Well logging identification method and system for fluid properties of tight sandstone reservoir

Abstract

The invention discloses a well logging identification method and system for fluid properties of a tight sandstone reservoir, and belongs to the technical field of gas reservoir development. The method sequentially comprises the following steps that S1, logging curve parameters of a gas testing interval of the tight sandstone reservoir are obtained, and dimensionless processing is carried out; s2, calculating the correlation degree of the logging curve parameters and the gas saturation of the gas test interval by using a grey correlation degree method, and selecting the logging curve parameters with high correlation degree as sensitive logging parameters; s3, carrying out standardization processing on the sensitive logging parameters, and extracting principal components by utilizing a principal component analysis method; and S4, establishing quantitative discrimination functions of reservoirs with different fluid properties by taking the extracted principal components as basic parameters, and performing reservoir fluid property identification. The method is suitable for tight sandstone reservoir fluid property identification, the problems that the tight sandstone reservoir logging response relation is complex, the fluid logging response characteristics are not obvious, and the reservoir fluid type is difficult to identify are effectively solved, the method is verified in a plurality of difficult wells of the Sulige gas field, and the method is effective.

Description

Technical Field

[0001] This invention belongs to the field of gas reservoir development technology and relates to a method for identifying the fluid properties of tight sandstone reservoirs. Specifically, it is a well logging method and system for identifying the fluid properties of tight sandstone reservoirs. Background Technology

[0002] Reservoir fluid property identification is a crucial task in well logging interpretation and evaluation of tight sandstone reservoirs, and improving its accuracy plays a vital role in the efficient development of tight sandstone gas reservoirs. Compared with conventional sandstone, tight sandstone reservoirs have undergone multiple stages of complex diagenetic alteration, typically exhibiting poor physical properties, diverse pore types, rapid changes in pore-throat structure, strong heterogeneity, and complex gas-water relationships. This leads to complex reservoir logging response relationships, indistinct fluid logging response characteristics, and prominent anomalies such as high water-resistivity layers and low gas-resistivity layers, severely impacting the effectiveness of well logging identification of reservoir fluid properties.

[0003] Currently, well logging methods for identifying reservoir fluid properties mainly include curve overlay, cross plot, nuclear magnetic resonance (NMR) logging, and array acoustic logging. Curve overlay primarily utilizes the influence of changes in natural gas content in the reservoir on neutron, density, and acoustic logging responses, using the neutron "mining effect" to qualitatively identify gas-bearing layers. However, in tight sandstone reservoirs, the electrical differences between gas and water layers are small, and the neutron "mining effect" in gas layers is not significant, easily leading to missed gas layers. Cross plotting primarily utilizes the differences in logging response characteristics of reservoirs with different fluid types, selecting logging parameters sensitive to fluid properties to create cross plots and identify reservoir fluid properties. This method can quickly and intuitively identify reservoir fluid properties and is currently the most widely used method. However, the cross-plot method is only applicable to reservoirs with medium to high physical properties. For tight sandstone with complex electrical characteristics, it is difficult to distinguish clear electrical boundaries between reservoirs with different fluid properties. Furthermore, the cross-plot method can only use two logging parameters simultaneously for reservoir fluid property identification, resulting in a limited number of interpretable parameters. Additionally, the reservoir information reflected by various logging curves (rock skeleton, physical properties, pore structure, and fluid properties, etc.) can overlap and interfere with each other, adding complexity and ambiguity to reservoir fluid identification. Methods based on array acoustic logging, such as the equivalent elastic modulus difference ratio method and the P-wave and S-wave velocity method, are only suitable for identifying the fluid properties of medium to high porosity and permeability sandstone reservoirs. The shift spectrum method and difference spectrum method based on nuclear magnetic resonance (NMR) logging show good results in identifying the fluid properties of tight sandstone reservoirs, but special logging methods like NMR are expensive per well, and not every well has special logging data. Therefore, there is an urgent need to develop a new logging interpretation method for tight sandstone reservoir fluid properties based on conventional logging data to improve the accuracy of tight sandstone reservoir fluid property identification. Summary of the Invention

[0004] The purpose of this invention is to provide a well logging identification method and system for fluid properties in tight sandstone reservoirs, in order to solve the problems that conventional well logging identification methods are not applicable to tight sandstone structures, while special well logging identification methods are expensive.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A well logging method for identifying fluid properties in tight sandstone reservoirs includes the following steps performed sequentially:

[0007] S1. Obtain the logging curve parameters of the gas testing section of the tight sandstone reservoir and perform dimensionless processing;

[0008] S2. The correlation between the logging curve parameters and the gas saturation of the gas-testing interval is calculated using the grey relational analysis method, and the logging curve parameters with high correlation are selected as sensitive logging parameters.

