A method for quantitatively characterizing pore structure of low-permeability reservoirs based on multi-peak fitting of T2 spectrum of nuclear magnetic resonance

Abstract

The application discloses a low-permeability reservoir pore structure quantitative characterization method based on a nuclear magnetic resonance T2 spectrum multi-peak fitting, relates to the technical field of reservoir pore structure characterization, and comprises the following steps: preprocessing a nuclear magnetic resonance T2 spectrum, and decomposing the data after preprocessing by means of Gaussian multi-peak fitting; combining the peak area, the peak center and the peak width of each sub-peak to quantitatively characterize the porosity contribution and the pore size range of micro-pores, mesopores and macropores; introducing the sub-peak parameters obtained by multi-peak fitting into a permeability prediction model, correcting the model for predicting the permeability based on NMR T2 relaxation data based on the multi-peak fitting results, constructing a permeability model, and more accurately predicting the permeability, which has important significance for promoting reservoir prediction.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of reservoir pore structure characterization, in particular to a low-permeability reservoir pore structure quantitative characterization method based on multi-peak fitting of nuclear magnetic resonance T2 spectrum. BACKGROUND

[0002] In the process of oil and gas exploration and development, accurate pore structure characterization plays a crucial role in evaluating reservoir quality, predicting permeability, and optimizing production schemes. At present, for low-permeability sandstone reservoirs, various experimental methods have been developed for pore structure characterization, such as nanometer CT scanning, scanning electron microscopy (SEM), high-pressure mercury injection (HPMI), nuclear magnetic resonance (NMR) measurement, and low-temperature nitrogen / carbon dioxide adsorption. These techniques can help researchers understand key indicators such as rock porosity, pore throat distribution, and pore connectivity to some extent.

[0003] Nuclear magnetic resonance (NMR) technology has the advantages of non-destructiveness, testing under near-in-situ conditions, and measuring results reflecting fluid properties in pores, making it a common pore structure evaluation method in oilfield sites and research laboratories. Among them, NMR T2 spectrum has a certain correspondence with pore size, and better reflects the characteristics of porosity distribution and pore size distribution.

[0004] In practical applications, traditional NMR analysis methods mainly rely on parameterization techniques (such as FFI, BVI, T 2cutoff , T 2gm , etc.) and qualitative peak type judgment (such as single peak, double peak, and triple peak) to divide or induce pore types. However, with the increasing complexity of oil reservoir types, this method still has shortcomings in distinguishing multiple types of pores (micro-pores, mesopores, and macropores) and accurately extracting the physical meaning of each sub-peak.

[0005] Defects and deficiencies of the prior art:

[0006] (1) Qualitative division relies on experience and has poor stability

[0007] In the early stage, many studies compared the NMR T2 spectrum shapes of saturated brine and centrifuged samples, or relied on researchers' experience to distinguish single peaks and double peaks to qualitatively identify different pore structures. This method relies on manual experience, is time-consuming, and often has poor stability, making it difficult to adapt to large-scale and diversified sample analysis.

[0008] (2) The existing parameter method has limited scope of application

[0009] Common NMR parameters (such as T 2cutoffThe methods (such as FFI, BVI, etc.) can only divide the pores into two categories of "bound fluid" and "free fluid", and it is difficult to further subdivide more types of pores such as micropores, mesopores and macropores. In addition, the parameters (such as T 2gm , T 2max , T 2mid , etc.) for overall description of T2 spectrum morphology are highly correlated with each other, and it is difficult to highlight the key differences in pore structure.

[0010] (3) Lack of clear explanation of the physical meaning of the parameters after multi-peak fitting

[0011] Although the previous studies involving multi-peak fitting can effectively improve the resolution of complex T2 spectrum, the physical meaning of the peak area, peak center and peak width of each sub-peak is not systematically explained, which limits the application of the method in engineering practice.

