Studying On Signal Modeling And Signal Processing Methods Of Coriolis Mass Flowmeter

Coriolis mass flowmeter is one of the most widely used and fastest growing flowmeters.It has many advantages such as direct mass flow measurement,unobstructed internal components,high measurement accuracy and flexible installation.Digital signal processing technology is the core of Coriolis mass flowmeter and it directly determines the measurement accuracy and the measurement stability.The mathematical model of sensor output signal is the basis of signal processing.Domestic and foreign scholars have proposed many signal processing methods for Coriolis mass flowmeter.But there is no research that compares and evaluates these signal processing methods based on different mathematical model and different applications to select the best signal processing method under different measurement conditions.The former signal models can not reflect all features of the sensor output signal with the in-depth research and are needed to be improved.For these reasons,the signal models of three typical measurement conditions are improved or established for single phase flow,batch flow and gas-liquid two-phase flow and different signal processing methods are compared and evaluated in order to maximize the measurement performance of Coriolis mass flowmeter in three kinds of typical measurement conditions.The signal processing methods and signal models proposed by domestic and foreign scholars are reviewed.The signal processing methods are classified as method in frequency domain and method in time domain according to the measurement principle.The advantages and disadvantages of different signal processing methods are analyzed and compared.It provides the basis for selecting signal processing method under different measurement conditions.The signal models are analyzed and the defects are pointed out.It provides the basis for improving the signal models.The single phase flow experiments are conducted with different types and diameters of the flow sensors.The characteristic parameters of phase difference,frequency and amplitude are extracted according to the flow sensor output signals under condition of single flow.The features of the characteristic parameters are analyzed to propose the method of determining the initial value and weight of the signal model.The best signal processing method SDTFT algorithm is selected by comparison to solve the key problems under single phase flow based on the improved signal model.The algorithm is implemented in TMS320F288335DSP chip.The experiment results show that the measurement error is within ±0.1%and the measurement repeatability is within 0.05%with the range ratio of 12:1.The batch flow experiments are conducted,and the characteristic parameters of phase difference,frequency and amplitude are extracted according to the flow sensor output signals.The function equations of characteristic parameters are established based on the analysis of the features of characteristic parameters.The best signal processing method quadrature demodulation algorithm is selected by comparison to solve the key problems under batch flow based on the improved signal model.The algorithm is implemented in TMS320F288335DSP chip.The experiment results show that the measurement error is within ±0.1%and the measurement repeatability is within 0.05%.The gas-liquid two-phase flow experiments are conducted and the characteristic parameters sequences of phase difference,frequency and amplitude are extracted according to the flow sensor output signals.The ARMA models of characteristic parameters sequences and the signal model are established.The best signal processing method zero-crossing algorithm is selected by comparison to solve the key problems under the gas-liquid two-phase flow based on the signal model.The measurement errors are corrected by BP artificial neural network to improve the measurement accuracy of the gas-liquid two-phase flow.and the measurement error is within ±1.5%.