The MFL(Magnetic Flux Leakage) type PIG(Pipelines Inspection Gauge) is commonly used to detect defects for underground pipeline as one of nondestructive testing instruments. This MFL type nondestructive testing has been proved to be a reliable and effective ways of detection of defects.
In-line inspection of MFL PIG is commonly used to examine a large portion of the long distance transmission pipeline which transports natural gas from gathering points to local companies. The operating pressure of this pipeline is over 20 atm and the diameter of pipe is more than 20 inches. So, MFL PIG is moved automatically from the front and back pressure difference. But in case of city gas pipes, because the operating pressure of pipeline is less than 10 atm and the diameter of pipe is less than 20 inches, previous MFL PIG is hard to be applicable, so it is necessary for in-pipe robot system to tow the MFL Inspection system.
The basic principle of MFL PIG is that it generates a strong magnetic field on the pipe wall to be magnetically saturated. That’s because the leakage field signal in the vicinity of defects on the pipe could be maximized. Because the common structure of MFL PIG consists of a fixed permanent magnet as the source of magneto-motive force, it cannot change the path of magnetic fields on the pipe. In this structure, the magnetic field is always distributed on the pipe wall and a strong magnetic attractive force is always generated between the pipe and MFL PIG in order to reduce the magnetic reluctance in an air gap. Even though it is enough to pass in straight pipelines, when MFL module passes through bent pipe or welded pipe joint, this module is too much on one side of pipe unsymmetrically and then the magnetic force between pipe and MFL module is increased sharply because of the reduction of an air gap distance so that the PIG module could be stuck eventually. Therefore, this paper proposes a new shunting structure of MFL PIG to decrease stuck forces on the pipe wall. The magnetic forces on the PIG with respect to gap distance are analyzed to estimate necessary traction forces of an in-pipe snake robot and magnetic stuck forces could be decreased remarkably by using new shunting structure. So, to estimate the total traction force of a driving module of in-pipe robot system, both the weight of entire robot and the magnetic force acting on the pipe must be considered. The distribution of magnetic field is computed using 3-D FEM to derive a magnetic force on the pipe.
In MFL type nondestructive testing, a strong magnetic field produced by the permanent magnet saturates the ferromagnetic pipeline so as to leak the magnetic field around the defect. Hall sensors equipped along the pipeline measure the leakage fields, then the size of defects could be estimated by the measured signals. Sensing signals contain the size and shape information of defects, so that it is necessary to make a decomposing or estimating algorithm for the sizing and shaping of defects in the pipeline maintenance. In conventional method, the defect size of depth is estimated simultaneously with axial length and circumferential width, which makes it hard to estimate defect depth correctly because the depth signals are closely related to the length and width signals. The depth estimations are inevitably affected by the estimation errors of defect length and width together. So, the previous work for depth estimation showed some errors especially in the large size of defect depth.
In this paper, an enhanced estimation method for the sizing of defect depth is presented. The functional relationship between signal amplitude and shape factors of defects are derived by the polynomial surface fitting with respect to defect’s length and width so as to decouple the errors. To derive a decomposing algorithm, magnetic leakage signals are computed by nonlinear 3-D FEM and measured by hall sensors from standard defects with 16 inches diameter pipe specimen.
This paper also introduced RFECT(Remote Field Eddy Current Testing) system for inspection of underground pipelines. The RFEC technique is generally used for examination of ferromagnetic and conductive tubes such as pipelines. One main advantage of this technique is the high sensitivity to detect a defect on the external surface of pipe wall even if systems are located inside the pipelines. This system consists of an exciting coil to generate the alternating magnetic field and receiving sensor coils for detecting defect signals. Both coils could be wound coaxially or vertically with respect to the tested pipe. Also, RFECT system has a good performance to detect defects on the outer wall of pipelines and to pass through the curved pipe easily because it can be made even smaller than the diameter of pipelines. So many of institutes and laboratories have studied on RFECT for the past 50 years. But, there is lack of discussion about a study on the distribution and pattern of eddy current flow and magnetic fields distribution in a pipe wall. So, in this paper, new aspects for the basic principle of RFECT and the interpretation of detecting defect signal are proposed by analyzing both the magnitude change of eddy current distribution and that of phase shift in a pipe wall. Based on the RFEC phenomenon, there is remote field zone that the circumferential eddy currents induced by the source field are spread in the external surface of the pipe. In general, this remote fields are located about two times of pipe diameters from the exciter coil. Hence, the receiving coil should be placed in the remote field zone to detect the defect signal efficiently. In previous study, the unknown defect size is being mainly estimated by using phase angle difference between sensing signal and reference input signal because the amplitude of sensing signals is too small. In such a small signals, eddy current signals imposing the size and shapes of defects are mixed with induced magnetic fields of defects, which causes the distortions of sensing signals.
To increase the sizing accuracy of defects, it is important to measure the peak amplitude variation of defect signals at receiving coils exactly and analyze the distribution patterns of eddy currents in the pipeline. From the measured coil signals, it is necessary to decompose background offset and predict the pattern of defect signal in order to estimate the defect size for the maintenance of underground pipelines eventually. In the process, the amplitude variations of induced magnetic field at sensor coils are closely related to defect length and width. So, it is important to analyze and eliminate distortions of measured RFECT signals. In this paper, the distribution of eddy currents on the pipe wall is computed by finite element method to predict the electromagnetic field distortion in the vicinity of the defect. And then, the effects of the induced magnetic field on defect signals in receiving coils are analyzed with respect to the different defect sizes. Simulated results in this paper agreed well with measuring ones.