[0012] (4) Lack of connection with subsequent reservoir permeability prediction

[0013] Some studies only stay at the peak decomposition stage of NMR T2 spectrum, and do not establish a close connection with subsequent permeability prediction, making it difficult to form a complete analysis and application process based on multi-peak fitting. SUMMARY

[0014] To solve the above technical problems, the present application provides a method for quantitative characterization of pore structure in low-permeability reservoirs based on multi-peak fitting of nuclear magnetic resonance T2 spectrum, which pre-processes the nuclear magnetic resonance T2 spectrum and decomposes the pre-processed data by Gaussian multi-peak fitting;

[0015] The peak area, peak center and peak width of each sub-peak are combined to quantitatively characterize the porosity and pore size distribution of micropores, mesopores and macropores;

[0016] The sub-peak parameters obtained by multi-peak fitting are introduced into the permeability prediction model, the permeability prediction model based on NMR T2 relaxation data is modified based on the multi-peak fitting results, and the permeability model is constructed.

[0017] In the preferred embodiment, the nuclear magnetic resonance T2 spectrum is pre-processed to obtain the nuclear magnetic porosity

[0018]

[0019] wherein, is the porosity of the i-th component T 2i of the transverse relaxation time T2, and n is the total number of components;

[0020] The actual cumulative distribution function F * of the nuclear magnetic porosity is:

[0021]

[0022] Actual porosity probability density function f * (logT2) is:

[0023]

[0024] In the preferred embodiment, suppose the pre-processed data distribution is formed by linear superposition of multiple Gaussian distributions, the porosity probability density function f (logT2) of the total peak is:

[0025]

[0026] Where y0 is the baseline, f j (logT2) is the porosity probability density function of the sub-peak, A j is the area of the jth sub-peak, G j (logT2) is the Gaussian distribution of the jth sub-peak, and m is the number of sub-peaks.

[0027] The area A j of the jth sub-peak and the NMR porosity Satisfy the following relationship:

[0028]

[0029] Where logT 2min and logT 2max are the minimum and maximum values of the log horizontal relaxation time T2, respectively; and the Gaussian distribution function is transformed as:

[0030]

[0031] Where w j is the full width at half maximum FWHM of the jth sub-peak; and c j is the center of the jth sub-peak.

[0032] In the preferred embodiment, the R-square value reflecting the degree of fitting is calculated as follows:

[0033]

[0034] Where N is the total number of data points of a nuclear magnetic resonance T2 spectrum after Gaussian multi-peak fitting, f k * (logT2) is the kth data value of the actual total peak porosity probability density function f k (logT2) is the kth data value of the total peak porosity probability density function after fitting, and is the average value of the porosity probability density function of the total peaks after Gaussian multi-peak fitting.

[0035] In the preferred embodiment, in the permeability prediction model, T 2gm :

[0036]

[0037] Taking logarithm on both sides simplifies to:

[0038]

[0039] log(T 2gm ) is expressed by the porosity probability density function of the total number of peaks obtained by Gaussian multi-peak fitting:

[0040]

[0041] In the preferred embodiment, logT' 2gm is redefined as the logarithmic geometric mean value based on the peak parameters of Gaussian multi-peak fitting:

[0042]

[0043] In the preferred embodiment, log(T 2gm ) is approximated by log(T' 2gm ) calculated from the peak parameters after Gaussian multi-peak fitting, and log(T' 2gm ) is applied to the permeability prediction model, and the permeability K equation becomes:

[0044]

[0045] The permeability prediction model is modified to exclude the peak parameters related to micropores, and the calculation is as follows:

[0046]

[0047] Where a, b and c are constants.

[0048] Compared with the prior art, the present application has the following beneficial technical effects:

[0049] 1. Quantitative decomposition of complex T2 spectrum, overcoming the limitations of relying on empirical qualitative division

[0050] The present application decomposes NMR T2 spectrum into several independent components through Gaussian multi-peak fitting, each component corresponds to different pore size range and physical property characteristics, greatly improving the ability to distinguish micro-pores, mesopores and macropores in complex reservoirs, and reducing the dependence on experience judgment of researchers.

