RU2527667C2 - Method for measurement of flow rate of two-phase three-component medium - Google Patents

Abstract

FIELD: oil and gas industry.

SUBSTANCE: invention relates to oil and gas industry and namely to the methods of oil well debit determination without preliminary gas separation from the well production. The method for measurement of flow rate of two-phase three-component medium involves calibration of a multiphase flow meter, processing of results of calibration works, synthesis of mathematical model of two-phase three-component medium movement, determination of interval of flow rate of liquid and oil gas at which there is an allowable error for calculation of flow rate of oil, water and oil gas. At that in process of calibration works and synthesis of mathematical model of two-phase three-component medium movement dependency of calibration point error is determined on the average weight of points in the teaching model an in process of the well operation sensor readings are recorded for the multiphase flow meter and calculation of the product flow rate is made for components of the oil well at the average weight of points in the teaching model when at calibration points a minimum value of root-mean-square deviation is observed between the calculated and measured values of fluid rate.

EFFECT: reducing error in measurement of the flow rate by components for the oil well product.

1 dwg, 2 tbl

Description

The invention relates to the oil industry, and in particular, to methods for determining the production rate of oil wells without preliminary separation of gas from well products.

A known method for determining the flow rate of oil wells, including sensing the flow of acoustic pulses, recording the pulses transmitted through the medium by the receiver in a limited controlled volume of the stream formed by a pair of "radiation source-receiver", fix the time it takes for the pulses to pass through the controlled volume and calculate the flow rate of the components based on the laws of two-phase movement ternary environment [1]. However, this method leads to significant errors in determining the flow rate of a two-phase medium due to the neglect of the effect of oil gas dissolved in oil and water at pressures and temperatures in the measured flow.

Closest to the proposed solution is a method of measuring the flow rate of a two-phase three-component medium, including calibrating a multiphase flow meter, processing the results of calibration work, synthesizing a mathematical model of the motion of a two-phase three-component medium, determining the interval of flow rates of liquid and oil gas, in which there is an allowable error in calculating the flow rates of oil and water and gas [2].

However, this method leads to errors in determining the component-wise flow rate of oil well production due to improper selection of the average weight of training points.

The objective of the proposed technical solution is to develop such a method for measuring the flow rate of a two-phase three-component medium, in the implementation of which errors could be eliminated due to an incorrect choice of the average weight of the training points.

The problem is solved in that in the process of calibrating a multiphase flow meter and synthesizing a mathematical model of the motion of a two-phase three-component medium, the dependence of the error of the test points on the average weight of the points of the training model is determined, and during the operation of the well, the sensors of the multiphase flow meter are read and component-wise flow rate is determined with the average weight of the training points at which there is a minimum error of the test points.

The technical result of the invention is to improve the accuracy of measuring the flow rate of a two-phase three-component medium.

The technical result is achieved by the fact that in the method of measuring the flow rate of a two-phase three-component medium, including calibrating a multiphase flow meter, processing the results of calibration work, synthesizing a mathematical model of the movement of a two-phase three-component medium, determining the interval of flow rates of liquid and oil gas, in which there is an allowable error in calculating the flow rate of oil, water and oil gas, in the process of calibrating a multiphase flow meter and synthesizing a mathematical model of the motion of a two-phase three-component of the medium, the dependence of the error of the test points on the average weight of the points of the training model is determined, and during the operation of the well, the readings of the sensors of the multiphase flow meter are taken and the component flow rate is calculated at the average weight of the training points, at which the minimum error of the test points takes place.

The method is implemented as follows. A multiphase flow meter is being calibrated. In the table of results of the calibration work, the flow rate of the liquid corresponds to the readings of gas saturation sensors, Doppler frequency shift, oil humidity, pressure and temperature.