[0009] S3. Standardize the sensitive logging parameters and extract the principal components using principal component analysis.

[0010] S4. Using the extracted principal components as basic parameters, establish quantitative discrimination functions for reservoirs with different fluid properties to identify reservoir fluid properties.

[0011] As a limitation, the dimensionless processing in step S1 is defined by the following formula:

[0012]

[0013] In the formula, X i The original logging curve parameters; X i ' represents the dimensionless logging curve parameters; X max X represents the maximum value of the original logging curve parameters. min This represents the minimum value of the original logging curve parameters.

[0014] As another limitation, step S2 specifically includes:

[0015] S21. Based on the dimensionless logging curve parameters and gas saturation data of the gas-testing section of the tight sandstone reservoir, establish a dataset matrix and calculate the grey relational degree.

[0016] S22. Sort the grey relational values ​​from largest to smallest, and select the top five logging curve parameters as sensitive logging parameters with strong correlation to reservoir fluid properties.

[0017] As a third limitation, step S3 specifically includes:

[0018] S31. Construct the original evaluation matrix X from the sensitive logging parameters:

[0019]

[0020] In the formula, n is the number of gas-testing intervals in the tight sandstone reservoir; p is the number of sensitive logging parameters; X np This is the p-th sensitive logging parameter of the n-th tight sandstone reservoir gas testing interval;

[0021] S32. Standardize the parameters in the original evaluation matrix X using the following equation:

[0022]

[0023] In the formula, It is the average value of the j-th sensitive logging parameter in all tested gas-bearing intervals; x ij is the j-th sensitive logging parameter in the i-th tight sandstone reservoir gas testing section; n is the number of samples in the tight sandstone reservoir gas testing section; It is the standard deviation of the j-th sensitive logging parameter; This is the standardized value of the j-th sensitive logging parameter in the i-th tight sandstone reservoir gas testing interval;

[0024] S33. Calculate the correlation coefficient matrix R of the standardized original evaluation matrix:

[0025]

[0026] In the formula, r ij Let be the covariance between the i-th standardized sensitive logging parameter and the j-th standardized sensitive logging parameter; p is the number of sensitive logging parameters; This is the standardized value of the i-th sensitive logging parameter in the k-th test gas section; This is the standardized value of the j-th sensitive logging parameter in the k-th test gas interval;

[0027] S34. Solve for the eigenvalues ​​and eigenvectors using the correlation coefficient matrix R:

[0028] |R-λ i E|=0;

[0029] (R-λ i E)u j =0

[0030] In the formula, R is the correlation coefficient matrix; λ i Let be the i-th eigenvalue; E is the identity matrix; u j For feature vectors;

[0031] S35. The eigenvalue λ i Sort by size from largest to smallest and calculate the cumulative variance contribution rate:

[0032]

[0033] In the formula, a m represents the cumulative variance contribution rate of the m-th eigenvalue; p represents the number of sensitive logging parameters.

[0034] S36. Calculate the principal component values ​​Y i :

[0035]

[0036] In the formula, u np This is the feature vector of the p-th evaluation variable for the nth test layer; Let p be the standardized evaluation variable.

[0037] As a further limitation, step S4 specifically includes:

[0038] Principal components with a cumulative variance contribution rate greater than 85% were selected as independent variables, and the fluid property type of the gas testing section of the tight sandstone reservoir was selected as the dependent variable. The Fisher discriminant method was used to establish a quantitative discriminant function for reservoirs with different fluid properties to identify reservoir fluid properties.

[0039] This invention also provides a well logging identification system for the fluid properties of tight sandstone reservoirs, comprising the following units in sequence from first to last:

[0040] The data preprocessing unit is used to obtain the logging curve parameters of the gas testing section of the tight sandstone reservoir and perform dimensionless processing.

[0041] The grey relational analysis unit is used to calculate the correlation between the logging curve parameters and the gas saturation of the gas-testing interval using the grey relational method, and selects the logging curve parameters with high correlation as sensitive logging parameters.