[0051] The pore components in T2 spectrum are usually distinguished by fixed values, but in this method, there is no fixed boundary between different types of pores, and they are all concentrated near the peak center. The farther away from the peak center, the fewer the types of pores. Figure 3 B illustrates the crossover phenomenon between different types of pores. For example, sub-peak 1 and sub-peak 2 intersect at log(T2) value 0.8, and the content of peak 3 is zero at this point. This observation shows that at this pore size, the sample contains equal amounts of micropores and mesopores, and no macropores exist. Compared with previous studies, this finding more accurately reflects the actual pore size distribution, providing a deeper understanding of the pore distribution pattern.

[0052] 2. Accurate extraction of multi-peak parameters and explanation of physical meaning, enhancing the interpretability of the results

[0053] Each sub-peak can be quantified by peak area, peak center, and peak width. Peak area reflects the proportion of porosity, peak center reflects the center of pore size distribution, and peak width represents the range of pore size distribution. This parameterized result has a clear rock physics meaning and is convenient for application in reservoir evaluation, numerical simulation, and field interpretation.

[0054] 3. Improve the permeability prediction model and improve the prediction accuracy

[0055] By introducing the sub-peak parameters obtained by multi-peak fitting into the permeability prediction model, the dependence of traditional methods on single T 2cutoff or T 2gm is effectively overcome. Figure 5 A shows the results of the original SDR model, with an R-square value of 0.8202, Figure 5 B shows the results of the modified model, indicating that the R-square value has increased to 0.8718. This shows that the modified model has higher accuracy and can more accurately predict permeability, which is of great significance for promoting reservoir prediction. BRIEF DESCRIPTION OF DRAWINGS

[0056] Figure 1 A represents the principle of Gaussian multi-peak fitting algorithm of NMR T2 spectrum; B represents the specific parameters A j , w j , and c j of sub-peak;

[0057] Figure 2 A represents the original NMR T2 spectrum of low permeability sandstone samples in Daluhu area; B represents the spectrum after Gaussian multi-peak fitting in Daluhu area;

[0058] Figure 3 A represents the original NMR T2 spectrum of low permeability sandstone samples in Mahaidong area; B represents the spectrum after Gaussian multi-peak fitting in Mahaidong area;

[0059] Figure 4 log(T 2gm ) vs. log(T 2gm ) crossplot for all low permeability sandstone samples in Mahidong area;

[0060] Figure 5 Fig. A represents the log permeability vs. coefficient relationship of the original SDR model in Mahidong area; Fig. B represents the three-dimensional relationship of the log permeability vs. coefficient of the modified SDR model in Mahidong area. DETAILED DESCRIPTION

[0061] The technical solutions in the embodiments of the present application will be apparently and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0062] The present application proposes a quantitative characterization method of pore structure of low permeability reservoir based on multi-peak fitting of NMR T2 spectrum for low permeability sandstone reservoir or other complex reservoirs. Firstly, the NMR T2 spectrum is preprocessed, and is regarded as linear superposition of several log Gaussian distributions, and the preprocessed data is decomposed by Gaussian multi-peak fitting; then, the pore volume contribution and pore size range of micro-pore, meso-pore and macro-pore are accurately characterized by combining the peak area A, peak center c and peak width w of each sub-peak; and the results are applied to permeability prediction.

[0063] 1. NMR T2 spectrum data preprocessing and conversion.

[0064] Preliminary analysis shows that when the transverse relaxation time T2 exceeds 1000 ms, micro-pores corresponding to micro-fractures are observed in the sandstone sample. These micro-fractures are negligible in quantity, and thus are excluded from the experimental data processing. In addition, since the original data represents the increment of porosity, it is converted into a porosity probability density function for more accurate data analysis. The calculation process is as follows.

