CN112096374A - A dynamic measurement error compensation method for oil holdup measurement by split flow method - Google Patents

Abstract

The invention belongs to the technical field of petroleum engineering and well logging, and in particular relates to a dynamic measurement error compensation method for measuring oil _holdup by a split flow method. , the measured actual oil holdup Vf _t , calculate the dynamic measurement error of oil holdup ΔV _t ; 2. Use grey system theory to identify and correct the abnormal value of the dynamic measurement error ΔV _t of oil holdup; 3. The dynamic measurement error of oil holdup ΔVm _t is calculated by numerical addition and anti-fold preprocessing, and the dynamic measurement error of oil holdup after processing is ΔVP _t ; 4. The dynamic measurement error of oil holdup after numerical addition and anti-fold preprocessing is ΔVP _t , establish the oil holdup dynamic measurement error prediction model; 5. Use the oil holdup dynamic measurement error prediction model to compensate the oil holdup dynamic measurement error caused by the diversion. The method compensates the dynamic measurement error of oil holdup measured by the split flow method, improves the accuracy of oil holdup measurement by the split flow method, and provides reliable oil holdup data for oilfield measurement.

Description

A dynamic measurement error compensation method for oil holdup measurement by split flow method

Technical field:

The invention belongs to the technical field of petroleum engineering and well logging, and particularly relates to a dynamic measurement error compensation method for measuring oil holdup by a shunt method.

Background technique:

With the use of water injection oil recovery technology and the extension of production time, most of my country's onshore oilfields have entered the middle and late stages of development, with extremely high water retention rates, some even reaching 95%, and 70% of newly discovered reserves are high water cut reserves. In the field of dynamic logging in oilfields, monitoring the oil holdup of each producing layer and providing information on reservoir production can optimize the reservoir engineering plan, evaluate the effects of fracturing, water plugging and other engineering operations; Measurement is of great significance for evaluating the development effect of a single well, optimizing the production system of oil wells, and ensuring the safe operation of oil production equipment.

However, in high water-cut oil wells, it is difficult to obtain accurate oil holdup by directly measuring oil holdup with traditional conductivity sensors due to the low oil holdup. At present, the shunt method is mostly used. The so-called shunt method refers to the separation of part of the water according to gravity, which increases the concentration of the oil, and then uses the conductivity probe sensor to measure the oil holdup. Theoretically, if the amount of water to be diverted can be accurately measured, the oil holdup can be accurately measured by the separation method, but the measurement of oil holdup is a dynamic measurement, and there is emulsification of oil and water during the measurement process, so it is unavoidable that a small amount of oil flows with the water. Therefore, dynamic measurement errors are generated, and the generation of dynamic measurement errors leads to inaccurate measurement of oil holdup. In order to realize the accurate measurement of oil holdup, it is necessary to perform error compensation on the oil holdup obtained by dynamic measurement.

Invention content:

The purpose of the present invention is to solve the problem of inaccurate oil holdup measurement caused by partial oil flowing off with water during the dynamic measurement of oil holdup by the existing split flow method, and to provide a dynamic measurement of oil holdup by the split flow method Error compensation method.

The technical scheme adopted in the present invention is: a dynamic measurement error compensation method for measuring oil holdup by a split flow method, and the compensation method comprises the following steps:

Step 1: Carry out the oil-water measurement experiment of the split flow method, set the theoretical oil holdup V _t of the t-th experiment, the actual oil holdup obtained from the measurement Vf _t , calculate the dynamic measurement error of oil holdup ΔV _t , t=1, 2, 3, ..., n, n is the total number of experiments performed;

The calculation method of oil holdup dynamic measurement error ΔV _i is:

ΔV _t =V _t -Vf _t

Step 2: Use the grey system theory to identify and correct the abnormal value of the dynamic measurement error ΔV _t of oil holdup, and the corrected dynamic measurement error of oil holdup is ΔVm _t ;

The grey system theory is used to identify and correct the abnormal value of the oil holdup dynamic measurement error ΔV _t as follows:

