JPH10281846A - Polyphase flowmeter by pattern recognition method utilizing coriolis flowmeter - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain the flow rate of a gas and a liquid based on a statistical method by using a Coriolis flowmeter and utilizing a mass flow rate and a density signal that fluctuate according to a void rate. SOLUTION: A mixed phase flow of a gas and a liquid is fed to a Coriolis flowmeter 1, the fluctuation waveform of a mass flow rate Qm and a density ρthat fluctuate according to a void rate is inputted to a statistical feature measurement part 6, thus obtaining a statistical feature grid cell as the logic sum of a statistical vector diagram as a statistical feature. On the other hand, sample patterns with a known mass flow rate Qm and a density ρ are compared, a position M of the gas flow rate (abscissa) and the fluid flow rate (ordinate) of the statistical feature grid cell is identified by a pattern recognition part 7, and the gas flow rate and the liquid flow rate are obtained according to the two parameters of the mass flow Qm and the density ρ.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multi-phase flow meter by a pattern recognition method using a Coriolis flow meter, and more particularly, to a multi-phase flow meter proportional to a Coriolis force obtained by flowing a gas-liquid multiphase flow through a Coriolis flow meter. Calculate the statistical characteristics of various parameters such as the flow rate phase difference signal, compare with the previously determined sample pattern of the parameter, identify the position of the sample pattern on the statistical grid cell, and with the gas phase flux It relates to a multi-phase flow meter for determining the liquid flux.

[0002]

2. Description of the Related Art In a flow rate measurement, various fluids having different temperatures, pressures, densities and viscosities are measured according to purposes. When the measurement fluid is a single-phase fluid, it is relatively easy to determine the volume flow rate in the reference state by correcting environmental conditions and physical property values.
For example, in the case of a gas-liquid multiphase flow, the flow rate of each phase is determined by a conventional flow rate measuring means after separating into a gas phase and a liquid phase in order to determine the flow rate of each phase. . However, in this method, the phase separation mechanism is large, expensive and requires a large installation space, and it is almost impossible to completely separate the gas phase from a gas-liquid mixed phase flow containing a high-viscosity liquid phase. It was not possible to determine the phase flow.

[0003] For this reason, Archer et al., "Software technology for flow measurement in horizontal two-phase flow" (SPEP
roduction Engneering, August 1991), a gas and a liquid whose flow rate was measured beforehand flow into a horizontal gas-liquid multi-phase flow path, and a plurality of absolute pressure gauges, differential pressure gauges and one capacitance type flow into the gas-liquid multi-phase flow path. A void rate meter and a conductorless microwave were attached, and stochastic features were extracted and classified using the speech recognition method from the pressure waveform and void rate waveform of the gas-liquid multiphase flow flowing through the gas-liquid multiphase flow path. It has been reported that the method can determine the gas flow rate and the liquid flow rate by identifying the reference flow rate form of the multiphase flow and determine the gas flow rate and the liquid flow rate with an error of 5 to 10% with respect to the reference gas flow rate and the liquid flow rate measured as the flow rate. I have.

However, the flow measurement method of Archer et al. Requires that a horizontal gas-liquid mixed-phase flow path be provided to achieve a uniform multi-phase flow. Therefore, a long horizontal flow path is required.
Applicants previously introduced the gas-liquid multiphase flow into the swirl device, flowed out to the venturi as a homogeneous gas-liquid multiphase flow, obtained the differential pressure signal between the throat of the venturi and the enlarged taper part, and obtained the differential pressure signal. We have proposed an apparatus that measures the statistical characteristics of a sample, compares it with a sample pattern that has been calibrated in advance, and measures the flow rates of gas and liquid by pattern recognition.

