitation electrode and measuring area was also studied.These simulation results could lead the theory innovation and application of PCPD.An improved multidirectional PCPD method was proposed.A concept of “limit angle” is given,and the effective monitoring area of the excitation electrode corresponds to the "limit angle".The number of excitation electrodes is increased according to the size of the limit angle,so that the crack depth of any angle can be calculated by using the depth solving equation of single angle crack calibration.An innovative coordinate method is proposed to solve the depth of bottom crack at any angle,which could improve the measurement accuracy and reduce the dependence on the electrode.The depth of crack can be determined by the values of measured assistant voltage and secondary voltage.However,this method is only suitable for monitoring the bottom surface cracks,and the surface cracks can be only solved by multidirectional PCPD.Defect classification is the premise of measuring the depth of crack.As PCPD dose not give the approach to classifiy and recognize diverse type defects,this thesis gives the eigenvector with combining the drag effect and PCPD.Moreover,a novel approach aim at classifying three typical defects(pit,crack,and general corrosion)has been inovatinally proposed,and the new classification method is based on support vector machine.The best classification method is selected among four commonly used classification algorithms.Verification of simulation test set and experiment validation have been implemented.Experiment results show that the classification accuracy of the novel method is more than 90%,which can identify defects well.The above research results were integrated together to verify the feasibility of solving depth of cracks in the pipeline by using the multi-direction pulsed potential drop method and the coordinate method.The measurement errors of both methods were less than 10%,which could effectively improve the measurement accuracy.These test results can provide guidance for the application and development of this technology.