[0042] The principal component analysis unit is used to standardize sensitive logging parameters and extract principal components using principal component analysis.

[0043] The fluid property discrimination unit for tight sandstone reservoirs is used to establish quantitative discrimination functions for reservoirs with different fluid properties by using the extracted principal components as basic parameters, and to identify the fluid properties of the reservoirs.

[0044] By adopting the above technical solution, the technical progress achieved by this invention compared with the prior art is as follows:

[0045] This invention provides a well logging method and system for identifying fluid properties in tight sandstone reservoirs. First, conventional well logging parameters with high sensitivity to changes in fluid properties in tight sandstone reservoirs are optimized using grey relational analysis, and principal component analysis is performed on them. Then, based on the extracted principal components, Fisher's discriminant method is used to establish quantitative discriminant functions for different fluid properties to identify reservoir fluid properties. Grey relational analysis can determine the correlation between different well logging parameters and reservoir fluid properties, allowing for the optimization of well logging parameters and improving the accuracy and efficiency of identifying fluid properties in tight sandstone reservoirs under different geological backgrounds. Principal component analysis integrates information from multiple sensitive well logging parameters, extracting principal components to establish Fisher's discriminant functions for reservoir fluid property identification. This reduces the influence of redundant information in well logging parameters and correlations between well logging curves, thus improving the accuracy of fluid identification. Attached Figure Description

[0046] Figure 1 This is a flowchart of the well logging identification method for fluid properties in tight sandstone reservoirs in Example 1;

[0047] Figure 2 This is a correlation coefficient diagram between different logging curve parameters of tight sandstone reservoirs in the comparative example;

[0048] Figure 3 This is a cross plot of acoustic transit time and formation resistivity of a tight sandstone reservoir in the comparative example.

[0049] Figure 4 This is a cross plot of compensated density and formation resistivity for a tight sandstone reservoir in the comparative example.

[0050] Figure 5 This is a cross-plot of compensated neutron and formation resistivity in a tight sandstone reservoir, as shown in the comparative example.

[0051] Figure 6 This is a schematic diagram of the well logging identification system for fluid properties in tight sandstone reservoirs in Example 2. Detailed Implementation

[0052] The present invention will be further described in detail below through specific embodiments. It should be understood that the described embodiments are only for explaining the present invention and do not limit the present invention.

[0053] Example 1

[0054] This embodiment discloses a well logging method for identifying the fluid properties of tight sandstone reservoirs, the flowchart of which is shown below. Figure 1 As shown, this well logging identification method was applied to a study area in the Sulige Gas Field of the Ordos Basin, specifically including the following steps performed sequentially:

[0055] S1. Obtain the logging curve parameters of the gas-testing section of the tight sandstone reservoir and perform dimensionless processing.

[0056] Various conventional logging parameters were obtained for the gas testing intervals of tight sandstone reservoirs in the development wells within the study area, including formation true resistivity (RT), acoustic transit time (AC), compensated neutron (CNL), and compensated density (DEN). Table 1 shows the logging parameters and single-layer gas production data for 112 gas testing intervals from 30 development wells in the study area.

[0057] Table 1. Overview of logging curve parameters and single-layer gas production data for the gas testing section.

[0058]

[0059]

[0060] In Table 1, porosity and permeability are based on well logging interpretation results, while clay content is calculated from natural gamma (GR) logging data using the following formula:

[0061]

[0062] In the formula, GCUR is the Hillcutter index, whose value is related to the geological age of the strata; the GCUR is 2 for older strata and 3.7 for younger strata; I GR V is the clay content index, whose value is equal to the relative value of the natural gamma curve; sh The content of clay is %.

[0063] The above logging curve parameters are dimensionless using the following formula:

[0064]

[0065] In the formula, X i The original logging curve parameters; X i ' represents the dimensionless logging curve parameters; X max X represents the maximum value of the original logging curve parameters. min This represents the minimum value of the original logging curve parameters.

[0066] S2. The correlation between logging parameters and gas saturation in the testing zone is calculated using the grey relational analysis method. Logging parameters with high correlation are selected as sensitive logging parameters.