[0065] The sum of the porosity increments of all sampling points is the porosity of the sample, which can be expressed by the following formula:

[0066]

[0067] wherein, is the NMR porosity of the sample, is the porosity of the i-th component T 2i of the transverse relaxation time T2 (also referred to as porosity increment), and there are n components.

[0068] Therefore, the actual cumulative distribution function F of NMR porosity * The following formula can be used to calculate:

[0069]

[0070] Finally, the actual porosity probability density function f * It can be calculated using the logarithmic differential of equation (2), as shown below:

[0071]

[0072] 2. Gaussian multi-peak fitting

[0073] In the initial stage of the Gaussian peak fitting algorithm, the experimental data is carefully observed to identify potential components or sub-peaks. Subsequently, a series of Gaussian distributions are used as models to accurately characterize these components, each representing an independent component. The algorithm uses a nonlinear least squares method to iteratively optimize the parameters of the Gaussian model, aiming to minimize the deviation between the model-fitted composite curve and the experimental data. The R-squared value is used to evaluate the fit between the processed distribution and the actual distribution, ensuring the accuracy and effectiveness of the model.

[0074] Assume the preprocessed data distribution is formed by the linear superposition of several Gaussian distributions. Figure 1 Figure A in the figure has the following mathematical expression:

[0075]

[0076] Where f(logT2) is the porosity probability density function obtained after applying Gaussian multi-peak fitting, summing the total peaks, and representing the actual porosity probability density f. * An exact fit; y0 is the baseline, in which case: y0 = 0; f j (logT2) is the porosity probability density function of the sub-peak, A j G represents the area of ​​each sub-peak, reflecting the weight assigned to each sub-peak. j (logT2) is the Gaussian distribution of the sub-peaks, and m is the number of sub-peaks.

[0077] Using equations (1) to (4) and the properties of the Gaussian distribution, the sub-peak area (A) can be deduced. j ) and NMR porosity The following relationship must be satisfied:

[0078]

[0079] Where logT 2min and logT 2maxThese are the minimum and maximum values ​​of the logarithmic transverse relaxation time T2, respectively.

[0080] In engineering applications, the Gaussian distribution function is transformed as follows:

[0081]

[0082] Among them, w j It is the full width at half maximum (FWHM) of the sub-peak; c j It is the center of Zifeng.

[0083] Therefore, the properties of each sub-peak ( Figure 1 Figure B) can be obtained through specific parameters A j w j and c j To determine.

[0084] The R-squared value, which reflects the goodness of fit, can be calculated in the following way:

[0085]

[0086] Where N is the total number of data points in a single NMR T2 spectrum after Gaussian multi-peak fitting, and f k * (logT2) is the k-th data value of the porosity probability density function of the actual total peak, f k (logT2) is the k-th data value of the porosity probability density function of the fitted total peak. It is the average value of the porosity probability density function of the total peak after Gaussian multi-peak fitting.

[0087] 3. Examples

[0088] This patent example uses low-permeability sandstone samples from the Daluhu area of ​​the Dongying Depression in the Bohai Bay Basin and the Mahaidong area of ​​the Qaidam Basin for demonstration and illustration.

[0089] like Figure 2 As shown, taking a representative sample from the Daluhu area as an example, two sub-peaks were finally fitted, and the sum of their peak areas was approximately equal to the NMR porosity (16.7%), indicating that the peak decomposition was consistent with the actual porosity. After data preprocessing, porosity values ​​greater than 100ms were removed. Figure 2 Figure A shows the transformation of the original T2 spectrum into the log(T2) spectrum. Figure 3 Figure B shows the graph, where the horizontal axis represents the logarithm of the T2 relaxation time and the vertical axis represents the probability density function of porosity.

[0090] contrast Figure 2As shown in Fig. A and Fig. B in the drawings, the log(T2) spectrum keeps similar shape with the original T2 spectrum, which is due to the similar interval of data points in log(T2) spectrum, resulting in the consistent relative amplitude of data points after differentiation.

