Method for predicting dolomite order degree of carbonate gas reservoir

Abstract

The invention discloses a carbonate rock gas reservoir dolomite order degree prediction method, which comprises the following steps: S1, well logging parameters of the depth of a stratum where a measured dolomite sample is located are obtained, and the measured dolomite order degree measured value of the measured dolomite sample is measured through an X-ray diffraction method; s2, the logging parameters and the dolomite order degree measured value are fitted, and a dolomite order degree prediction model is constructed; s3, well logging parameters of the depth of the stratum where a to-be-measured dolomite sample is located are obtained, and the dolomite order degree of the to-be-measured dolomite sample is obtained through prediction in combination with the dolomite order degree prediction model. According to the method, the dolomite order degree can be accurately predicted through the established dolomite order degree prediction model, so that the dolomite order degree distribution condition of the non-cored well is obtained, and technical support is provided for carbonate rock gas reservoir development.

Description

Technical Field

[0001] The invention relates to the technical field of carbonate gas reservoir exploration and development, and in particular to a method for predicting the order degree of dolomite in carbonate gas reservoirs. Background Art

[0002] Carbonate rocks play an extremely important role in the natural gas production in the Sichuan Basin. The carbonate rock system accounts for 85% of the total conventional gas resources in the Sichuan Basin, and its proven reserves account for 70%. It is the main system for realizing the efficient exploration and development of natural gas in the Sichuan Basin. After more than 60 years of continuous exploration, a number of large and medium-sized gas fields have been obtained in carbonate rock systems such as the Middle Permian in the southern Sichuan Basin, the Weiyuan Sinian System, the Carboniferous System in the high and steep structural belt in the eastern Sichuan Basin, the Permian System in eastern Sichuan in the northern Sichuan Basin, and the Triassic reef beach. Among them, dolomite is the most important reservoir rock and an important target for the exploration and development of carbonate gas reservoirs.

[0003] The formation and evolution process of dolomite has a complex nucleation and precipitation dynamics mechanism. So far, predecessors have published many hypotheses and results on the genesis of dolomite, including microbial genesis, penecontemporaneous dolomitization, seepage reflux, hydrothermal genesis, burial dolomitization, etc. Among them, the order of dolomite can effectively reflect the structural characteristics of dolomite crystals. The order of dolomite will vary in different geological environments and diagenetic processes. Therefore, it is often used to analyze the genesis of dolomite and is an important indicator for judging the genesis of dolomite. How to accurately identify the genesis and evolution process of dolomite is an important scientific issue in the field of geology. If it is not clearly understood, it will inevitably hinder the further exploration and development of carbonate gas reservoirs. Therefore, accurate prediction of the order of dolomite in carbonate gas reservoirs can effectively guide oil and gas field companies in the exploitation of carbonate gas reservoirs.

[0004] At present, the order of dolomite in carbonate gas reservoirs is mainly tested by X-ray diffraction method, but this method can only measure the order of dolomite samples in the coring well section, and cannot obtain the distribution of dolomite order in non-coring wells. In addition, coring wells are relatively limited, small in number and expensive. Summary of the invention

[0005] In view of the above problems, the present invention aims to provide a method for predicting the order of dolomite in carbonate gas reservoirs.

[0006] The technical solution of the present invention is as follows:

[0007] A method for predicting the order of dolomite in carbonate gas reservoirs comprises the following steps:

[0008] S1: obtaining logging parameters of the formation depth of the measured dolomite sample, wherein the measured dolomite order value of the measured dolomite sample has been measured by an X-ray diffraction method;

[0009] S2: Fitting the logging parameters with the measured values ​​of the dolomite order degree to construct a dolomite order degree prediction model;

[0010] S3: obtaining the logging parameters of the formation depth where the dolomite sample to be tested is located, and combining the dolomite order prediction model to predict the dolomite order of the dolomite sample to be tested.

[0011] Preferably, in step S1, the logging parameters include acoustic time difference, compensated neutron, compensated density, natural gamma and resistivity.

[0012] Preferably, step S1 or step S1 and step S3 further include a step of preprocessing the logging parameters to remove abnormal values ​​in the logging parameters.

