For the sake of realizing the safety detection of natural gas and petroleum pipeline welds, this paper designs a ferromagnetic pipeline weld magnetic flux leakage detector based on the calculation of the magnetic circuit of the detection probe, with the magnetization direction perpendicular to the traveling direction. The traditional pipeline magnetic flux leakage detection device uses a detection system mode in which the magnetization direction is parallel to the direction of travel. However, due to the structural characteristics of the weld, the traditional detection system mode is not applicable. Since the weld magnetic flux leakage detector needs to travel along the direction of the weld, the detector designed in this paper rotates the magnetizer 90 degrees along the direction of the weld seam so that the magnetization direction is perpendicular to the direction of travel, breaking through the technical barrier that make traditional magnetic flux leakage detection devices unsuitable for weld detection. The detection device includes a magnetizing structure, a data sampling device, and a driving and traveling device. The magnetic flux leakage signal collected by the detector is converted into a digital image in the form of a grayscale matrix. Using mathematical morphology and chain code algorithms in image processing technology, a pipeline weld defect inversion software system is developed, and a preliminary quantitative analysis of pipeline weld defects is achieved. The application of this technology enables the inspection and protection of oil and gas pipeline welds throughout their life cycle, broadens the scope of existing inspection objects, and is of great safety significance for ensuring national public security.

In order to study the internal flow characteristics of the electric submersible pump (ESP) when the gas–liquid two-phase flow is conveyed by the variable frequency variable speed operation and the change of the imported gas content, the impeller of the Q10# ESP is taken as the research object, based on the Eulerian-Eulerian non-homogeneous phase. The flow model, the unsteady Reynolds time-averaged N-S equation, and the standard k-ε turbulence model are used for transient simulation calculations of the gas–liquid two-phase flow in the impeller of the ESP. Calculations show that with the rotation of the impeller, the gas phase is unevenly distributed in the flow channel. The gas phase is mainly concentrated on the inlet side of the flow channel near the front cover, and the gas phase exhibits periodic aggregation and diffusion in the flow channel. When the impeller speed increases, the period of periodic accumulation and diffusion of gas in the flow channel is shortened and the gas concentration in the impeller decreases, the overall flow velocity in the flow channel increases, and the pressure difference between the inlet and outlet increases. The pressure difference between the two sides of the blade is proportional to the speed of the impeller, and the fluctuation frequency of the blade surface also increases. As the gas content increases, the maximum concentration of gas phase in the flow channel increases. The area occupied by the high concentration of gas phase in the flow channel expands toward the blade’s working surface, and periodically accumulates, diffuses, and grows. The gas-liquid splitting area shrinks toward the front cover side and the pump. The internal pressure increases slightly, the main flow velocity increases, and the vortex action range increases.