Research On Phase Difference Estimation Of Coriolis Mass Flowmeter

The measuring tube of the Coriolis mass flowmeter performs simple harmonic vibration under the action of the exciter,so that the sensors at both ends are subjected to the action of Coriolis force to produce relative displacement,thereby realizing the direct measurement of the mass flow.The Coriolis mass flowmeter has the advantages of high stability,high measurement accuracy and versatility.Not only the measurement of mass flow,but also density and temperature can be obtained,and the industrial accuracy can reach one thousandth.However,the Coriolis mass flowmeter is susceptible to noise interference,has high requirements on the processing process,and the density changes slowly with speed,fluid medium,etc.,which reduces the measurement accuracy of mass flow.This paper firstly introduces the background,significance and research status of the research topic at home and abroad.The structure,advantages and disadvantages of Coriolis flowmeter are expounded.Then,research is carried out to improve the extraction accuracy of mass flow of Coriolis flowmeter.(1)The mass flow of a Coriolis mass flowmeter is obtained by density and phase difference.To improve the extraction accuracy of mass flow,the phase difference extraction algorithm can be started with,and the mass flow accuracy can be improved by improving the extraction accuracy of phase difference.This paper proposes to apply digital up-conversion technology to the measurement of flowmeter phase difference.Aiming at this method,this paper firstly uses MATLAB to verify the algorithm,and then designs the algorithm in blocks.The whole algorithm is divided into signal source module,frequency mixing module,filter module and phase extraction module,each module is programmed and verified.Experiments are set up to compare the method with other traditional phase difference extraction algorithms to verify the accuracy of the method.(2)Both phase difference accuracy and density are important factors that affect mass flow accuracy.The phase difference estimation algorithm is studied in the first point.Density is determined by the natural frequency of the Coriolis flowmeter structure,which is discussed in this chapter.This paper builds a modal analysis simulation model with the help of ANSYS and uses three different structures of elastic bending pipe to analyze the natural frequency.To explore the influence of different velocities and different fluid media on the natural frequency,a flow-solid coupling simulation model is built,and finally it is concluded that the fluid velocity and the fluid medium affect the natural frequency,and the natural frequency change slowly with time.(3)According to the conclusion of the second point,simply improving the phase difference estimation accuracy is not the best way to improve the mass flow estimation accuracy.In this paper,the Hilbert transform and SVR are used to build a machine learning model,with the phase difference as the feature and the mass flow as the label value to directly establish the relationship between the phase difference and the mass flow.The algorithm first filters the obtained sensor signal,and then obtains the phase difference and correlation coefficient as model features through Hilbert transform and correlation analysis.After building the SVR machine learning model,the particle swarm algorithm is used to optimize the hyperparameters of the model,and finally the extraction of mass flow is realized.