[0013] Preferably, the pretreatment specifically comprises the following sub-steps:

[0014] The average value of the original data of each logging parameter is obtained respectively, and the deviation and probability error between the original data of the logging parameter and the average value are calculated by the following formula:

[0015]

[0016] Where: D is the deviation; x i is the value of the original data of the i-th logging parameter; is the average value of the original data of the logging parameters; p is the probability error; n is the total number of original data of each logging parameter;

[0017] The D / p value of each logging parameter is calculated respectively. If the D / p value is greater than the D / p standard value, the corresponding logging parameter is an abnormal value.

[0018] Preferably, the D / p standard value is obtained by the following standard:

[0019] Among n original data, the deviation values ​​of all original data are obtained. If the deviation value of a certain original data is D1, when the probability of occurrence of all deviations equal to or greater than D1 is less than 1 / (2n), then the original data is an abnormal value.

[0020] Preferably, step S2 further includes the step of reducing the dimension of the logging parameters to obtain the parameters after dimension reduction; when fitting, the parameters after dimension reduction are fitted with the measured values ​​of the dolomite order degree.

[0021] Preferably, the dimension reduction of the logging parameters specifically includes the following sub-steps:

[0022] The raw data of the logging parameters are standardized using the following formula to obtain a standardized data matrix:

[0023]

[0024] Where: x i ′ j is the standardized x ij ;x ij is the parameter of row i and column j; is the average value of all parameters in the jth column; n is the total number of original data of each logging parameter;

[0025] The correlation coefficient is calculated according to the standardized data matrix to obtain a correlation coefficient matrix; the correlation coefficient is calculated by the following formula:

[0026]

[0027] Where: R ij is the correlation coefficient; m is the total number of rows / columns of the standardized data matrix; k represents the kth row / column of the standardized data matrix; x ik is the parameter of the i-th row and k-th column in the standardized data matrix; is the average value of all parameters in the i-th row of the standardized data matrix; x jk is the parameter of the jth column and the kth row in the standardized data matrix; is the average value of all parameters in the jth column of the standardized data matrix;

[0028] Establish a feature matrix |R-λE|=0, and obtain the eigenvalue λ of the feature matrix, cumulatively calculate the sum of the eigenvalues ​​from large to small, select the parameter whose sum of eigenvalues ​​is greater than the eigenvalue threshold as the parameter after dimensionality reduction, and set the expression of the parameter after dimensionality reduction to:

[0029] F i =x i AC+ x i+1 CNL+x i+2 DEN+x i+3 GR+x i+4 RT(i=1、2、3……) (5)

[0030] Where: F i is the i-th dimension reduction parameter; x i 、x i+1 、x i+2 、x i+3 、x i+4 is the corresponding eigenvalue λ i The characteristic vector of; AC is the acoustic time difference; CNL is the compensation neutron; DEN is the compensation density; GR is the natural gamma; RT is the resistivity;

[0031] Establish the equation system (R-λE)X=0, and find the non-zero solution of the equation system. The non-zero solution obtained is the corresponding eigenvalue λ i The eigenvector x i 、x i+1 、x i+2 、x i+3 、x i+4 , substitute the feature vector into the expression of the reduced dimension parameter to obtain the reduced dimension parameter.

[0032] Preferably, the eigenvalue threshold is 85% of the total eigenvalues.

[0033] Preferably, when there are only two parameters after dimension reduction and only one fitting is performed, the formula for one fitting is:

[0034] Y=aX1+bX2+c (6)

[0035]

[0036] In the formula: Y is the dependent variable parameter, dimensionless; a, b, c are all fitting coefficients, dimensionless; X1, X2 are independent variable parameters, dimensionless; Y i is the i-th data of the dependent variable parameter; X 1i , X 2i They are the i-th data of the first independent variable parameter and the second independent variable parameter respectively; are the average values ​​of the dependent variable parameter, the first independent variable parameter, and the second independent variable parameter respectively;

[0037] The dolomite order prediction model obtained by a fitting is:

[0038] Q1=aF1+bF2+c (10)

[0039] Where: Q1 is the predicted value of dolomite order after a first fitting, dimensionless; F1 and F2 are the parameters after dimensionality reduction, dimensionless.

[0040] Preferably, when there are only two parameters after the dimension reduction, and a secondary fitting is performed on the basis of the primary fitting, the formula for the secondary fitting is:

[0041] Y=d+eX (11)

[0042]

[0043] Where: e and d are both fitting coefficients, dimensionless; X i is the i-th data of the independent variable parameter; is the mean value of the independent variable parameter;

[0044] The dolomite order prediction model obtained by performing quadratic fitting is:

[0045] Q2=e(aF1+bF2+c)+d (14)

[0046] Where: Q2 is the predicted value of dolomite order by quadratic fitting, dimensionless.