则ΔV_t为动态测量异常值，并且此处的值采用

修正；The dynamic measurement error of oil holdup ΔV _t satisfies

Then ΔV _t is the dynamic measurement outlier, and the value here is

amend;

中，γ为灰辨别系数，

为ΔV_t的灰预测值，在精密测量中，γ一般取2.3，但是在分流法测量持油率动态测量误差补偿异常值识别中，γ取值2.3不满足需求，本方法从跳跃度定义上分析将灰辨别系数将γ定义为：

, γ is the gray discrimination coefficient,

is the gray predicted value of ΔV _t . In precision measurement, γ generally takes 2.3, but in the identification of abnormal value of dynamic measurement error compensation of oil holdup measured by shunt method, γ value of 2.3 does not meet the requirements. This method is based on the definition of jump degree. The analysis defines the grey discrimination coefficient γ as:

The oil holdup dynamic measurement error ΔV _t outlier identification model can be expressed as:

的表达式如下：According to the GM(1,1) model,

The expression is as follows:

where α is the development coefficient, μ is the amount of ash action, and the methods for obtaining α and μ are:

z⁽¹⁾(t)为GM(1，1)背景序列，z⁽¹⁾(t)＝0.5ΔV_t+0.5ΔV_t-1，t＝2，3，...，n；原始矩阵Y的表达式为：Among them, B is the grayscale matrix of the GM(1,1) model, and the expression is:

z ⁽¹⁾ (t) is the GM(1,1) background sequence, z ⁽¹⁾ (t)=0.5ΔV _t +0.5ΔV _t-1 , t=2,3,...,n; original matrix Y The expression is:

Step 3: Perform arithmetic addition and anti-fold preprocessing on the corrected dynamic measurement error of oil holdup ΔVm _t , and the dynamic measurement error of oil holdup after numerical addition and anti-fold pre-processing is ΔVP _t ;

The methods of adding and de-folding preprocessing are as follows:

Because the gray system theory requires the data to be roughly concave and the modeling data cannot be negative, and the corrected oil holdup dynamic measurement error ΔVm _t has a negative value and is uneven, in order to change the corrected oil holdup dynamic measurement error ΔVm The concave-convexity and positive and negative properties of _t need to be pre-processed by anti-folding and numbering;

If the corrected dynamic measurement error of oil holdup ΔVm _t is concave and does not contain negative values, then it is not necessary to do the pre-processing of the reverse fold and number, that is: ΔVP _t =ΔVm _t ;

If the corrected dynamic measurement error ΔVm _t of oil holdup is concave and contains negative values, no preprocessing is required at this time, namely:

ΔVP _t =ΔVm _t +{|ΔVm ₁ |, |ΔVm ₂ |,…,|ΔVm _t |,…,|ΔVm _n |}max;

If the corrected oil holdup dynamic measurement error ΔVm _t exhibits convexity and does not contain negative values, then it is necessary to perform anti-fold preprocessing, namely: ΔVP _t =ΔVm _t ;

If the corrected dynamic measurement error of oil holdup ΔVm _t is convex and contains negative values, it is necessary to do the pre-processing of the reverse fold and number, namely:

ΔVP _t =−ΔVm _t +{|ΔVm ₁ |, |ΔVm ₂ |,…,|ΔVm _t |,…,|ΔVm _n |}max;

In the formula, {|ΔVm ₁ |, |ΔVm ₂ |, …, |ΔVm _t |, …, |ΔVm _n |}max are the data operation coefficients, meaning |ΔVm ₁ |, |ΔVm ₂ |, …, |ΔVm The largest value among _t |, …, |ΔVm _n |;

Step 4: Establish a prediction model of oil holdup dynamic measurement error according to the numerical addition and the oil holdup dynamic measurement error ΔVP _t after anti-fold preprocessing;

The method of establishing the oil holdup dynamic measurement error prediction model is as follows:

When the oil holdup is less than 5%, the GM(1,1) model predicts the dynamic measurement error of oil holdup more accurately, but when the oil holdup is in the interval [5%, 10%], the GM(1,1) model predicts The dynamic measurement error accuracy of oil holdup is low; while the NDGM(1, 1) model predicts that when the oil holdup is in the interval [5%, 10%], the prediction accuracy is higher, and when the oil holdup is less than 5%, the prediction accuracy is higher. Therefore, neither the GM(1,1) model nor the NDGM(1,1) model can be directly applied to the prediction of oil holdup dynamic measurement error;

This method establishes the oil holdup dynamic measurement error prediction model including the modeling weight λ as follows:

为数加与反褶预处理后的持油率动态测量误差ΔVP_t的GM(1，1)模型预测值，其解法与

的解法相同；In the formula,

is the predicted value of the GM(1, 1) model of the dynamic measurement error ΔVP _t of oil holdup after numerical addition and anti-fold preprocessing, and its solution is the same as

The solution is the same;

为数加与反褶预处理后的持油率动态测量误差ΔVP_t的NDGM(1，1)模型预测值，其表达式如下：

It is the predicted value of the NDGM(1, 1) model of the dynamic measurement error ΔVP _t of oil holdup after numerical addition and anti-fold preprocessing, and its expression is as follows:

In the formula, β ₁ , β ₂ and β ₃ are the multiple coefficients, linear coefficients and intercepts of the NDGM(1, 1) model, respectively. The solution methods are as follows:

(β ₁ , β ₂ , β ₃ ) ^T = (A ^T A) ^-1 A ^T M

In the formula, A is the NDGM(1,1) gray matrix, M is the NDGM(1,1) background matrix, and its expression is as follows:

和持油率动态测量误差ΔVP_t的值误差最小平方和获得，其表达式如下：β ₄ is the initial predicted value correction factor, which can be solved by solving the predicted value

The value of the oil holdup dynamic measurement error ΔVP _t is obtained by the least square sum of the value error, and its expression is as follows:

In the formula, j is the temporary statistical coefficient;

The modeling weight λ unconstrained optimization solution model is as follows:

Obtain the value of the modeling weight λ by derivation;

Step 5: Use the oil holdup dynamic measurement error prediction model to compensate for the oil holdup dynamic measurement error caused by the diversion;

The method of using the oil holdup dynamic measurement error prediction model to compensate the holdup error caused by the diversion is as follows:

where Vs _t is the oil holdup after compensation.

The beneficial effects of the present invention are as follows: a dynamic measurement error compensation method for oil holdup measurement by the split flow method is provided, which solves the problem of inaccurate oil holdup measurement caused by partial oil flowing off with water during the dynamic measurement of oil holdup by the existing split flow method. question. Its main advantages are as follows:

(1) This method uses the grey discrimination method to identify and correct the dynamic measurement error of the oil rate, avoids the abnormal measurement value caused by the adhesion of oil bubbles to the conductive surface, and improves the compensation accuracy;

(2) A prediction model of oil holdup dynamic measurement error including modeling weight λ is established, which overcomes the situation of GM(1,1) model at high oil holdup and NDGM(1,1) model at low oil holdup The problem of low compensation accuracy is solved, and the compensation accuracy is improved.

Description of drawings:

Fig. 1 is the schematic diagram of split flow method in the embodiment one;

Fig. 2 is the error curve of measuring oil holdup by adopting the GM(1,1) model compensation shunt method in Example 1 (theoretical oil holdup is 3%);

Fig. 3 is the error curve of measuring oil holdup by adopting the NDGM (1,1) model compensation shunting method in Example 1 (theoretical oil holdup is 3%);

Fig. 4 is the error curve of the measurement of oil holdup by adopting the GM(1,1) model compensation shunting method in the first embodiment (theoretical oil holdup is 9%);

Fig. 5 is the error curve of the measurement of oil holdup by adopting the NDGM (1,1) model compensation shunting method in Example 1 (theoretical oil holdup is 9%);

FIG. 6 is a graph showing the error curve (theoretical oil holdup 3%) of measuring holdup by using the method compensation shunting method in Example 1;

FIG. 7 is a graph showing the error curve of the oil holdup (theoretical oil holdup 9%) measured by the method compensation shunting method in Example 1. FIG.

