| With the gradual transformation of oil well production from low water cut to high water cut,the accuracy and reliability of the three-phase holdup parameter measurement for downhole oil,gas and water has gradually become the focus of research in related fields,and has also become an important indicator and guide for oil well production.A vital data source for tuning oilfield development.In order to solve the problem of measuring the phase holdup of oil,gas and water three-phase flow,the subject is innovatively designed based on optical fiber measurement technology and conductivity measurement technology.A fiber-optic conductance integrated probe sensor is proposed,and a classification and prediction model for the three-phase flow of oil,gas and water is proposed based on the time series pulse signal data collected by the sensor.First,due to the lack of effective research on the response characteristics and geometric parameter optimization of the fiber-optic conductance integrated sensor proposed by the research group.Therefore,this paper uses the COMSOL finite element numerical simulation method to model the sensor and analyze the change of the potential sensitivity in the measurement area of the fiber optic conductance integrated probe sensor with different structural parameters.The optimal shape of the outer wall of the insulating tube is determined by the change of the impact force of the flow field on the outer wall of the insulating tube under different shapes.In summary,the optimal structural parameters of the fiber optic conductance integrated probe sensor are determined.Secondly,a light GBM classification prediction model based on multi-dimensional features is studied for the analysis of time-series pulse fluctuation signals.The model intends to use time-frequency domain,statistics,chaos,approximate entropy and other analysis methods to obtain multi-dimensional characteristic parameters sensitive to gas holdup changes.And based on the Light GBM model,the timing pulse signals under different working conditions are predicted.Finally,the optical fiber conductance integrated probe sensor is developed in kind,and a static experimental operation platform is built.By comparing the potential changes of the two when they touch the bubbles in the water phase,the accuracy and reliability of the optical fiber conductance integrated probe sensor measurement after optimizing the structural parameters are further verified.Taking the optical fiber measurement time series pulse signal as an example,based on the Light GBM multi-classification prediction model,the data collected from the horizontal simulated well dynamic experiment in the Fengshou Base of Daqing Petroleum Corporation was learned,and the multi-classification evaluation indicators such as the accuracy of the model were obtained,which fully proved that the model can be used.Accurately predict the phase holdup of each phase of the three-phase flow. |