As a method for measuring the flow rate of gas and liquid in a gas-liquid multi-phase flow, a venturi and a volume flow meter are connected in series in a gas-liquid multi-phase flow path, and the sum of the gas flow rate and the liquid flow rate obtained by the volume flow meter is measured. Agar proposes a method for obtaining a gas flow rate and a liquid flow rate by utilizing the fact that the differential pressure of the venturi is proportional to the flow rate ratio between the gas flow rate and the liquid flow rate.

[0006]

A method for measuring the statistical characteristics of a differential pressure signal of a venturi by introducing a gas-liquid multi-phase flow homogenized by a swirler into the venturi can reduce the size of the device. Although there is an advantage, the only parameter that represents a statistical feature is the differential pressure signal, which limits the measurement accuracy, and requires more parameters of the input signal for higher accuracy.

Further, in the method of connecting the volume flow meter and the venturi, since the volume flow meter has a movable portion, the volume flow meter has a reliability in that it may stop if it is mixed with foreign substances or used under severe conditions. I have anxiety.

The present invention has been made in view of the above-mentioned circumstances, and uses a Coriolis flow meter as a flow measuring means having no movable part. The Coriolis flow meter can measure up to 20% void ratio in the case of, for example, an ice cream stock solution, which is homogenized as the limit specification of the gas-liquid multiphase flow. The gain becomes unstable and the flow rate and density cannot be measured. In the present invention, utilizing the above-mentioned problems of the Coriolis flowmeter, by measuring the statistical characteristics of the signal waveforms of the flow rate and the density, the statistical characteristics of the gas-liquid multiphase flow having various void fractions calibrated in advance are obtained. It is intended to determine the gas flow rate and the liquid flow rate by pattern recognition as compared with a sample pattern.

[0009]

According to a first aspect of the present invention, there is provided a Coriolis flowmeter for driving a measurement tube supported at two points at a natural frequency, a gas-liquid multiphase flow through the Coriolis flowmeter, A flow rate phase difference signal corresponding to the Coriolis force acting on the flow rate and a vibration waveform having a frequency corresponding to the density of the gas-liquid multi-phase flow are detected, and the temperature-corrected flow rate signal and density signal of the gas-liquid multi-phase flow are output. A converter to determine the statistical characteristics of the flow rate signal and the density signal as the strength of the feature vector, and compare the gas phase flux and the liquid phase flux sample pattern in a predetermined gas-liquid multiphase flow, A multi-phase flow calculator for identifying a position of the sample pattern on a statistical grid cell to obtain a gas flux and a liquid flux of the gas-liquid multi-phase flow.

According to a second aspect of the present invention, in the multi-phase flow meter according to the first aspect of the present invention, a differential pressure gauge for detecting a differential pressure is provided at both ends of the measurement pipe, and the multi-phase flow meter is provided. In addition to the strength of the feature vector of the flow rate signal and the density signal obtained by the flow calculator, the strength of the feature vector of the differential pressure signal detected by the differential pressure gauge is obtained, and the sample pattern includes the differential pressure signal. is there.

According to a third aspect of the present invention, there is provided a multi-phase flow meter based on a pattern recognition method using a Coriolis flow meter according to the second aspect, wherein the flow signal, the density signal, and the difference obtained by the multi-phase flow calculator are obtained. In addition to the pressure signal, the strength of a characteristic vector of driving power for driving the measurement tube of the Coriolis flowmeter is determined, and the driving power is included in the sample pattern.

According to a fourth aspect of the present invention, there is provided a multi-phase flow meter based on the pattern recognition method using the Coriolis flow meter according to any one of the first to third aspects, wherein the liquid phase is water and oil in the gas-liquid multi-phase flow. In the case of a three-phase fluid, an oil / moisture meter is provided downstream of the Coriolis flow meter, and the flow rates of water and oil in the liquid phase are determined.

According to a fifth aspect of the present invention, there is provided a multi-phase flow meter based on the pattern recognition method using the Coriolis flow meter according to any one of the first to fourth aspects, wherein the gas-liquid mixed phase flow is provided upstream of the Coriolis flow meter. And a mixing means for uniformly mixing the gas phase and the liquid phase.