Detailed ways:

Example 1

A dynamic measurement error compensation method for measuring oil holdup by a shunt method, the compensation method comprises the following steps:

Step 1: Carry out the oil-water measurement experiment of the split flow method, set the theoretical oil holdup V _t of the t-th experiment, the actual oil holdup obtained from the measurement Vf _t , calculate the dynamic measurement error of oil holdup ΔV _t , t=1, 2, 3, ..., n, n is the total number of experiments performed;

The calculation method of oil holdup dynamic measurement error ΔV _i is:

ΔV _t =V _t -Vf _t

Step 2: Use the grey system theory to identify and correct the abnormal value of the dynamic measurement error ΔV _t of oil holdup, and the corrected dynamic measurement error of oil holdup is ΔVm _t ;

The grey system theory is used to identify and correct the abnormal value of the oil holdup dynamic measurement error ΔV _t as follows:

则ΔV_t为动态测量异常值，并且此处的值采用

修正；The dynamic measurement error of oil holdup ΔV _t satisfies

Then ΔV _t is the dynamic measurement outlier, and the value here is

amend;

中，γ为灰辨别系数，

为ΔV_t的灰预测值，在精密测量中，γ一般取2.3，但是在分流法测量持油率动态测量误差补偿异常值识别中，γ取值2.3不满足需求，本方法从跳跃度定义上分析将灰辨别系数将γ定义为：

, γ is the gray discrimination coefficient,

is the gray predicted value of ΔV _t . In precision measurement, γ generally takes 2.3, but in the identification of abnormal value of dynamic measurement error compensation of oil holdup measured by shunt method, γ value of 2.3 does not meet the requirements. This method is based on the definition of jump degree. The analysis defines the grey discrimination coefficient γ as:

The oil holdup dynamic measurement error ΔV _t outlier identification model can be expressed as:

的表达式如下：According to the GM(1,1) model,

The expression is as follows:

where α is the development coefficient, μ is the amount of ash action, and the methods for obtaining α and μ are:

z⁽¹⁾(t)为GM(1，1)背景序列，z⁽¹⁾(t)＝0.5ΔV_t+0.5ΔV_t-1，t＝2，3，...，n；原始矩阵Y的表达式为：Among them, B is the grayscale matrix of the GM(1,1) model, and the expression is:

z ⁽¹⁾ (t) is the GM(1,1) background sequence, z ⁽¹⁾ (t)=0.5ΔV _t +0.5ΔV _t-1 , t=2,3,...,n; original matrix Y The expression is:

Step 3: Perform arithmetic addition and anti-fold preprocessing on the corrected dynamic measurement error of oil holdup ΔVm _t , and the dynamic measurement error of oil holdup after numerical addition and anti-fold pre-processing is ΔVP _t ;

The methods of adding and de-folding preprocessing are as follows:

Because the gray system theory requires the data to be roughly concave and the modeling data cannot be negative, and the corrected oil holdup dynamic measurement error ΔVm _t has a negative value and is uneven, in order to change the corrected oil holdup dynamic measurement error ΔVm The concave-convexity and positive and negative properties of _t need to be pre-processed by anti-folding and numbering;

If the corrected dynamic measurement error of oil holdup ΔVm _t is concave and does not contain negative values, then it is not necessary to do the pre-processing of the reverse fold and number, that is: ΔVP _t =ΔVm _t ;

If the corrected dynamic measurement error ΔVm _t of oil holdup is concave and contains negative values, no preprocessing is required at this time, namely:

ΔVP _t =ΔVm _t +{|ΔVm ₁ |, |ΔVm ₂ |,…,|ΔVm _t |,…,|ΔVm _n |}max;

If the corrected oil holdup dynamic measurement error ΔVm _t exhibits convexity and does not contain negative values, then it is necessary to perform anti-fold preprocessing, namely: ΔVP _t =ΔVm _t ;

If the corrected dynamic measurement error of oil holdup ΔVm _t is convex and contains negative values, it is necessary to do the pre-processing of the reverse fold and number, namely:

ΔVP _t =−ΔVm _t +{|ΔVm ₁ |, |ΔVm ₂ |,…,|ΔVm _t |,…,|ΔVm _n |}max;

In the formula, {|ΔVm ₁ |, |ΔVm ₂ |, …, |ΔVm _t |, …, |ΔVm _n |}max are the data operation coefficients, meaning |ΔVm ₁ |, |ΔVm ₂ |, …, |ΔVm The largest value among _t |, …, |ΔVm _n |;

Step 4: Establish a prediction model of oil holdup dynamic measurement error according to the numerical addition and the oil holdup dynamic measurement error ΔVP _t after anti-fold preprocessing;

The method of establishing the oil holdup dynamic measurement error prediction model is as follows:

When the oil holdup is less than 5%, the GM(1,1) model predicts the dynamic measurement error of oil holdup more accurately, but when the oil holdup is in the interval [5%, 10%], the GM(1,1) model predicts The dynamic measurement error accuracy of oil holdup is low; while the NDGM(1, 1) model predicts that when the oil holdup is in the interval [5%, 10%], the prediction accuracy is higher, and when the oil holdup is less than 5%, the prediction accuracy is higher. Therefore, neither the GM(1,1) model nor the NDGM(1,1) model can be directly applied to the prediction of oil holdup dynamic measurement error;

This method establishes the oil holdup dynamic measurement error prediction model including the modeling weight λ as follows:

为数加与反褶预处理后的持油率动态测量误差ΔVP_t的GM(1，1)模型预测值，其解法与

的解法相同；In the formula,

is the predicted value of the GM(1, 1) model of the dynamic measurement error ΔVP _t of oil holdup after numerical addition and anti-fold preprocessing, and its solution is the same as

The solution is the same;

为数加与反褶预处理后的持油率动态测量误差ΔVP_t的NDGM(1，1)模型预测值，其表达式如下：

It is the predicted value of the NDGM(1, 1) model of the dynamic measurement error ΔVP _t of oil holdup after numerical addition and anti-fold preprocessing, and its expression is as follows:

In the formula, β ₁ , β ₂ and β ₃ are the multiple coefficients, linear coefficients and intercepts of the NDGM(1, 1) model, respectively. The solution methods are as follows:

(β ₁ , β ₂ , β ₃ ) ^T = (A ^T A) ^-1 A ^T M

In the formula, A is the NDGM(1,1) gray matrix, M is the NDGM(1,1) background matrix, and its expression is as follows:

和持油率动态测量误差ΔVP_t的值误差最小平方和获得，其表达式如下：β ₄ is the initial predicted value correction factor, which can be solved by solving the predicted value

The value of the oil holdup dynamic measurement error ΔVP _t is obtained by the least square sum of the value error, and its expression is as follows:

In the formula, j is the temporary statistical coefficient;

The modeling weight λ unconstrained optimization solution model is as follows:

Obtain the value of the modeling weight λ by derivation;

Step 5: Use the oil holdup dynamic measurement error prediction model to compensate for the oil holdup dynamic measurement error caused by the diversion;

The method of using the oil holdup dynamic measurement error prediction model to compensate the holdup error caused by the diversion is as follows:

where Vs _t is the oil holdup after compensation.

As shown in Figures 1-7, Figure 1 is a schematic diagram of the split flow method, including a current collecting umbrella 1, a liquid split outlet 2, a flow sensor 3 and a fluid pipe wall 4.