[0091] As shown in Fig. A and Fig. B in the drawings, the log(T2) spectrum keeps similar shape with the original T2 spectrum, which is due to the similar interval of data points in log(T2) spectrum, resulting in the consistent relative amplitude of data points after differentiation. Figure 2 As shown in Fig. B in the drawings, the fitted total peak has high coincidence with the actual data, and the R-square value in Table 2 is 0.997, which indicates that the fitted curve effectively reflects the characteristics of the original data. The total peak is composed of two linearly superimposed sub-peaks, each of which represents different pore types. Comparing these sub-peaks with the centrifugal porosity increment curve in Fig. A, it is found that peak 1 mainly represents the immobile water associated with micropores, and peak 2 mainly represents the mobile water associated with mesopores. Figure 2 As shown in Fig. A and Fig. B in the drawings, the log(T2) spectrum keeps similar shape with the original T2 spectrum, which is due to the similar interval of data points in log(T2) spectrum, resulting in the consistent relative amplitude of data points after differentiation.

[0092] And peaks 2 and 3 mainly represent the mobile water associated with mesopores and macropores

[0093] As shown in Fig. A and Fig. B in the drawings, the log(T2) spectrum keeps similar shape with the original T2 spectrum, which is due to the similar interval of data points in log(T2) spectrum, resulting in the consistent relative amplitude of data points after differentiation. Figure 3 As shown in Table 1, taking the representative sample in Mahaidong area as an example, three sub-peaks are finally fitted, and the sum of peak areas is approximately equal to the NMR porosity (21.11%). As shown in Fig. A and Fig. B in the drawings, the log(T2) spectrum keeps similar shape with the original T2 spectrum, which is due to the similar interval of data points in log(T2) spectrum, resulting in the consistent relative amplitude of data points after differentiation.

[0094] As shown in Fig. B in the drawings, the fitted total peak has high coincidence with the actual data, and the R-square value in Table 1 is 0.9975, which indicates that the fitted curve effectively reflects the characteristics of the original data. Unlike the sample in Daluhu area, the total peak of the sample in Mahaidong area is composed of three linearly superimposed sub-peaks. Comparing these sub-peaks with the centrifugal porosity increment curve in Fig. A, it is found that peak 1 mainly represents the immobile water associated with micropores, and peaks 2 and 3 mainly represent the mobile water associated with mesopores and macropores. Figure 3 As shown in Fig. A and Fig. B in the drawings, the log(T2) spectrum keeps similar shape with the original T2 spectrum, which is due to the similar interval of data points in log(T2) spectrum, resulting in the consistent relative amplitude of data points after differentiation. Figure 3 As shown in Fig. A and Fig. B in the drawings, the log(T2) spectrum keeps similar shape with the original T2 spectrum, which is due to the similar interval of data points in log(T2) spectrum, resulting in the consistent relative amplitude of data points after differentiation.

[0095] The above two examples show that the multi-peak fitting pore structure quantitative characterization method can be applied to the analysis of samples from different areas.

[0096] Table 1 Gaussian multi-peak fitting results of representative sample in Mahaidong area

[0097]

[0098]

[0099] 4. Explanation of the physical meaning of multi-peak parameters

[0100] As mentioned above, each sub-peak in the Gaussian distribution can be accurately quantified by three peak parameters A, w, c. The calculated results of the samples from Mahai East area are shown in Table 2.

[0101] Table 2 Data of NMR T2 spectrum of all samples in Mahai East area after Gaussian multi-peak fitting

[0102]

[0103]

[0104]

[0105] A is the area of the sub-peak, reflecting the porosity of each type of pore, the larger the area, the higher the porosity Figure 1 . According to equation (5), the sum of the areas of each sub-peak is approximately equal to the NMR porosity, and their difference depends on the accuracy of the Gaussian multi-peak fitting. From Table 1 and Fig. B in Figure 2 , it can be seen that the micro-pore, meso-pore and macro-pore porosities of the sample are 11.826%, 6.796% and 2.671% respectively, and their sum is 21.293%, which is close to the NMR porosity of 21.11%.