[0067] S21. Using the dimensionless logging curve parameter data of the gas-testing section of the tight sandstone reservoir as a subsequence and the gas saturation data of the reservoir in the gas-testing section as the parent sequence, arrange them sequentially with the parent sequence first and the subsequences last to establish a dimensionless dataset matrix:

[0068]

[0069] In the formula, m represents the number of types of logging curve parameters; n represents the number of samples from the gas testing interval of the tight sandstone reservoir; X' t (i) is the dimensionless logging curve parameter data matrix, i.e., the subsequence; Y t This is a dimensionless gas saturation data matrix, i.e., the parent sequence; (Y) t ,X' t (i) is a dimensionless dataset matrix consisting of the parent sequence and the child sequence data.

[0070] Based on the established dimensionless dataset matrix, the grey relational coefficient is calculated according to the following formula:

[0071]

[0072] In the formula, ξ i ρ is the grey relational coefficient of the i-th logging curve parameter; ρ is the resolution coefficient, whose value is between 0 and 1, typically 0.5; min t min i |X' t (i)-Y t | represents the minimum value of the difference matrix between each logging curve parameter data series and the gas saturation data series; max t max i |X' t (i)-Y t | represents the maximum value of the difference matrix between each logging curve parameter data series and the gas saturation data series; |X' t (i)-Y t | represents the absolute value of the difference between each logging curve parameter data series and the gas saturation data series.

[0073] The average value of the correlation coefficient between each logging curve parameter and the gas saturation of the reservoir in the testing section is calculated using the following formula, which is the grey relational degree:

[0074]

[0075] In the formula, r i Let be the correlation coefficient of the i-th logging curve parameter; is the correlation coefficient between the logging curve data value in the k-th row of the i-th column and the corresponding gas saturation data value; n is the number of samples in the gas testing section of the tight sandstone reservoir.

[0076] S22. Sort the grey relational degree values ​​from largest to smallest. The results are shown in Table 2:

[0077] Table 2. Grey Relational Ranking Table

[0078] order Well logging curve parameters Grey relational degree order Well logging curve parameters Grey relational degree 1 Sound wave time difference 0.783 5 mud content 0.726 2 Porosity 0.775 6 Penetration 0.586 3 Compensating neutrons 0.761 7 Formation resistivity 0.514 4 Compensation density 0.753

[0079] The top five logging parameters were selected as sensitive logging parameters with strong correlation to reservoir fluid properties, namely, sonic transit time (AC), porosity (Por), compensated neutron (CNL), compensated density (DEN), and shale content (Vsh).

[0080] S3. Standardize the sensitive logging parameters and extract principal components using principal component analysis.

[0081] S31. Construct the original evaluation matrix X from the selected sensitive logging parameters:

[0082]

[0083] In the formula, n is the number of gas-testing intervals in the tight sandstone reservoir; p is the number of sensitive logging parameters; X np This is the p-th sensitive logging parameter of the n-th tight sandstone reservoir gas testing interval;

[0084] S32. The parameters in the original evaluation matrix X are standardized using the following equation to eliminate differences in dimensions and orders of magnitude between parameters of different types of well logging curves:

[0085]

[0086] In the formula, It is the average value of the j-th sensitive logging parameter in all tested gas-bearing intervals; x ij is the j-th sensitive logging parameter in the i-th tight sandstone reservoir gas testing section; n is the number of samples in the tight sandstone reservoir gas testing section; It is the standard deviation of the j-th sensitive logging parameter; This is the standardized value of the j-th sensitive logging parameter in the i-th tight sandstone reservoir gas testing interval;

[0087] S33. Calculate the correlation coefficient matrix R of the standardized original evaluation matrix:

[0088]

[0089] In the formula, r ij Let be the covariance between the i-th standardized sensitive logging parameter and the j-th standardized sensitive logging parameter; p is the number of sensitive logging parameters; This is the standardized value of the i-th sensitive logging parameter in the k-th test gas section; This is the standardized value of the j-th sensitive logging parameter in the k-th test gas interval;

[0090] S34. Using the correlation coefficient matrix R, the eigenvalues ​​and eigenvectors are solved according to the following formula, and the results are shown in Table 3:

[0091] |R-λ i E|=0;

[0092] (R-λ i E)u j =0;

[0093] In the formula, R is the correlation coefficient matrix; λ i Let be the i-th eigenvalue (i≤p); E is the identity matrix; u j For feature vectors;

[0094] S35. The eigenvalue λ i Sort from largest to smallest, λ1≥λ2≥…≥λ p If the value is greater than 0, the cumulative variance contribution rate is calculated using the following formula, and the results are shown in Table 3:

[0095]

[0096] In the formula, a m represents the cumulative variance contribution rate of the m-th eigenvalue; p represents the number of sensitive logging parameters.