The oil holdup measurement experiment was carried out on a simulated well of oil, gas and water three-phase flow, and the theoretical oil holdup V _t was 3% and 9% respectively in two experiments; after several experiments, when the theoretical oil holdup V _t was 3%, 15 The experimental results were 2.1%, 2.2%, 2.1%, 2.3%, 2.7%, 2.6%, 2.5%, 2.2%, 2.8%, 2.6%, 2.6%, 2.2%, 2.5%, 2.4%, 2.3%; When the theoretical oil holdup V _t is 9%, the results of 15 experiments are 8.6%, 8.7%, 8.5%, 8.3%, 8.1%, 8.2%, 8.7%, 8.5%, 8.4%, 8.3%, 8.1%, 8.2%, 8.4%, 8.6%, 8.4%. The GM(1,1) and NDGM(1,1) models are used to compensate the measurement results of the theoretical oil holdup V _t of 3%. The compensation results are shown in Figures 2 and 3. It can be seen from Figure 2 and Figure 3 that the maximum error between the measured value and the true value after GM(1,1) compensation is 0.21%, and the average error is 0.12%, while the maximum error between the measured value and the true value after NDGM(1,1) compensation is is 0.70%, and the average error is 0.47%. The GM(1,1) and NDGM(1,1) models are used to compensate the theoretical oil holdup V _t of 9%. The compensation results are shown in Figures 4 and 5. It can be seen from Figure 4 and Figure 5 that the maximum error between the measured value and the true value after GM(1,1) compensation is 0.6%, and the average error is 0.52%, while the maximum error between the measured value and the true value after NDGM(1,1) compensation is is 0.16%, and the average error is 0.13%.

This method is used to compensate the same data with theoretical oil holdup V _t of 3% and 9%, respectively, and the compensation results are shown in Figure 6 and Figure 7, respectively. It can be seen from Figure 6 that the maximum error between the measured value and the true value after compensation by this method is 0.11%, and the average error is 0.09%; from Figure 7, it can be seen that the maximum error between the measured value and the true value after compensation by this method is 0.12%, and the average error is 0.10% %. Therefore, the method improves the measurement accuracy and is superior to the existing model, so the patent provides reliable oil holdup data for oilfield measurement.

In order to solve the problem of inaccurate oil holdup measurement due to partial oil flowing out with water during the dynamic measurement of oil holdup by the existing split flow method, a dynamic measurement error compensation method for oil holdup measurement by the split flow method was proposed. The invention improves the ash recognition and correction model, and uses the ash recognition and correction model to identify and correct the abnormal value of the dynamic measurement error of oil holdup; After processing, the prediction accuracy is improved; a prediction model of oil holdup dynamic measurement error including the modeling weight λ is established, and the oil holdup measurement results are compensated. The patent compensates for the dynamic measurement error of the oil holdup measured by the split flow method, improves the accuracy of the oil holdup measurement by the split flow method, and provides reliable oil holdup data for oilfield measurement.

The above content is a further detailed description of the present invention in combination with specific embodiments, and it cannot be considered that the present invention is limited to the above-mentioned specific implementations. On the premise of not departing from the overall idea of the present invention and the protection of the claims, some simple deductions or substitutions can also be made, which should be regarded as belonging to the protection scope of the present invention.

Claims (1)

1. A dynamic measurement error compensation method for measuring oil holdup by a shunt method is characterized by comprising the following steps: the compensation method comprises the following steps:

the method comprises the following steps: carrying out oil-water measurement experiment by a flow splitting method, and setting the theoretical oil holdup of the t-th experiment as V_tAnd the measured actual oil holdup is Vf_tCalculating the dynamic measurement error delta V of the oil holdup_tWhere t is 1, 2, 3, …, n, n is the total number of experiments performed;

dynamic measurement error delta V of oil holdup_iThe calculation method comprises the following steps:

ΔV_t＝V_t-Vf_t

step two: dynamic measurement error delta V of oil holdup_tAbnormal values are identified and corrected by adopting a grey system theory, and the corrected dynamic measurement error of the oil holdup is delta Vm_t；