[0014]

FIG. 1 is a block diagram for explaining an embodiment of the present invention. In the drawing, reference numeral 1 denotes a Coriolis flow meter, 2 denotes an inflow pipe, 3 denotes an outflow pipe, and 4 denotes a Coriolis flow meter. 1 is a converter, 5 is a multiphase flow calculator, 6 is a statistical feature measuring unit, 7
Is a pattern recognition unit, 8 is a learning unit, 9 is a sample pattern, and 10 is an output.

A Coriolis flowmeter 1 shown in FIG. 1 has a U-shaped pipe connected to an inflow pipe 2 and an outflow pipe 3 through which a gas-liquid multiphase flow flows.
The measuring tube 1a has a V-shaped measurement tube 1a.
It is driven at a natural frequency f by a driving means (not shown) in a direction perpendicular to the paper surface with 1c as a fulcrum. The natural frequency f of the measuring tube 1a changes according to the fluid density ρ of the gas-liquid multiphase flow flowing inside. Measuring tube 1a driven at natural frequency f
Generates a torsion θ proportional to the Coriolis force, and detects this torsion as a time difference signal τ.
The temperature T of the gas-liquid multiphase flow and the natural frequency f are transmitted to the converter 4. The Coriolis flow meter 1 shown in FIG.
Has been described with reference to the U-shaped measuring tube 1a, the measuring tube may have any shape, and may be composed of one or a plurality of tubes.

The time difference signal τ transmitted to the converter 4 is a value proportional to the mass flow rate Qm of the gas-liquid multiphase flow, and the temperature T of the gas-liquid multiphase flow is the value of the measuring tube 1a related to the natural frequency f. It is a value that changes the Young's modulus, and the mass flow rate Qm and the fluid density ρ are obtained by correcting in the converter 4 based on the temperature T,
The signal is transmitted from the converter 4 to the statistical feature measuring unit 6 of the multiphase flow calculator 5.

The multi-phase flow calculator 5 includes a statistical feature measuring unit 6,
It comprises a pattern recognition unit 7, a learning unit 8, and a sample pattern 9, and is configured to output the liquid flow rate and the gas flow rate identified by the pattern recognition unit 7 from an output terminal 10.

The statistical feature measuring unit 6 measures the statistical feature by inputting the mass flow rate Qm and the fluctuation value of the density ρ generated by measuring the gas-liquid multi-phase flow. Mass flow Q resulting from the ratio of
This is to analyze the fluctuation waveform of m and density ρ. The analysis of the statistical features determines the amplitude domain features and the frequency domain features, and is performed on the fluctuation waveforms of the mass flow rate Qm and the density ρ, respectively.

The amplitude area feature is characterized by a standard deviation σ, which is a measure for determining how much the instantaneous value for each data digitally sampled from the fluctuation waveform is different from the average value, and the average value of the distribution of the fluctuation waveform. A distortion coefficient γ ₁ , which indicates the degree of asymmetry, and a sharpness coefficient γ ₂ , which indicates the degree of variation in the distribution of the fluctuation waveform. Power spectral density S for showing in area
It is. The statistical feature of the fluctuation waveform composed of the amplitude domain feature and the frequency domain feature described above is called a waveform feature vector,
For a two-dimensional coordinate system in which the horizontal axis represents the gas flux (apparent flow rate of gas: hereinafter referred to as gas flow rate) and the vertical axis represents the liquid flux (apparent flow rate of liquid: hereinafter referred to as liquid flow rate), A map-like feature vector diagram in which the intensity of the feature vector is displayed at the same height is obtained.

Standard deviation σ, distortion coefficient γ ₁ , sharpness coefficient γ ₂
And feature vector diagrams such as power spectral density S
The logical sum on the horizontal axis and vertical axis of each feature vector diagram represents the class in the feature space, based on the pattern of the fluctuation waveform determined by the gas-liquid mixing ratio and the flow rate of the gas-liquid mixed fluid, The gas flow rate and liquid flow rate that are the reference for the class are unknown.