[0106] w is the full width at half maximum (FWHM) of different sub-peaks, indicating the width of the Gaussian distribution and the pore size range within each pore type Figure 1 . The narrower the sub-peak, the more concentrated the pore size distribution; the wider the sub-peak, the more dispersed the distribution.

[0107] c is the center of different sub-peaks, representing the central value of each pore type Figure 1 . The smaller the peak center, the smaller the pore of that type, and the larger the peak center, the larger the pore.

[0108] 5. Permeability prediction and application

[0109] There are several models for predicting permeability based on NMR T2 relaxation data, among which the most commonly used are Coates and SDR models.

[0110] In the Coates model, permeability depends on the T 2cutoff value, which separates the free fluid index FFI from the bound water volume BVI:

[0111]

[0112] where K is the permeability (mD), NMR effective porosity (%), C is a formation-specific constant that reflects the correlation between rock pore throat and pore size. FFI is the fraction of formation volume occupied by free flowing fluids, and BVI is the fraction of formation volume occupied by immobile capillary bound fluids.

[0113] In the SDR model, permeability depends on T 2gm values:

[0114]

[0115] where T 2gm is the geometric mean of the T2 distribution, and a is a coefficient that depends on the formation type.

[0116] The key in the Coates model is to obtain FFI and BVI using T 2cutoff values. Typically, T 2cutoff is determined by comparing NMR T2 spectra of centrifuged and saturated samples (Figure A in Figure 1 ). In the absence of centrifuge experimental data, T 2cutoff is used. This approach has limitations when centrifuge data are lacking or when there are significant differences in formation lithology.

[0117] In the SDR model, the key is the calculation of T 2gm , which is given by:

[0118]

[0119] Taking the logarithm of both sides of equation (10) simplifies to:

[0120]

[0121] According to equation (5), the log(T 2gm ) calculation can be expressed by the porosity probability density function of the total number of peaks obtained by the Gaussian multi-peak fitting:

[0122]

[0123] Equation (12) shows that log(T 2gm ) is approximately equal to the expectation of the Gaussian function f(log T2) divided by the sum of the peak areas.

[0124] From the properties of the Gaussian function, the expectation and variance of a linear combination of random variables are functions of the means and variances of the individual variables. This means that the expectation of the Gaussian multi-peak fitting function is a weighted average of the expectations of each sub-peak, with the weights being the peak areas A. Therefore, the following equation is satisfied:

[0125]

[0126] In this study, logT' 2gm It is redefined as the log-geometric mean of the peak parameters based on Gaussian multi-peak fitting, to distinguish it from the actual log-geometric mean:

[0127]

[0128] Table 2 lists all sample log(T) 2gm ) and logT' 2gm The calculation results show a strong linear correlation between the two. Figure 4 ), R 2 The value is 0.9999. This difference is mainly attributed to the error in Gaussian multi-peak fitting and computational factors such as the spacing of data points.

[0129] Based on the above analysis, log(T) 2gm The logT' can be calculated using the peak parameters after Gaussian multi-peak fitting. 2gm To approximate it, we use logT'. 2gm Applying this to the SDR model, the permeability equation becomes:

[0130]

[0131] Equation (15) shows that permeability is closely related to peak parameters A and c. Studies have shown that micropores, represented by bound water, contribute the least to permeability, while permeability is mainly affected by mesopores and macropores, represented by movable water.

[0132] Based on this, the SDR model was modified to exclude peak parameters related to micropores, and the calculations are as follows:

[0133]

[0134] Where a, b, and c are constants.

[0135] The calculation results of the model are as follows Figure 5 As shown.