Dynamic measurement error delta V of oil holdup_tThe method for identifying and correcting the abnormal value by adopting the gray system theory comprises the following steps:

dynamic measurement error delta V of oil holdup_tSatisfy the requirement of

Then Δ V_tAbnormal values are measured for dynamics, andvalues herein adopt

Correcting;

in the formula, γ is a gray discrimination coefficient,

is DeltaV_tThe method analyzes from the jerk definition that the gray discrimination coefficient defines γ as:

dynamic measurement error delta V of oil holdup_tThe outlier identification model can be expressed as:

according to the GM (1, 1) model,

the expression of (a) is as follows:

in the formula, alpha is a development coefficient, mu is a gray acting amount, and the acquisition method of alpha and mu is as follows:

wherein, B is a GM (1, 1) model gray matrix, and the expression is:

z⁽¹⁾(t) is GM (1, 1) background sequence, z⁽¹⁾(t)＝0.5ΔV_t+0.5ΔV_t-1T is 2, 3,. n; the expression of the original matrix Y is:

step three: dynamically measuring error delta Vm of corrected oil holdup_tPerforming a number-addition and inverse-pleat preprocessing operation, wherein the dynamic measurement error of the oil holdup after the number-addition and inverse-pleat preprocessing is delta VP_t；

The method of the number addition and inverse pleat preprocessing operation is as follows:

if the corrected oil holdup dynamic measurement error delta Vm_tExhibits concavity and contains no negative values, in which case no anticracking and no pretreatment is required, i.e.: Δ VP_t＝ΔVm_t；

If the corrected oil holdup dynamic measurement error delta Vm_tExhibits concavity and contains negative values, in which case no number plus pretreatment is required, i.e.:

ΔVP_t＝ΔVm_t+{|ΔVm₁|，|ΔVm₂|，…，|ΔVm_t|，…，|ΔVm_n|}_max；

if the corrected oil holdup dynamic measurement error delta Vm_tExhibits convexity without negative values, and requires a reverse-folding pretreatment, namely: Δ VP_t＝ΔVm_t；

If the corrected oil holdup dynamic measurement error delta Vm_tThe convex character is convex and contains negative values, and then inverse folding and number addition pretreatment are needed, namely:

ΔVP_t＝-ΔVm_t+{|ΔVm₁|，|ΔVm₂|，…，|ΔVm_t|，…，|ΔVm_n|}_max;

wherein { |. DELTA.vm₁|，|ΔVm₂|，…，|ΔVm_t|，…，|ΔVm_n|}_maxFor operating on dataCoefficient, meaning | Δ Vm₁|，|ΔVm₂|，…，|ΔVm_t|，…，|ΔVm_nThe largest value of |;

step four: dynamically measuring error delta VP according to oil holdup after number addition and anti-pleat pretreatment_tEstablishing an oil holdup dynamic measurement error prediction model;

the method for establishing the oil holdup dynamic measurement error prediction model comprises the following steps:

establishing a dynamic measurement error prediction model of the oil holdup containing a modeling weight lambda as follows:

in the formula (I), the compound is shown in the specification,

dynamically measuring error delta VP of oil holdup after number addition and anti-pleat pretreatment_tThe predicted value of the GM (1, 1) model, its solution and

the solution method is the same;

dynamically measuring error delta VP of oil holdup after number addition and anti-pleat pretreatment_tThe predicted value of the NDGM (1, 1) model of (1) is expressed as follows:

in the formula, beta₁、β₂And beta₃Respectively is a multiple coefficient, a linear coefficient and an intercept of the NDGM (1, 1) model, and the solving method is as follows:

(β₁，β₂，β₃)^T＝(A^TA)^-1A^TM

wherein A is NDGM (1, 1) gray matrix, M is NDGM (1, 1) background matrix, and the expression is as follows:

β₄for the initial predicted value correction factor, the predicted value can be corrected by solving the predicted value

And oil holdup dynamic measurement error Δ VP_tThe value of (a) is obtained as the least squares sum of the errors, and is expressed as follows:

in the formula, j is a temporary statistical coefficient;

the model building weight lambda unconstrained optimization solution model is as follows:

obtaining a value of a modeling weight lambda by derivation;

step five: compensating the dynamic measurement error of the oil holdup caused by shunting by using a dynamic measurement error prediction model of the oil holdup;

the method for compensating the rate holding error caused by shunting by using the dynamic oil holding rate measurement error prediction model comprises the following steps:

in the formula, Vs_tThe compensated oil holdup is obtained.

Images