Experimental work is carried out to determine the dependence of the error of the test points on the average weight of the points of the training model. The weights of each experimental point (stats.w) can be obtained using the robustfit function in the MatLab environment. This function implements an iterated weighted least squares method. The weights of each experimental point at the current iteration are calculated using a biquadratic function of the residue vector calculated at the previous iteration. Using this algorithm allows you to set a lower weight for observations that have a greater deviation from the regression model relative to the rest. The average weight of the points of the training model changes in experimental work due to changes in the number of experimental points with one or another weight. In this case, the approximation of the experimental points is based on the regression dependence.

As a result of the experimental work, the dependence of the error of the test points on the average weight of the experimental points of the training model is found. The average weight of the training points is found at which there is a minimum of the average absolute error of the test points. During the operation of the well, the readings of the sensors of the multiphase flowmeter are taken, and the calculation of the flow rate of the two-phase three-component medium is carried out with the average weight of the training points at which the smallest error is observed.

An example of a specific implementation of the method is illustrated by the calibration work materials of the UltraFlow device. Moreover, the fluid flow rate was adopted as a response.

The following regression model is used:

where y is the fluid flow, m ³ / day;

x ₁ - readings of the gas saturation sensor;

x ₂ - Doppler frequency shift, Hz;

x ₃ - readings of the water cut sensor;

x ₄ - pressure in the measurement interval, MPa;

x ₅ - flow temperature, ° C.

The data on the training sample are given in table 1. Based on these data, a regression dependence was synthesized (1).

Table 1 Fluid flow rate, m ³ / day Gas saturation sensor readings Doppler frequency shift, Hz Water saturation sensor readings Pressure, MPa Temperature ° C 10.37 0.4127 4197.4 33020,3 0.1521 24 15.25 0.7052 4280.2 28121.1 0.1489 24.6 15,20 0.5992 4291.3 27702.1 0.1499 24.2 10.6 0.4493 4314.9 26523.1 0,1507 25.7 15.33 0.6734 4417.7 27658.7 0.1491 25 10.01 0.4928 4525.4 25362.4 0,1502 25 15.25 0.6247 4634,1 27641.3 0.1499 25,2 15.24 0.7759 4893 28190.4 0.1486 24.6 15.18 0.7961 5219.9 28354.8 0.1485 24.3 10.12 0.7049 5497.3 25720.4 0.1483 25.1 10.1 0.69 5708.1 32,745.9 0.1494 24.9 10.11 0.7598 5735.2 26197.5 0.1481 25,2 10.33 0.7611 5811.4 26,261.7 0.1488 25,2 10.37 0.7620 5862.0 26196 0.1485 25.1 10,2 0.7304 5881.4 32659.6 0.1489 24.7 10.26 0.7670 5961.8 32432 0.1487 24.7 15.15 0.4976 4076.3 27759.1 0,1512 25,4

Experimental points that did not participate in the synthesis of model (1) are called test points. The error of the test points is used to select the type of regression dependence. During normal well operation, the test points are points whose input variables determine the dependent variable (in this case, fluid flow). Data on the test points are given in table 2.

table 2 Fluid flow rate, m ³ / day Gas saturation sensor readings Doppler frequency shift, Hz Water saturation sensor readings Pressure, MPa Temperature ° C 10.85 0,3019 3054.6 26491 0.1524 25.6 15.11 0.3485 3120 28150.4 0.1528 25.3 10.1 0.3427 3559.9 25847.6 0.1515 25,2 10.16 0.3528 3621.7 25505.5 0,1506 26.5 15.34 0.6437 4083.8 27742.8 0.1492 24.5 15.26 0.5909 4,194.3 27733.9 0.1488 24.9 10.22 0.221 2650 32415.8 0.155 24.7

Using the data in Tables 1 and 2, the dependence of the average absolute error of the test points on the average weight of the training points is constructed. At the same time, in the Matlab environment, the weight of each experimental point of the training sample (stats.w) can be obtained using the robustfit function. The average weight of points in the training sample mean (stats.w) can be changed by adding or removing training points. So, for example, with the number of training points 18 (Table 1), the average weight of the training points mean (stats.w) = 0.9775. In this case, the average absolute error of the test points (Table 2) mae (eryo) = 0.0132. The 19th point is added to the training set.