[0097] Table 3. Overview of Eigenvalues, Eigenvectors, and Variance Contribution Rates of Sensitive Logging Parameters

[0098]

[0099] S36. Use the following formula to obtain the principal component value Y. i :

[0100] Y1=0.527AC′-0.543DEN′+0.04CNL′+0.601Por′+0.253V′ sh

[0101] Y2=0.068AC′-0.058DEN′+0.729CNL′+0.121Por′-0.668V′ sh

[0102] Y3=0.716AC′+0.634DEN′+0.174CNL′-0.145Por′+0.182V′ sh

[0103] Y4=-0.17AC′+0.516DEN′-0.299CNL′+0.742Por′-0.254V′ sh

[0104] Y5=-0.418AC′+0.183DEN′+0.589CNL′+0.229Por′+0.626V′ sh

[0105] In the formula, AC', DEN', CNL', Por', V' sh These are the standardized values ​​of logging parameters for sonic transit time, compensated density, compensated neutrons, porosity, and clay content, respectively.

[0106] The cumulative variance contribution rate of principal components Y1 and Y2 is 86.99%, which means that principal components Y1 and Y2 can effectively represent 86.99% of the information of the five original sensitive logging parameters, with small data information loss and no correlation between them.

[0107] S4. Using the extracted principal components as basic parameters, establish quantitative discrimination functions for reservoirs with different fluid properties to identify reservoir fluid properties.

[0108] Principal components with a cumulative variance contribution rate greater than 85% were selected as independent variables, and the fluid property type of the gas-testing interval in tight sandstone reservoirs was selected as the dependent variable. Quantitative discriminant functions for reservoirs with different fluid properties were established using the Fisher discriminant method.

[0109] For air layers:

[0110] F(1)=0.428AC'-0.453DEN'-0.647CNL'+0.446Por'+0.863V' sh -2.183

[0111] Differential gas layer:

[0112] F(2)=-0.333AC'+0.363DEN'+0.79CNL'-0.337Por'-0.952V' sh -2.236

[0113] Gas and water in the same layer:

[0114] F(3)=-0.274AC'+0.287DEN'+0.298CNL'-0.293Por'-0.441V' sh -1.793

[0115] Aquifer:

[0116] F(4)=-0.625AC'+0.659DEN'+1.11CNL'-0.643Por'-1.42V' sh -3.067

[0117] Water layer:

[0118] F(5)=-1.241AC'+1.307DEN'+1.674CNL'-1.308Por'-2.307V' sh -5.961

[0119] In the formula, F(i) is the discriminant function for the i-th fluid property type reservoir; AC', DEN', CNL', Por', V' sh These are the standardized values ​​of logging parameters for sonic transit time, compensated density, compensated neutrons, porosity, and clay content, respectively.

[0120] Based on the Fisher posterior probability maximum discrimination rule (i.e. the obtained function value is the largest), the logging curve parameters of tight sandstone reservoirs with unknown fluid properties are substituted into each discrimination function, and the obtained function values ​​are compared to determine the fluid properties of the unexploited target reservoir and identify the reservoir fluid properties.

[0121] The following well logging identification method was applied to identify reservoir fluid properties in 112 gas-bearing test sections within the study area, including gas layers, gas-bearing layers, poorly gas-bearing layers, gas-water co-layers, water-bearing gas layers, and water layers, to verify the accuracy of the well logging identification method. The results are shown in Table 4.

[0122] Table 4 Reservoir fluid property identification results

[0123] reservoir types Number of samples Number of positive judgments Compliance rate / % air layer 65 60 92.3 Differential gas layer 7 6 85.7 Gas and water in the same layer 18 13 72.2 Aquifer 13 11 84.6 water layer 9 7 77.8 total 112 97 86.6

[0124] As shown in Table 4, the back-judgment statistics of the identification results of 112 gas-testing test sections in the study area, including gas-bearing sections, gas-differential gas sections, gas-water co-layers, water-bearing gas sections, and water sections, show that the overall number of positive judgments (i.e., the number of fluid properties determined by this logging identification method that match the actual situation) is 97, with an overall consistency rate of 86.6%. Among them, the consistency rate of gas layer identification is as high as 92.3%.