The sample pattern 9 is calibration data obtained by experiments in advance to determine a class of a feature space as a logical sum of the feature vectors obtained by the statistical feature measuring unit 6. That is, the class of the feature space is determined using the gas / liquid mixture ratio and the flow rate as parameters.

The learning unit 8 calculates a parameter used for classifying the class of the feature space obtained by the statistical feature measuring unit 6 from a set of calibration data whose class is determined in advance by the sample pattern 9.

The pattern recognition section 7 identifies the class of the feature space of the statistical feature measurement section 6 based on the pattern recognition based on the class of the feature space classified by the learning section 8.

FIG. 2 is a diagram for explaining a statistical feature grid cell according to the present invention. Assuming that a point identified by the pattern recognition unit 7 is a position M on the statistical feature grid cell shown in FIG. The gas flow rate and the liquid flow rate for the position M can be measured by image processing using a computer. Moreover, the position M is based on the analysis of the mass flow rate Qm, the density ρ, and two statistical features, and the parameter is doubled as compared with the case where only the conventional pressure signal is used, so that the accuracy and reliability are improved. High gas flow and liquid flow can be obtained.

FIG. 3 is a block diagram for explaining an embodiment of the second aspect of the present invention. In FIG.
Reference numeral 2 denotes a differential pressure converter, and portions having the same operation as in FIG. 1 are denoted by the same reference numerals as in FIG. In FIG. 3, the pattern recognition unit 7, the learning unit 8, and the sample pattern 9 constituting the multi-phase flow computing unit 5 are not shown in order to use the same one as the multi-phase flow computing unit 5 shown in FIG.

According to the second aspect of the present invention, the statistical characteristic is obtained by adding not only the two parameters of the mass flow rate Qm and the density ρ but also the fluctuating differential pressure ΔP of the Coriolis flowmeter 1.
That is, the differential pressure transmitter 11 is attached to both ends of the measuring pipe 1a via the conduits 11a and 11b,
The differential pressure ΔP of the Coriolis flow meter 1 is transmitted to the statistical characteristic measuring unit 6 via the, and the statistical characteristic of the differential pressure ΔP is obtained. A feature, a frequency vector feature, and a feature vector diagram of a differential pressure ΔP waveform are obtained. Therefore, in the sample pattern 9, in addition to the calibration data of the mass flow rate Qm and the density ρ, the differential pressure Δ
The calibration data of the P waveform is also included, and the gas flow rate and the liquid flow rate are identified in the pattern recognition unit 7 in terms of the mass flow rate Qm and the density ρ and the statistical characteristic grid cell of the differential pressure ΔP. Therefore, the points of the statistical feature grid cell are identified by further adding the feature vector diagram of the fluctuation waveform of the differential pressure ΔP, so that the accuracy is higher than in the case of claim 1.
Highly reliable gas and liquid flow rates are required.

According to a third aspect of the present invention, the mass flow rate Qm, the density ρ and the differential pressure ΔP are mentioned as statistical features in the second aspect of the invention. In addition to this, the driving power W of the measuring tube 1a is further reduced. In addition.

As described above, when a gas-liquid multiphase flow having a high void ratio is passed through the Coriolis flowmeter 1 of a type that drives the measuring tube 1a at a natural frequency, the drive gain becomes unstable and measurement becomes impossible. Conversely, an unstable element of the drive gain is significant in a statistical method. That is, by determining the statistical characteristics of the driving power W of the measuring tube 1a, the points of the gas flow rate and the liquid flow rate on the statistical grit cell can be identified, and the points of the mass flow rate Qm, the density ρ and the differential pressure ΔP can be identified. By adding to the statistical features, more accurate and highly reliable gas flow rates and liquid flow rates can be obtained. Also in this case, the driving power W
Include.