B1 = [15.01 0.4986 4012.5 27650 0.145 27].

Then mean (stats.w) = 0.9609, mae (epo) = 0.0111, etc.

The input variables of the model are: readings of gas saturation sensors, Doppler frequency shift, oil moisture, pressure in the measurement interval, flow temperature. When processing the experimental data, the dependence of the average absolute error of the test points on the average weight of the points of the training sample was obtained (see Fig.). As can be seen from the above figure, there is a minimum of error in the calculated values of the fluid flow rate with an average weight of points of the training model 0.926-0.928. In this case, there is a significant decrease in the error compared to the error outside this interval of average weight.

An exploded flow rate of oil well production is determined as follows.

The moisture content of oil is determined by the regression dependence [3]:

where z ₁ - flow rate, m ³ / day;

z ₂ - readings of the oil humidity sensor;

z ₃ - Doppler frequency shift, Hz;

z ₄ - readings of the gas saturation sensor;

z ₅ is the ratio of flow temperature to pressure, ° C / MPa.

Water consumption is determined by the dependence of Q _in = Y · Wl.

In this case, oil consumption is determined by the formula

Q _n = YQ _c .

Gas saturation of the flow is determined by the regression dependence [4]:

where W ₁ - flow rate, m ³ / day;

W ₂ - readings of gas saturation sensors; oil moisture;

W ₃ - Doppler frequency shift, Hz;

W ₄ - readings of oil humidity sensors;

W ₅ is the ratio of flow temperature to pressure, ° C / MPa.

Associated gas consumption is determined by the formula

The application of the proposed technical solution will significantly reduce the error in calculating the flow rate of oil well products when using a multiphase flow meter.

Information sources

1. RF patent No. 2138023 "Method for determining the flow rate of components of a multiphase medium." // Melnikov V.I., Drobkov V.P. - 1999.09.20.

2. Pismarov M.N. Calculation of the flow of a three-component medium during the calibration of a multiphase flow meter. Innovations and actual problems of engineering and technology: All-Russian scientific-practical conference of young scientists in 2 volumes / M.N. Pismarov, K.Yu. Plessovsky; under the editorship of A.A. Bolshakova. - Saratov: Saratov State Technical University, 2009. - T.1. - 360 p. - S. 110-112.

3. Goryunov A.N. Determination of oil moisture according to the readings of the sensors of the UltraFlow device / A.N. Goryunov, T.V. Kalinina, O.B. Kachalov. // Innovative educational technologies and methods for their implementation: Proceedings of the IX All-Russian Scientific and Practical Conference (Arzamas, January 27, 2012). - M.: Publishing House of SSU, 2012. - S.306-308.

4. Baranova A.V. The choice of a mathematical model for calculating the gas saturation of a stream using the Ultraflow device / A.V. Baranova, O.B. Kachalov, A.S. Spiridonova // Innovative educational technologies and methods for their implementation: Proceedings of the IX All-Russian Scientific and Practical Conference (Arzamas, January 27, 2012). - M.: Publishing House of SSU, 2012 .-- S.301-303.

Claims (1)

A method of measuring the flow rate of a two-phase three-component medium, including calibrating a multiphase flow meter, processing the results of calibration work, synthesizing a mathematical model of the motion of a two-phase three-component medium, determining the interval of flow rates of liquid and oil gas, in which there is an allowable error in calculating the flow rates of oil, water and oil gas, characterized in that in the process of calibrating a multiphase flow meter and synthesizing a mathematical model of the motion of a two-phase three-component medium, the error of the test points versus the average weight of the points of the training model, and during the operation of the well, the readings of the sensors of the multiphase flow meter are taken and the component flow rate is determined with the average weight of the training points at which the minimum error of the test points takes place.