[0047] The beneficial effects of the present invention are:

[0048] The dolomite order prediction model established in the present invention can accurately predict the dolomite order, thereby obtaining the dolomite order distribution of non-coring wells, and providing technical support for the development of carbonate gas reservoirs. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0050] Figure 1 It is a correlation diagram between the measured value of dolomite order and the first fitting value in a specific embodiment. DETAILED DESCRIPTION

[0051] The present invention is further described below in conjunction with the accompanying drawings and embodiments. It should be noted that, in the absence of conflict, the embodiments in this application and the technical features in the embodiments can be combined with each other. It should be noted that, unless otherwise specified, all technical and scientific terms used in this application have the same meanings as those generally understood by those of ordinary skill in the art to which this application belongs. The words "including" or "comprising" and the like used in the disclosure of the present invention mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects.

[0052] The present invention provides a method for predicting the order of dolomite in carbonate gas reservoirs, comprising the following steps:

[0053] S1: obtaining logging parameters of the formation depth where the measured dolomite sample is located, wherein the measured dolomite order degree measured value of the measured dolomite sample has been measured by X-ray diffraction method.

[0054] In a specific embodiment, the logging parameters include acoustic wave time difference, compensated neutron, compensated density, natural gamma and resistivity.

[0055] In a specific embodiment, this step also includes the step of preprocessing the logging parameters to remove abnormal values ​​in the logging parameters. Optionally, the preprocessing specifically includes the following sub-steps:

[0056] The average value of the original data of each logging parameter is obtained respectively, and the deviation and probability error between the original data of the logging parameter and the average value are calculated by the following formula:

[0057]

[0058] Where: D is the deviation; x i is the value of the original data of the i-th logging parameter; is the average value of the original data of the logging parameters; p is the probability error; n is the total number of original data of each logging parameter;

[0059] The D / p value of each logging parameter is calculated respectively. If the D / p value is greater than the D / p standard value, the corresponding logging parameter is an abnormal value.

[0060] In a specific embodiment, the D / p standard value is obtained by the following standard: among n raw data, the deviation values ​​of all raw data are obtained. If the deviation value of a certain raw data is D1, when the probability of occurrence of all deviations equal to or greater than D1 is less than 1 / (2n), then the raw data is an abnormal value. The standard of this embodiment can be used to establish a D / p standard value table corresponding to the number of raw data n, and the results are shown in Table 1:

[0061] Table 1D / p standard value table

[0062] n 5 10 15 20 50 100 D / p 2.5 2.9 3.2 3.3 3.8 4.2

[0063] It should be noted that n in Table 1 is only some conventional values, and the D / p standard values ​​corresponding to other n values ​​can be obtained through the standards of the above embodiments.

[0064] It should be noted that the purpose of preprocessing the logging parameters is mainly to remove outliers in the logging parameters, reduce errors, and avoid outliers affecting the prediction accuracy of the final model. In addition to the preprocessing method in the above embodiment, other methods for identifying and removing outliers in the prior art can also be applied to the present invention, such as box plot detection method, mathematical statistical method (mean, standard deviation, etc.), discrete point test method, etc.

[0065] S2: Fitting the logging parameters with the measured values ​​of the dolomite order degree to construct a dolomite order degree prediction model.

[0066] In a specific embodiment, this step also includes the step of reducing the dimension of the logging parameters to obtain the parameters after dimension reduction; when fitting, the parameters after dimension reduction are fitted with the measured values ​​of the dolomite order degree.

[0067] In the above embodiment, by reducing the dimension of the logging parameters, the parameters required for modeling can be reduced, and only a small number of parameters can be used to include the information of previous multiple parameters, which simplifies the dolomite order prediction model and can also make accurate predictions.