[0125] Comparative Example

[0126] The Person correlation coefficients between different logging parameters were calculated using Origin software, and a correlation heatmap was plotted. The results are as follows: Figure 2 As shown.

[0127] Depend on Figure 2It is known that there are varying degrees of correlation among the acoustic transit time (AC), porosity (Por), compensated neutron (CNL), compensated density (DEN), clay content (Vsh), permeability (Perm), and formation resistivity (RT) of the reservoir test sections in the study area. This will increase the degree of overlap between data points of reservoirs with different fluid properties in conventional cross plots and reduce the accuracy of conventional cross plot method in identifying reservoir fluid properties.

[0128] If traditional cross-plot methods are used to identify reservoir fluid properties in the test gas sections of this study area, such as cross-plots of acoustic transit time (AC) and formation resistivity (RT), cross-plots of compensated density (DEN) and formation resistivity (RT), and cross-plots of compensated neutron (CNL) and formation resistivity (RT), the results are as follows: Figure 3 , Figure 4 and Figure 5 As shown.

[0129] Traditional cross-plot methods can only use two logging parameters to identify reservoir fluid properties, resulting in a limited number of evaluation parameters and potential overlap between these parameters. Therefore, they are difficult to apply to identifying fluid properties in tight sandstone reservoirs. Figure 3 , Figure 4 and Figure 5 It is known that RT, AC, DEN, and CNL logging parameters can all reflect the rock properties, fluid type, and fluid saturation of the reservoir to varying degrees. However, the information reflected by these logging parameters is repetitive and similar, leading to correlations among them. Therefore, the logging parameter data points of the tight sandstone reservoirs with different fluid properties in the He8 section of this study area show significant overlap, and there are no clear boundaries between data of different fluid properties, making it difficult to accurately identify gas layers.

[0130] This invention provides a well logging identification method for fluid properties in tight sandstone reservoirs. First, it uses the grey relational analysis method to select conventional well logging parameters that are highly sensitive to changes in fluid properties in tight sandstone reservoirs. Then, it performs principal component analysis on these parameters, orthogonally transforming the high-dimensional data of the five original sensitive well logging parameters into five independent principal component variables. The principal components Y1 and Y2, with a cumulative variance contribution rate of 86.99%, not only integrate most of the information from the five original sensitive well logging parameters but also exhibit no correlation among themselves, overcoming the problems of limited parameters and correlation among parameters in conventional cross-plot methods. Using principal components Y1 and Y2 as identification parameters, and establishing quantitative discriminant function models for reservoirs with different fluid properties using the Fisher discriminant method, the accuracy of fluid property identification is greatly improved. This contributes to increasing the development efficiency and economic benefits of tight sandstone gas reservoirs, and is of great significance for exploration and development.

[0131] Example 2

[0132] This embodiment discloses a well logging identification system for the fluid properties of tight sandstone reservoirs, the structural schematic of which is shown below. Figure 6 As shown, the following units are included in the order from first to last:

[0133] Data preprocessing unit 1 is used to obtain the logging curve parameters of the gas testing section of the tight sandstone reservoir and perform dimensionless processing.

[0134] Grey relational analysis unit 2 is used to calculate the correlation between the logging curve parameters and gas saturation of the gas-testing interval using the grey relational method, and select logging curve parameters with high correlation as sensitive logging parameters.

[0135] Principal component analysis unit 3 is used to standardize sensitive logging parameters and extract principal components using principal component analysis.

[0136] The tight sandstone reservoir fluid property discrimination unit 4 is used to establish a quantitative discrimination function for reservoirs with different fluid properties by using the extracted principal components as basic parameters, and to identify the fluid properties of the reservoir.

Claims

1. A well logging method for identifying the fluid properties of tight sandstone reservoirs, characterized in that, This includes the following steps performed sequentially: S1. Obtain the logging curve parameters of the gas testing section of the tight sandstone reservoir and perform dimensionless processing; S2. The correlation between the logging curve parameters and the gas saturation of the gas-testing interval is calculated using the grey relational analysis method, and the logging curve parameters with high correlation are selected as sensitive logging parameters. S3. Standardize the sensitive logging parameters and extract the principal components using principal component analysis. S4. Using the extracted principal components as basic parameters, establish quantitative discrimination functions for reservoirs with different fluid properties to identify reservoir fluid properties.

