Two-phase Flow Holdup Adaptive Estimation Based On Kalman Filter

Two phase flow widely exists in nature and industrial production.With the development of scientific research and industrial production,parameter measurement become a highlight spot due to the increasing demand with great significance in engineering practices.With the development of manufacturing technology,accuracy of oil-water two-phase process parameter measurement has been increasing to meet the requirements of energy saving,noninvasive measurement and process control.Water holdup is a critical process parameter in oil-water two-phase flow and gas-water two-phase flow.Owing to the complexity of flow structure and time-variant fluctuation during the measurement process,there exists a big challenge for process parameter measurement and direct modeling based on sensor signal.As a result,this paper does research on state estimation according to measured data to offer a solution for water holdup measurement.In this paper,both capacitance sensor and conductance sensor are used to experiment,and upstream and downstream measured data enrich modeling information for oil-water two-phase water holdup estimation and gas-liquid two-phase water holdup estimation.The research works is as follows:(1)Overall estimation strategy of two-phase flow.Aiming at the lack of state transition description,this paper develops adaptive state transition matrix based on sensor structure in order to improve estimation model's pertinence and accuracy.(2)Oil-water two-phase flow water holdup estimation modeling.Inserting adaptive state transition method to measurement process for linear and nonlinear methods.Contextual information is used for judging the boundary of phase inversion,choosing applicable sensor to improve accuracy and expand the measurement range.(3)Gas-water two-phase flow water holdup estimation modeling.Gas-water two-phase flow has more complex characteristics of flow structure with more time-variant fluctuation than oil-water two-phase flow.Aiming at higher frequency of state transition,this paper make adjustment for estimation model from aspect of system model matching.And adaptive estimation model improve measurement accuracy by overcoming the influence from complicated noise.