[0068] In a specific embodiment, the dimension reduction of the logging parameters specifically includes the following sub-steps:

[0069] The raw data of the logging parameters are standardized using the following formula to obtain a standardized data matrix:

[0070]

[0071] Where: x i ′ j is the standardized x ij ;x ij is the parameter of row i and column j; is the average value of all parameters in the jth column; n is the total number of original data of each logging parameter;

[0072] The correlation coefficient is calculated according to the standardized data matrix to obtain a correlation coefficient matrix; the correlation coefficient is calculated by the following formula:

[0073]

[0074] Where: R ij is the correlation coefficient; m is the total number of rows / columns of the standardized data matrix; k represents the kth row / column of the standardized data matrix; x ik is the parameter of the i-th row and k-th column in the standardized data matrix; is the average value of all parameters in the i-th row of the standardized data matrix; x jk is the parameter of the jth column and the kth row in the standardized data matrix; is the average value of all parameters in the jth column of the standardized data matrix;

[0075] Establish a feature matrix |R-λE|=0, and obtain the eigenvalue λ of the feature matrix, cumulatively calculate the sum of the eigenvalues ​​from large to small, select the parameter whose sum of eigenvalues ​​is greater than the eigenvalue threshold as the parameter after dimensionality reduction, and set the expression of the parameter after dimensionality reduction to:

[0076] F i =x i AC+ xi+1 CNL+x i+2 DEN+x i+3 GR+x i+4 RT(i=1、2、3……) (5)

[0077] Where: F i is the i-th dimension reduction parameter; x i 、x i+1 、x i+2 、x i+3 、x i+4 is the corresponding eigenvalue λ i The characteristic vector of; AC is the acoustic time difference; CNL is the compensation neutron; DEN is the compensation density; GR is the natural gamma; RT is the resistivity;

[0078] Establish the equation system (R-λE)X=0, and find the non-zero solution of the equation system. The non-zero solution obtained is the corresponding eigenvalue λ i The eigenvector x i 、x i+1 、x i+2 、x i+3 、x i+4 , substitute the feature vector into the expression of the reduced dimension parameter to obtain the reduced dimension parameter.

[0079] In a specific embodiment, the eigenvalue threshold is 85% of the total eigenvalue. It should be noted that the eigenvalue threshold affects the dimensionality reduction effect. The larger the eigenvalue threshold, the worse the dimensionality reduction effect, but it contains more information about the previous multi-parameters; the smaller the eigenvalue threshold, the better the dimensionality reduction effect, but it contains less information about the previous multi-parameters, which will affect the prediction accuracy of the established model. The specific value of the eigenvalue threshold can be selected according to the target prediction accuracy.

[0080] In a specific embodiment, when there are only two parameters after dimension reduction and only one fitting is performed, the formula for one fitting is:

[0081] Y=aX1+bX2+c (6)

[0082]

[0083] In the formula: Y is the dependent variable parameter, dimensionless; a, b, c are all fitting coefficients, dimensionless; X1, X2 are independent variable parameters, dimensionless; Y i is the i-th data of the dependent variable parameter; X 1i , X 2i They are the i-th data of the first independent variable parameter and the second independent variable parameter respectively; are the average values ​​of the dependent variable parameter, the first independent variable parameter, and the second independent variable parameter respectively;

[0084] The dolomite order prediction model obtained by a fitting is:

[0085] Q1=aF1+bF2+c (10)

[0086] Where: Q1 is the predicted value of dolomite order after a first fitting, dimensionless; F1 and F2 are the parameters after dimensionality reduction, dimensionless.

[0087] In a specific embodiment, when there are only two parameters after the dimension reduction, and a secondary fitting is performed on the basis of the primary fitting, the formula for the secondary fitting is:

[0088] Y=d+eX (11)

[0089]

[0090] Where: e and d are both fitting coefficients, dimensionless; X i is the i-th data of the independent variable parameter; is the mean value of the independent variable parameter;

[0091] The dolomite order prediction model obtained by performing quadratic fitting is:

[0092] Q2=e(aF1+bF2+c)+d (14)

[0093] Where: Q2 is the predicted value of dolomite order by quadratic fitting, dimensionless.

[0094] In the above embodiment, the present invention performs a secondary fitting on the basis of the primary fitting, which can further reduce the error and improve the prediction accuracy of the dolomite order prediction model.

[0095] S3: obtaining the logging parameters of the formation depth where the dolomite sample to be tested is located, and combining the dolomite order prediction model to predict the dolomite order of the dolomite sample to be tested.

