No abstract

The mechanical damage of the pipeline will cause local dents of the pipe wall, accompanied by different levels of localized stresses. In this paper, the magnetic flux leakage detection technology is used to simulate and test the local dents of pipeline mechanical damage. The magnetic field distribution of the magnetic leakage detection is observed in the local stress loading state of the pipeline mechanical damage. Meanwhile, it is modeled and analyzed the equivalent change and detection effect of the magnetic flux density in the leakage magnetic field in the state that the pipeline mechanical damage is at different degree stress loading. Then the experiment is conducted with reference to the simulation results. Simulated artificial injuries with different deformation degrees of local dents are processed on the Φ323.9 steel pipe. It conducted magnetic flux leakage detection on local dents of pipeline mechanical damage by parameters of different DC magnetization strength. It showed that the leakage magnetic field effect of 5% OD pipeline mechanical damage local dents is better than 0.5% OD and 1% OD.

The ability to characterise corrosion and gouging associated with metal loss and the identification of gouge type from metal loss defect types are two of the primary obstacles affecting magnetic flux leakage (MFL) internal inspection technology. Gouges in pipelines are not extraordinarily severe; however, the depth of corrosion increasing due to a certain corrosion rate can be quite serious. In this paper, a novel theoretical model combined with a new approach for the analysis of the signals that distinguish gouging and corrosion using a low magnetisation level is presented for MFL detection, since the traditional MFL internal detection tools are insensitive to stress characteristics in the case of saturation magnetisation. A two-stage finite element (FE) model for the prediction of magnetic flux leakage resulting from two types of defect is built. In the first stage, the stress distribution associated with gouging is obtained from a solid mechanics model and, in the second stage, the stress distribution is incorporated into a magnetic finite element model by mapping the stress levels to permeability. The possibility of detection and identification of corrosion and gouging using the MFL technique at low-level magnetisation was confirmed by experimentally comparing the characteristics of the MFL signals for each defect type.

For the analysis of the magnetic flux leakage detection data in pipelines, a single information source data analysis method is used to determine the pipeline characteristics with uncertainty. A multi-source information fusion data analysis technology is proposed. This paper makes full use of the information collected by the multi-source sensors of the magnetic leakage internal detector, and adopts distributed and centralized multi-source information fusion analysis technology. First, pre-analyze and judge the information data of the auxiliary sensors (speed, pressure, temperature) of the internal magnetic flux leakage detector. Then, the data of the main sensor, ID / OD sensor, axial mileage sensor, and circumferential clock sensor of the magnetic flux leakage detector are analyzed separately. Finally, the RBF neural network + least squares support vector machine (LSSVM)fusion analysis technology is adopted to realize the fusion analysis of multi-source information. The results show that this method can effectively improve the quality and reliability of data analysis compared with traditional single information source data analysis.