Oil-water Two-phase Flow Velocity And Its Distribution Measurement With Ultrasonic Doppler Method

Multiphase flow is commonly encountered in nature and industrial production and energy utilization.Liquid-liquid two-phase flow is a representative flow process,in which the oil-water two-phase flow is widely existed in power,pipeline transportation,and chemical/petroleum industries.The accurate measurement of the flow status and flow parameters is essential to understand the fluid mechanics.Flow velocity and its distribution are key parameters to characterize flow status in the scientific research and the safety and efficient operation of industrial process.However,the nonlinear/time-varying complex flow process and the "phase inversion"phenomenon in oil-water two-phase flow increase the difficulty of flow parameters measurement.Due to different sensing principles and application conditions,the conventional velocity detection techniques,such as the cross-correlation,conductance probe and optical LDV/PIV technique,are difficult to conduct online and non-intrusive measurement of the velocity and its distribution in most oil-water two-phase flow patterns.Ultrasonic Doppler velocity measurement technique,consisting of continuous wave ultrasonic Doppler(CWUD)and pulsed wave ultrasonic Doppler(PWUD)method,has features of non-invasion,low cost,high safety,simple physical principle,and easy implementation.It has been widely used in single-phase flow and gas-liquid bubbly flow owing to their simple flow field and high acoustic impedance difference between the phases.In oil-water two-phase flow,the non-uniform and dispersed complex flow field and low acoustic impedance difference make the scattering mechanism coupling with fluid property,droplet size,multiple scattering and other complex physical effects,complicating the Doppler velocity estimation.As summarizing the oil-water two-phase flow velocity measurement techniques and analyzing the ultrasonic scattering through theoretical and simulative study,the thesis focuses on non-intrusive and accurate measurement of oil-water two-phase flow velocity and its distribution with ultrasonic Doppler method.The thesis applies the CWUD method to obtain spatial-averaged Doppler shift related to the flow velocity of dispersed phase,and the PWUD method to obtain spatial-temporal Doppler shift related to its velocity distribution.By analyzing the features of Doppler spectrum as well as flow/acoustic field,the measurement model is established with Doppler shift,to measure the superficial flow velocity,one-dimensional(1-D)velocity profile and two-dimensional(2-D)cross-sectional velocity distribution.The main work is summarized as follows:1)According to the proposed detection method,the specified ultrasonic Doppler sensors are designed,and a set of lab-scale testing system is developed,consisting of CWUD/single-line PWUD module,multi-line PWUD module,capacitance/conductance module,NI-PXI data acquisition module and Lab VIEW control software module.Among them,the capacitance/conductance sensor provides the phase fraction needed in measurement model and individual phase velocity calculation.They are tested on an experimental facility of horizontal oil-water two-phase flow with pipe diameter of 50mm to collect sufficient testing signal under typical flow patterns(ST&MI flow,Do/w&w flow,Dw/o&Do/w flow,o/w flow and w/o flow).The overall superficial velocity ranges from 0.23 to 3.17 m/s and the water fraction ranges from 3.33%to 96.54%.2)To study the Doppler spectrum characteristic in oil-water two-phase flow and obtain the superficial flow velocity,a same-side dual-chip CWUD method is proposed and the sensor structure is optimized based on theoretical calculation and numerical simulation of ultrasonic propagation.The modulation of flow field on Doppler signal is investigated through the qualitative and quantitative analysis on Doppler spectrum/shift.On this basis,by introducing "ablation study"method to select the effective flow pattern features from Doppler spectrum,a DDAG(Decision Directed Acyclic Graph)based multi-class SVM flow pattern classification model is established,and the averaged identification accuracy for all flow patterns can reach the level of 95%.Based on the dual-slope linear relationship between the averaged Doppler shift and flow status,an overall superficial velocity measurement model is established.Experiment result shows the relative measurement error is less than 4%.3)To measure and parameterize the 1-D velocity profile and obtain more accurate superficial flow velocity,a single-line PWUD method is proposed.For ensuring the signal-noise-ratio of Doppler shift information in low amplitude scattering signal,the transducer structure is optimized and the ultrasonic pulse Doppler method(UDM)method are implemented to measure velocity profile.According to the velocity profile feature under different flow statuses,a dimensionless velocity distribution model is established for parametrized representation of oil-water two-phase flow velocity field.Considering the interphase slip or non-slip,two overall superficial velocity measurement models are established.The non-slip-based model regards the droplet’s average velocity as the overall superficial velocity,delivering a relative measurement error of about 6%.Combined with the flow structure and droplet force analysis,the slip-based model takes relative velocity between two phases into account,and reducing the measurement error to less than 3%.4)To further improve the accuracy of 1-D velocity profile accuracy and obtain 2-D cross-sectional velocity distribution of non-uniform and non-axisymmetric flow field,a multi-line PWUD method is proposed.By analyzing the effect of flow/acoustic field on scattering signal and Doppler effect,four key factors causing the velocity profile measurement error are determined.A novel "one-transmitting-three-receiving"(1T-3R)velocity correction model is designed to reduce the velocity profile error by using three adjacent transducers.Based on the sound propagation distance and Doppler angle,a more accurate local velocity is obtained with a weighted summation of three Doppler velocities.By 2-D interpolating the multiple velocity profiles from PWUD sensor array,the cross-sectional velocity distribution is obtained.To validate the effectiveness and performance of the proposed 1T-3R method,a novel Doppler signal numerical computation model is developed with given flow/acoustic field properties,testing system parameters and noise input.The numerical results show the relative measurement error of 1-D and 2-D velocity distribution measured with 1T-3R method are reduced by 75%and 80%compared with the single-line PWUD method.And a vertical experimental set-up of oil droplets rising is designed to experimentally verify the 1T-3R method,obtained cross-sectional velocity distributions are consistent with experimental observations.