FIG. 4 is a block diagram for explaining an example of an oil / moisture meter according to the fourth aspect of the present invention. In the figure, 13 is an oil / moisture meter, 14 is a microwave transmitter, and 15 is a microwave receiver. , 16 is a differential amplifier, and 17 is reference energy.

According to a fourth aspect of the present invention, when the liquid phase in the gas-liquid multiphase flow is a multiphase flow of oil and water, the oil flow rate and the water flow rate are determined from the liquid flow rates obtained in the first to third aspects. Is what you want.

The oil-moisture meter 13 shown in FIG. 4 is a well-known one. For example, a microwave transmitter 14 and a microwave receiver 15 are connected to a gas-liquid multiphase flow flowing through the outlet pipe 3 shown in FIG. They are installed at predetermined intervals. The microwave transmitted from the microwave transmitter 14 and received by the microwave receiver 15 is absorbed by water in the gas-liquid multiphase flow, attenuated, and input to one of the differential amplifiers 16,
This is compared with the reference energy 17. The differential amplifier 16 increases the output of the microwave transmitter 14 so that the comparison signal becomes zero, and detects the oil concentration and the water concentration from the output of the differential amplifier 16 at this time.

The flow rates of oil and water in the multiphase flow of air, oil and water can be obtained by multiplying the flow rate of the liquid phase in the gas-liquid multiphase flow by the ratio of the oil concentration to the water concentration. Note that the oil / moisture meter 13 is an example, and it is possible to obtain the oil content and the water concentration from the capacitance type for obtaining the capacitance from the dielectric constant difference of the multiphase flow, or the conductivity type obtained from the conductivity of the multiphase flow. it can.

FIG. 5 is a block diagram for explaining an embodiment of the fifth aspect of the present invention. In FIG. 5, reference numeral 18 denotes a homogenizer, and portions having the same functions as those in FIG. The same reference numbers are given.

In the figure, a homogenizer 18 is a mixing device for converting a gas-liquid mixed-phase flow flowing into the Coriolis flowmeter 1 into a substantially constant gas-liquid mixed-phase flow. The gas phase in the gas-liquid mixed-phase flow flows into the Coriolis flowmeter 1 as fine bubbles by reversely rotating the gas-liquid mixed-phase flow, so that the gas-liquid mixed-phase flow always has a stable mixing ratio.

The gas-liquid multiphase flow is classified into classes according to the magnitudes of the gas flux and the liquid flux as shown in the flow morphology map of Taiter and Dukier. Changes, bubble flow, hierarchical flow, wavy flow,
A shell flow, an annular flow, etc. are formed. Therefore, since the flow classified in this way flows directly into the Coriolis flowmeter, there is a possibility that the driving of the measurement tube 1a becomes unstable. Therefore, by providing the homogenizer 18 upstream of the Coriolis flowmeter 1, Enables the measurement of gas-liquid multiphase flows with a higher void fraction.

[0036]

【The invention's effect】

According to the first aspect of the present invention, when a gas-liquid multiphase flow is caused to flow through a Coriolis flowmeter, it becomes impossible to measure when the void fraction is increased, and the statistical characteristics of the fluctuation waveform of the flow rate Qm and the density ρ are used. Since a point on the statistical feature grid cell is identified by comparing the sample pattern with a sample pattern as the strength of the vector, a point is conventionally identified only from the strength of the feature vector of the differential pressure signal. Since the identification is performed according to twice the statistical information, the measurement accuracy of the gas flow rate and the liquid flow rate is improved, and the reliability is improved.

Effect corresponding to the second aspect: The statistical feature of the fluctuation waveform of the differential pressure ΔP of the Coriolis flowmeter is added to the first aspect of the present invention, so that the first aspect of the present invention is further measured. Accuracy and reliability are improved.