[0096] In a specific embodiment, taking a carbonate gas reservoir as an example, the dolomite order of a dolomite sample to be tested is predicted using the method for predicting the dolomite order of a carbonate gas reservoir according to the present invention, specifically comprising the following steps:

[0097] (1) The logging parameters of the formation depth of the measured dolomite sample in a carbonate gas reservoir and the measured dolomite order value are obtained. The results are shown in Table 2:

[0098] Table 2 Logging parameters of dolomite samples from a carbonate gas reservoir and measured values ​​of dolomite order

[0099]

[0100]

[0101] (2) preprocessing the logging parameters to remove abnormal values ​​in the logging parameters;

[0102] First, calculate the average value of each parameter in Table 2 The deviation D of each parameter from the corresponding mean value and its D 2 ; Secondly, calculate the probability error p of each parameter; then calculate the D / p value of each parameter; finally, compare the D / p value of each parameter with the D / p standard value. If its value is greater than the standard value, it is an abnormal value and is discarded. The calculation results are shown in Table 3 and Table 4:

[0103] Table 3 Preprocessing calculation results Table 1

[0104]

[0105]

[0106] Table 4 Preprocessing calculation results Table 2

[0107]

[0108]

[0109] In this embodiment, the total number n of raw data of each logging parameter is 18. According to the acquisition standard of the D / p standard value, it can be known that the corresponding D / p standard value is 3.2. It can be seen from Tables 3 and 4 that only the RT data in sample 10 is abnormal, and its RT = 247.628Ω·m, D = 131.874, D / p = 3.962>3.2, so the RT in this sample is an abnormal value. In order to reduce the error of the prediction equation, sample 10 is discarded.

[0110] (3) performing dimension reduction on the logging parameters to obtain dimension-reduced parameters;

[0111] First, the parameter data are standardized to eliminate the influence of the parameters on the level and dimension, and the correlation coefficient matrix of the standardized data is obtained. The results are shown in Table 5:

[0112] Table 5 Correlation coefficient matrix

[0113] parameter AC CNL DEN GR RT AC 1 0.752 -0.768 0.849 -0.789 CNL 0.752 1 -0.758 0.83 -0.828 DEN -0.768 -0.758 1 -0.869 0.749 GR 0.849 0.83 -0.869 1 -0.859 RT -0.789 -0.828 0.749 -0.859 1

[0114] Next, we build the feature matrix. The result is as follows:

[0115]

[0116] The eigenvalues ​​of the characteristic matrix are obtained, and the result is λ1=4.223; λ 2＝ 0.282; λ3=0.238; λ 4＝ 0.166;λ5=0.09.

[0117] in Therefore, the first two parameters after dimensionality reduction can represent the original five parameters of AC, CNL, DEN, GR, and RT.

[0118] Then establish the system of equations, the result is as follows:

[0119]

[0120] The non-zero solution of the equation system is obtained, and the result is x1=0.905; x2=0.907; x3=-0.902; x4=0.96; x5=-0.92; x6=-0.131; x7=0.291; x8=0.332; x9=-0.08; x 10 =-0.252.

[0121] Finally, the obtained non-zero solution is substituted into the expression of the dimension-reduced parameter shown in formula (5) to obtain the dimension-reduced parameter. The result is as follows:

[0122] F1=0.905AC+0.907CNL-0.902DEN+0.96GR-0.92RT (18)

[0123] F2=-0.131AC+0.291CNL+0.332DEN-0.08GR-0.252RT (19)

[0124] (4) using the dimension-reduced parameters and the measured values ​​of the dolomite order to perform a fit, and constructing a fitted dolomite order prediction model;

[0125] In this embodiment, the dolomite order prediction model obtained by performing a fitting construction using the fitting formulas shown in equations (6) to (9) is:

[0126] Q1=-0.003252F1+0.012F2+1.2973 (20)

[0127] The dolomite order prediction model of formula (20) is used to predict the dolomite order of the dolomite sample to be tested. The measured order of the dolomite sample to be tested and the related logging data are shown in Table 6. The prediction results are as follows: Figure 1 As shown in Table 7:

[0128] Table 6 Measured order of dolomite samples and related logging data

[0129] Samples to be tested Order AC(μs / m) CNL(V / V) <![CDATA[DEN(g / cm 3 )]]> GR(API) RT(Ω·m) D1 0.61 70.845 14.458 2.653 87.325 87.412 D2 0.91 51.083 7.201 2.672 49.461 187.722 D3 0.77 61.719 8.462 2.648 52.491 172.241 D4 0.79 53.693 9.148 2.681 61.512 152.708 D5 0.46 66.781 16.249 2.617 127.547 69.937 D6 0.52 62.695 15.794 2.638 121.376 85.142 D7 0.74 64.371 11.977 2.665 62.387 138.482 D8 0.51 66.218 15.912 2.621 112.794 79.375

