Study On The Measurement Mechanism Of Two-Phase Flow By Nuclear Magnetic Resonance/Capacitance Sensor

Two-phase flow widely exists in the process of oil production wells and surface pipeline transportation.Two-phase flow parameter information reflects the change of single well productivity in oil field and affects the management and optimization of reservoir.The traditional detection method has some limitations due to the complexity and variability of the dynamic characteristics of the two-phase flow velocity and flow pattern.NMR technology and capacitance method are used in two-phase flow detection due to their advantages of nonintrusion,accurate measurement and low cost.In order to accurately measure the information of different two-phase flow parameters,this paper proposes a measurement system combining nuclear magnetic resonance and capacitance,and studies the sensor,detection circuit,measurement mechanism and indoor test.The main research work and achievements of this paper are as follows:(1)We study two types of sensors with six different structures,and determine the structural model of the sensor by analyzing the sensor performance indicators such as capacitance variation range,sensitivity and sensitivity uniformity,and phase holdup measurement results.On this basis,we use the sensitivity uniformity and the capacitance difference between full field and empty field to optimize the sensor parameters,and determine that the electrode width of the double-ring sensor is 22 mm,the electrode spacing is 16 mm,and the electrode thickness is 0.6mm.In the meanwhile,the double-ring sensor and a 12-electrode array sensor were developed for two-phase flow measurement,and a CDC-based micro-capacitance detection circuit was completed to verify the reliability,stability and noise immunity of the measurement system,reducing the average relative error of the sensor measurement to 0.4%.(2)We use multi-physics simulation software to model and analyze the RF coil,and select the coil structure of the nuclear magnetic sensor.By comparing and analyzing the magnetic field distribution uniformity of the solenoid coil and the birdcage coil,we finally chose the birdcage coil with integrated transceiver.The coil performance is optimized by capacitance scanning,and the distribution of the central magnetic field before and after the shield is further analyzed.The results show that increasing the shield can improve the uniformity of the magnetic field distribution.(3)A two-phase flow parameter prediction model based on BP neural network and MLP multi-layer perceptron neural network is proposed.We analyze the influence of temperature,inclination angle and flow pattern on phase holdup.At the same time,BP neural network is used to establish detection models for horizontal wells with different well depths,small inclinations and different flow patterns.It can be seen from the results that the error between the predicted value and the true value is small,so as to realize the prediction of phase holdup in the pipeline.We also study the influence of flow velocity and phase holdup on T2 spectrum,and use MLP multi-layer perceptron neural network to establish phase holdup and flow velocity detection model to realize the prediction of two-phase flow velocity in pipeline.(4)We use a double-ring sensor and a 12-electrode array sensor,a data acquisition system,and a data display system to build an indoor detection platform to analyze the variation of dielectric constant and capacitance values in the pipeline.The flow velocity and phase holdup are simulated by simulation,and the characteristic parameters of T2 spectrum are analyzed to realize the measurement of two-phase flow velocity.