2. The well logging identification method for fluid properties in tight sandstone reservoirs according to claim 1, characterized in that, The dimensionless processing in step S1 is calculated using the following formula: In the formula, X i The original logging curve parameters; X i ' represents the dimensionless logging curve parameters; X max X represents the maximum value of the original logging curve parameters. min This represents the minimum value of the original logging curve parameters.

3. The well logging identification method for fluid properties in tight sandstone reservoirs according to claim 1, characterized in that, Step S2 specifically includes: S21. Based on the dimensionless logging curve parameters and gas saturation data of the gas-testing section of the tight sandstone reservoir, establish a dataset matrix and calculate the grey relational degree. S22. Sort the grey relational values ​​from largest to smallest, and select the top five logging curve parameters as sensitive logging parameters with strong correlation to reservoir fluid properties.

4. A well logging method for identifying the fluid properties of tight sandstone reservoirs according to any one of claims 1-3, characterized in that, Step S3 specifically includes: S31. Construct the original evaluation matrix X from the sensitive logging parameters: In the formula, n is the number of gas-testing intervals in the tight sandstone reservoir; p is the number of sensitive logging parameters; X np This is the p-th sensitive logging parameter of the n-th tight sandstone reservoir gas testing interval; S32. Standardize the parameters in the original evaluation matrix X using the following equation: In the formula, It is the average value of the j-th sensitive logging parameter in all tested gas-bearing intervals; x ij is the j-th sensitive logging parameter in the i-th tight sandstone reservoir gas testing section; n is the number of samples in the tight sandstone reservoir gas testing section; It is the standard deviation of the j-th sensitive logging parameter; This is the standardized value of the j-th sensitive logging parameter in the i-th tight sandstone reservoir gas testing interval; S33. Calculate the correlation coefficient matrix R of the standardized original evaluation matrix: In the formula, r ij Let be the covariance between the i-th standardized sensitive logging parameter and the j-th standardized sensitive logging parameter; p is the number of sensitive logging parameters; This is the standardized value of the i-th sensitive logging parameter in the k-th test gas section; This is the standardized value of the j-th sensitive logging parameter in the k-th test gas interval; S34. Solve for the eigenvalues ​​and eigenvectors using the correlation coefficient matrix R: |R-λ i E|=0; (R-λ i E)u j =0 In the formula, R is the correlation coefficient matrix; λ i Let be the i-th eigenvalue; E is the identity matrix; u j For feature vectors; S35. The eigenvalue λ i Sort by size from largest to smallest and calculate the cumulative variance contribution rate: In the formula, a m represents the cumulative variance contribution rate of the m-th eigenvalue; p represents the number of sensitive logging parameters. S36. Calculate the principal component values ​​Y i : In the formula, u np This is the feature vector of the p-th evaluation variable for the nth test layer; Let p be the standardized evaluation variable.

5. The well logging identification method for fluid properties in tight sandstone reservoirs according to claim 4, characterized in that, Step S4 specifically includes: Principal components with a cumulative variance contribution rate greater than 85% were selected as independent variables, and the fluid property type of the gas testing section of the tight sandstone reservoir was selected as the dependent variable. The Fisher discriminant method was used to establish a quantitative discriminant function for reservoirs with different fluid properties to identify reservoir fluid properties.

6. A well logging system for identifying the fluid properties of tight sandstone reservoirs, characterized in that, Including the following units in order from first to last: The data preprocessing unit is used to obtain the logging curve parameters of the gas testing section of the tight sandstone reservoir and perform dimensionless processing. The grey relational analysis unit is used to calculate the correlation between the logging curve parameters and the gas saturation of the gas-testing interval using the grey relational method, and selects the logging curve parameters with high correlation as sensitive logging parameters. The principal component analysis unit is used to standardize sensitive logging parameters and extract principal components using principal component analysis. The fluid property discrimination unit for tight sandstone reservoirs is used to establish quantitative discrimination functions for reservoirs with different fluid properties by using the extracted principal components as basic parameters, and to identify the fluid properties of the reservoirs.

Images