[0130] Table 7 Dolomite order prediction results of the tested dolomite samples

[0131] sample Orderly measured value <![CDATA[F1]]> <![CDATA[F2]]> Orderly prediction value Relative error D1 0.61 78.25 -33.21 0.64 4.92％ D2 0.91 -74.87 -54.97 0.88 3.30％ D3 0.77 -46.93 -52.35 0.82 6.49％ D4 0.79 -26.97 -46.88 0.82 3.80％ D5 0.46 130.92 -30.98 0.50 8.70％ D6 0.52 106.87 -33.91 0.54 3.85％ D7 0.74 -0.80 -43.95 0.77 4.05％ D8 0.51 107.25 -32.20 0.56 9.80％

[0132] It can be seen from Table 7 that the order degree prediction value of the dolomite order degree prediction model of the present invention is consistent with the measured order degree value, with a small error, and the average relative error is only 4.49%, which can meet general work needs.

[0133] It should be noted that, in this embodiment, the dolomite order prediction model obtained by fitting the parameters after dimensionality reduction can accurately predict the dolomite order. Then the dolomite order prediction model obtained by directly fitting the logging parameters without dimensionality reduction can more accurately predict the dolomite order. However, the prediction model without dimensionality reduction requires more parameters and is more complicated.

[0134] (5) performing a secondary fitting on the basis of the primary fitting to construct a dolomite order prediction model of the secondary fitting;

[0135] It can be seen from Table 7 that there is a certain error between the predicted value of the dolomite order prediction model obtained by the first fitting and the measured value of the order. In order to reduce this error, the dolomite order prediction model obtained by the second fitting of the fitting formula shown in equations (11)-(13) is:

[0136] Q2=0997(-0.003252F1+0.012F2+1.2973)-0.013 (21)

[0137] The dolomite order prediction model of formula (21) is used to predict the dolomite order of the dolomite sample to be tested. The results are shown in Table 8:

[0138] Table 8 The measured values ​​of the order degree of the dolomite samples to be tested, the two predicted values ​​and the relative error analysis results

[0139]

[0140]

[0141] It can be seen from Table 8 that the error between the order degree prediction value of the dolomite order degree prediction model obtained by quadratic fitting and the measured order degree value is smaller, with an average error of only 3.16%.

[0142] In summary, the present invention can accurately predict the order of dolomite in non-coring wells by establishing a dolomite order prediction model, thereby obtaining the distribution of dolomite order in non-coring wells. Compared with the prior art, the present invention has significant progress.

[0143] The above description is only a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been disclosed as a preferred embodiment as above, it is not intended to limit the present invention. Any technician familiar with the profession can make some changes or modifications to equivalent embodiments of equivalent changes using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the scope of the technical solution of the present invention.

Claims

1. A method for predicting the order of dolomite in carbonate gas reservoirs, characterized in that: The following steps are involved: S1: obtaining logging parameters of the formation depth of the measured dolomite sample, wherein the measured dolomite order value of the measured dolomite sample has been measured by an X-ray diffraction method; S2: Fitting the logging parameters with the measured values ​​of the dolomite order degree to construct a dolomite order degree prediction model; S3: obtaining the logging parameters of the formation depth where the dolomite sample to be tested is located, and combining the dolomite order prediction model to predict the dolomite order of the dolomite sample to be tested.

2. The method for predicting the order degree of dolomite in carbonate gas reservoirs according to claim 1, characterized in that: In step S1, the logging parameters include acoustic time difference, compensated neutron, compensated density, natural gamma and resistivity.

3. The method for predicting the order degree of dolomite in carbonate gas reservoirs according to claim 1, characterized in that: Step S1 or step S1 and step S3 also include a step of preprocessing the logging parameters to remove abnormal values ​​in the logging parameters.

4. The method for predicting the order degree of dolomite in carbonate gas reservoirs according to claim 3, characterized in that: The preprocessing specifically includes the following sub-steps: The average value of the original data of each logging parameter is obtained respectively, and the deviation and probability error between the original data of the logging parameter and the average value are calculated by the following formula: Where: D is the deviation; x i is the value of the original data of the i-th logging parameter; is the average value of the original data of the logging parameters; p is the probability error; n is the total number of original data of each logging parameter; The D / p value of each logging parameter is calculated respectively. If the D / p value is greater than the D / p standard value, the corresponding logging parameter is an abnormal value.