[0136] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered illustrative and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for quantitative characterization of pore structure in low-permeability reservoirs based on multi-peak fitting of nuclear magnetic resonance T2 spectra, characterized in that, The nuclear magnetic resonance T2 spectrum was preprocessed, and the preprocessed data was decomposed by Gaussian multi-peak fitting. The porosity and pore size distribution of micropores, mesopores and macropores were quantitatively characterized by combining the peak area, peak center and peak width of each sub-peak. By incorporating the sub-peak parameters obtained from multi-peak fitting into the permeability prediction model, and based on the multi-peak fitting results, the permeability prediction model based on NMRT2 relaxation data is modified to construct a permeability model; the nuclear magnetic resonance T2 spectrum is preprocessed to obtain the nuclear magnetic porosity. In the penetration rate prediction model, T is calculated. 2gm : It is the i-th component T of the transverse relaxation time T2 2i Porosity, where n is the total number of components; Taking the logarithm of both sides simplifies to: log(T 2gm The porosity probability density function is represented by the total number of peaks obtained through Gaussian multi-peak fitting: logT 2min and logT 2max These are the minimum and maximum values ​​of the logarithmic transverse relaxation time T2, respectively, f * (logT2) is the actual porosity probability density function, c j It is the center of the j-th sub-peak, A j is the area of ​​the j-th sub-peak, m is the number of sub-peaks, and f(logT2) is the porosity probability density function of the total peaks; logT' 2gm Redefining it as the log-geometric mean of the peak parameters based on Gaussian multi-peak fitting:

2. The method for quantitative characterization of pore structure in low-permeability reservoirs based on multi-peak fitting of nuclear magnetic resonance T2 spectra according to claim 1, characterized in that, Preprocessing of the nuclear magnetic resonance T2 spectrum yields the nuclear magnetic porosity. in, It is the i-th component T of the transverse relaxation time T2 2i Porosity, where n is the total number of components; The actual cumulative distribution function F of nuclear magnetic resonance porosity * for: Actual porosity probability density function f * (logT2) is:

3. The method for quantitative characterization of pore structure in low-permeability reservoirs based on multi-peak fitting of nuclear magnetic resonance T2 spectra according to claim 2, characterized in that, Suppose that the preprocessed data distribution is formed by the linear superposition of multiple Gaussian distributions, and the porosity probability density function f(logT2) of the total peak is: Where y0 is the baseline, f j (logT2) is the porosity probability density function of the sub-peak, A j G is the area of ​​the j-th sub-peak. j (logT2) is the Gaussian distribution of the j-th sub-peak, and m is the number of sub-peaks; The area A of the j-th sub-peak j and NMR porosity The following relationship must be satisfied: Among them, logT 2min and logT 2max These are the minimum and maximum values ​​of the logarithmic transverse relaxation time T2; the Gaussian distribution function is transformed as follows: Among them, w j It is the full width at half maximum (FWHM) of the j-th sub-peak; c j It is the center of the j-th sub-peak.

4. The method for quantitative characterization of pore structure in low-permeability reservoirs based on multi-peak fitting of nuclear magnetic resonance T2 spectra according to claim 3, characterized in that, The R-squared value, which reflects the goodness of fit, is calculated as follows: Where N is the total number of data points in a single NMR T2 spectrum after Gaussian multi-peak fitting, and f k * (logT2) is the k-th data value of the porosity probability density function of the actual total peak, f k (logT2) is the k-th data value of the porosity probability density function of the fitted total peak. It is the average value of the porosity probability density function of the total peak after Gaussian multi-peak fitting.

5. The method for quantitative characterization of pore structure in low-permeability reservoirs based on multi-peak fitting of nuclear magnetic resonance T2 spectra according to claim 1, characterized in that, log(T 2gm The log(T') calculated using the peak parameters after Gaussian multi-peak fitting 2gm To approximate, log(T') 2gm When applied to the permeability prediction model, the permeability K equation becomes: The permeability prediction model was modified to exclude peak parameters related to micropores, and the calculations are as follows: Where a, b, and c are constants.

Images