Effect corresponding to the third aspect: Since the statistical feature of the waveform of the driving power W for driving the Coriolis flowmeter is added to the second aspect, the measurement accuracy and the second aspect are further improved. Reliability is improved.

Advantages Corresponding to Claim 4: Claims 1 to 3
When the liquid determined by the invention of the above is a mixed phase flow of oil and water, the ratio of oil concentration to water concentration can be measured using an oil / moisture meter, so that the oil flow rate and the water flow rate can be separately measured. .

Advantages Corresponding to Claim 5: Claims 1 to 4
In the invention described in the section, the mixing means of the multiphase flow is provided upstream of the Coriolis flowmeter, so that the gas-liquid multiphase flow having a high void ratio can be measured.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating an embodiment of the present invention.

FIG. 2 is a diagram for explaining a statistical feature grid cell according to the present invention.

FIG. 3 is a block diagram for explaining an embodiment of the invention of claim 2;

FIG. 4 is a block diagram for explaining an example of an oil-moisture meter according to the invention of claim 4;

FIG. 5 is a block diagram for explaining an embodiment of the invention of claim 5;

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Coriolis flowmeter, 2 ... Inflow pipe, 3 ... Outflow pipe, 4 ... Converter of Coriolis flowmeter 1, 5 ... Multiphase flow calculator, 6 ... Statistical feature measurement unit, 7 ... Pattern recognition unit, 8 ... Learning Part, 9 ...
Sample pattern, 10 ... output, 11 ... differential pressure transmitter, 1
2 ... Differential pressure converter, 13 ... Oil moisture meter, 14 ... Microwave transmitter, 15 ... Microwave receiver, 16 ... Differential amplifier, 17 ... Reference energy, 18 ... Homogenizer.

Claims (5)

[Claims] 1. A Coriolis flowmeter for driving a measurement tube supported at two points at a natural frequency, a gas-liquid multiphase flow flowing through the Coriolis flowmeter, and a flow rate corresponding to a Coriolis force acting on the measurement tube. A converter that detects a phase difference signal and a vibration waveform having a frequency corresponding to the density of the gas-liquid multiphase flow, and outputs a temperature-corrected flow signal and a density signal of the gas-liquid multiphase flow; and The statistical feature of the signal is obtained as the strength of the feature vector, and the gas-phase flux and the liquid-phase flux in a predetermined gas-liquid multiphase flow are compared with sample patterns, and the position of the sample pattern on the statistical grid cell is determined. And a multi-phase flow calculator for determining the gas flux and the liquid flux of the gas-liquid multi-phase flow by a pattern recognition method using a Coriolis flow meter. 2. A differential pressure gauge for detecting a differential pressure is provided at both ends of the measuring pipe, and the differential pressure is detected by the differential pressure gauge in addition to the strength of the characteristic vector of the flow rate signal and the density signal obtained by the multiphase flow calculator. 2. The multi-phase flow meter according to claim 1, wherein a strength of a feature vector of the differential pressure signal is obtained, and the sample pattern includes the differential pressure signal. 3. In addition to the flow rate signal, the density signal, and the differential pressure signal obtained by the multiphase flow calculator, the strength of a characteristic vector of drive power for driving the measurement tube of the Coriolis flowmeter is obtained, and the sample is obtained. 3. The multi-phase flow meter according to claim 2, wherein the driving power is included in a pattern by a pattern recognition method using a Coriolis flow meter. 4. In the gas-liquid multi-phase flow, when the liquid phase is a three-phase fluid of water and oil, an oil-moisture meter is provided downstream of the Coriolis flow meter, and each of the liquid-phase water and oil is provided. 4. A multi-phase flow meter based on a pattern recognition method using a Coriolis flow meter according to claim 1, wherein a flow rate is obtained. 5. A mixing means for uniformly mixing a gas phase and a liquid phase of the gas-liquid mixed phase flow upstream of the Coriolis flow meter. Multi-phase flow meter by pattern recognition method using Coriolis flow meter.