5. The method for predicting the order degree of dolomite in carbonate gas reservoirs according to claim 4, characterized in that: The D / p standard value is obtained by the following standard: Among n original data, the deviation values ​​of all original data are obtained. If the deviation value of a certain original data is D1, when the probability of occurrence of all deviations equal to or greater than D1 is less than 1 / (2n), then the original data is an abnormal value.

6. The method for predicting the order degree of dolomite in carbonate gas reservoirs according to claim 2, characterized in that: Step S2 also includes the step of reducing the dimension of the logging parameters to obtain the parameters after dimension reduction; when fitting, the parameters after dimension reduction are fitted with the measured values ​​of the dolomite order degree.

7. The method for predicting the order degree of dolomite in carbonate gas reservoirs according to claim 6, characterized in that: The dimension reduction of the logging parameters specifically includes the following sub-steps: The raw data of the logging parameters are standardized using the following formula to obtain a standardized data matrix: Where: x i ′ j is the standardized x ij ;x ij is the parameter of row i and column j; is the average value of all parameters in the jth column; n is the total number of original data of each logging parameter; The correlation coefficient is calculated according to the standardized data matrix to obtain a correlation coefficient matrix; the correlation coefficient is calculated by the following formula: Where: R ij is the correlation coefficient; m is the total number of rows / columns of the standardized data matrix; k represents the kth row / column of the standardized data matrix; x ik is the parameter of the i-th row and k-th column in the standardized data matrix; is the average value of all parameters in the i-th row of the standardized data matrix; x jk is the parameter of the jth column and the kth row in the standardized data matrix; is the average value of all parameters in the jth column of the standardized data matrix; Establish a feature matrix |R-λE|=0, and obtain the eigenvalue λ of the feature matrix, cumulatively calculate the sum of the eigenvalues ​​from large to small, select the parameter whose sum of eigenvalues ​​is greater than the eigenvalue threshold as the parameter after dimensionality reduction, and set the expression of the parameter after dimensionality reduction to: F i =x i AC+x i+1 CNL+x i+2 DEN+x i+3 GR+x i+4 RT(i=1、2、3……) (5) Where: F i is the i-th dimension reduction parameter; x i 、x i+1 、x i+2 、x i+3 、x i+4 is the corresponding eigenvalue λ i The characteristic vector of; AC is the acoustic time difference; CNL is the compensation neutron; DEN is the compensation density; GR is the natural gamma; RT is the resistivity; Establish the equation system (R-λE)X=0, and find the non-zero solution of the equation system. The non-zero solution obtained is the corresponding eigenvalue λ i The eigenvector x i 、x i+1 、x i+2 、x i+3 、x i+4 , substitute the feature vector into the expression of the reduced dimension parameter to obtain the reduced dimension parameter.

8. The method for predicting the order degree of dolomite in carbonate gas reservoirs according to claim 7, characterized in that: The eigenvalue threshold is 85% of the total eigenvalues.

9. The method for predicting the order degree of dolomite in carbonate gas reservoirs according to claim 7, characterized in that: When there are only two parameters after dimension reduction and only one fitting is performed, the formula for one fitting is: Y=aX1+bX2+c (6) In the formula: Y is the dependent variable parameter, dimensionless; a, b, c are all fitting coefficients, dimensionless; X1, X2 are independent variable parameters, dimensionless; Y i is the i-th data of the dependent variable parameter; X 1i , X 2i They are the i-th data of the first independent variable parameter and the second independent variable parameter respectively; are the average values ​​of the dependent variable parameter, the first independent variable parameter, and the second independent variable parameter respectively; The dolomite order prediction model obtained by a fitting is: Q1=aF1+bF2+c (10) Where: Q1 is the predicted value of dolomite order after a first fitting, dimensionless; F1 and F2 are the parameters after dimensionality reduction, dimensionless.

10. The method for predicting the order degree of dolomite in carbonate gas reservoirs according to claim 9, characterized in that: When there are only two parameters after dimension reduction, and a second fitting is performed on the basis of the first fitting, the formula for the second fitting is: Y=d+eX (11) Where: e and d are both fitting coefficients, dimensionless; X i is the i-th data of the independent variable parameter; is the mean value of the independent variable parameter; The dolomite order prediction model obtained by performing quadratic fitting is: Q2=e(aF1+bF2+c)+d (14) Where: Q2 is the predicted value of dolomite order by quadratic fitting, dimensionless.

Images