A thin reservoir prediction method based on pre-stack frequency extension processing

Abstract

This invention discloses a thin reservoir prediction method based on pre-stack frequency extension processing. The technical solution is based on improving the resolution of seismic data processing, conducting reservoir petrophysical analysis using measured shear wave data, constructing a petrophysical model, clarifying sensitive parameters for thin reservoir identification, and bridging the gap between well logging and seismic data. For seismic data, pre-stack seismic deterministic inversion of sand group thickness resolution and pre-stack statistical high-resolution inversion of single sand body thickness are performed to calculate sensitive parameter data volumes. Based on petrophysical quantities, the spatial characterization accuracy of thin sand bodies is effectively improved. This invention is highly operable and easy to apply, showing good application prospects for ultra-deep and ultra-thin sand bodies similar to those in the Cretaceous and Triassic strata of the Tarim Basin's Tahe Oilfield. It is also applicable to the identification of sweet spots and fine spatial characterization of ultra-deep thin reservoirs in the central and western basins, which is of great significance for improving drilling success rates and production capacity construction.

Description

Technical Field

[0001] This invention relates to the field of thin reservoir prediction, and more specifically, to a method for thin reservoir prediction based on pre-stack frequency extension processing. Background Technology

[0002] Thin layers are generally defined as strata smaller than the seismic resolution limit. With the continuous advancement of exploration and development, thin layers are receiving increasing attention. As reservoirs, they contain substantial oil and gas resources; as mudstone caprocks, they control the migration and enrichment of oil and gas under certain geological conditions. Given the characteristics of thin layers—thinness, large variations, poor lateral continuity, and strong heterogeneity—improving the effectiveness of thin layer identification, tracking, and evaluation is currently a research hotspot.

[0003] Many scholars both domestically and internationally have conducted extensive research on thin reservoir prediction methods from both qualitative and quantitative perspectives. Qualitative methods primarily rely on geological understanding, comprehensively utilizing seismic multi-attribute data, seismic facies classification, and well logging data to describe reservoir distribution patterns and characteristics. Combined with actual drilling information, favorable reservoir distribution ranges are statistically determined. These methods are currently relatively mature and have shown significant application results (Yang Zhanlong et al. (2019)). For example, in the Songliao Basin, using "two-width, one-height" seismic data and the root mean square amplitude attribute to predict thin sand bodies significantly improved the accuracy rate (Nietal, 2018); in the Junggar Basin, the western slope of Mahu Lake near Karamay... Groups have used minimum interferometric frequency data to predict the boundaries of single thin layers relatively well (Liu Huaqing et al., 2018). However, these methods have limited accuracy in reservoir characterization and are difficult to predict the thickness of individual sand bodies in thin interbedded sand bodies, which cannot meet the needs of high-precision rolling development and production (Widess, 2020; Yuan Cheng et al., 2021). Quantitatively, methods such as pre-stack and post-stack inversion are mainly used, along with rock physical analysis scales, to quantitatively describe reservoir thickness, porosity, and spatial distribution through geostatistical methods (Huang Xude, 1994).

[0004] The Carboniferous reservoirs in the Tarim Basin are mainly composed of CK1 stratigraphic traps and CK5 lithological traps. Multiple well logging, testing, and production data indicate that the Carboniferous system has good evaluation and development potential. The main problems in thin reservoir prediction in this study area are as follows: First, the low resolution of seismic data makes it difficult to meet the required reservoir prediction accuracy; second, the presence of updip pinch-out strata traps in the area makes reservoir characterization difficult due to strong T50 reflections; third, the reservoirs are deep, thin, and exhibit rapid lateral variations, with overlapping P-wave impedances of sandstone and mudstone, making quantitative reservoir prediction difficult and hindering a clear understanding of favorable reservoir distribution characteristics. Due to limitations in vertical seismic resolution and strong interference from adjacent layers, seismic attributes in thin interbedded sandstone and mudstone strata cannot accurately reflect the true changes in the reservoirs, making it difficult to achieve a detailed reservoir description.

[0005] For the reasons mentioned above, this invention aims to provide a thin reservoir prediction method based on pre-stack frequency extension processing. Based on compressed sensing frequency extension processing, rock physical analysis and modeling, geostatistical inversion, and pre-stack simultaneous inversion technology, this method improves the original seismic data and enables subsequent reservoir prediction and analysis. It solves the problem of low resolution of the original seismic data and difficulty in accurately predicting thin layers, thereby improving the accuracy and precision of ultra-thin reservoir prediction. Summary of the Invention

[0006] To address the shortcomings of existing technologies, the present invention aims to provide a thin reservoir prediction method based on pre-stack frequency extension processing. This method is based on improving the resolution of seismic data, conducting reservoir petrophysical analysis using measured shear wave data, constructing a petrophysical model, identifying sensitive parameters for thin reservoir identification, and bridging the gap between well logging and seismic data. For the seismic data, pre-stack seismic deterministic inversion of sand group thickness resolution and pre-stack statistical high-resolution inversion of single sand body thickness are performed respectively, calculating sensitive parameter data volumes. Based on petrophysical quantities, this effectively improves the spatial characterization accuracy of thin sand bodies.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a thin reservoir prediction method based on pre-stack frequency extension processing, comprising the following steps:

[0008] S1. Based on compressed sensing algorithm, frequency conversion processing of seismic physical exploration data is used to improve the resolution of geophysical exploration data.

[0009] S2. Rock Physics Modeling: Based on the data processed in S1, a rock physics model is established and rock physics parameters are assigned. Reservoir physics and rock physics analysis are performed throughout the well based on the rock physics model. Based on the analysis results, several elastic parameters sensitive to lithology and fluid type are determined. The elastic parameters include P-wave velocity Vp, S-wave velocity Vs, P-wave impedance Pimp, S-wave impedance Simp, acoustic resistance AI, shear stiffness Si, Poisson's ratio Pr, P-wave / S-wave velocity ratio Vp / Vs, bulk modulus λρ, shear modulus uρ, bulk elastic coefficient K, and shear modulus G.

[0010] S3. Pre-stack deterministic inversion: Based on the rock physics modeling and the data processed in S1, a simultaneous pre-stack inversion is performed to solve for each elastic parameter. Based on the inversion results, the quantitative cutoff values ​​of each elastic parameter are interpreted according to the rock physics interpretation scale to directly identify the reservoir structure and possible fluid interfaces.

[0011] S4. For reservoirs with a thickness greater than or equal to 10m, quantitative description is performed based on the interpretation in S3. For reservoirs with a thickness less than 10m, S5 is performed.

[0012] S5. Pre-stack high-resolution statistical inversion: Using the data processed in S1, Markov chain Monte Carlo algorithm inversion is performed on each elastic parameter; following the constrained sparse pulse inversion workflow, the data processed in S1 is integrated and quality monitored, and the constrained sparse pulse inversion is used as the prior distribution for Markov chain Monte Carlo algorithm inversion. Based on the Markov chain Monte Carlo algorithm inversion results, combined with S3, the reservoir with a thickness of less than 10m is interpreted and quantitatively described.

[0013] The present invention is further configured such that: before inversion in S5, discrete attribute lithofacies simulation test is performed on the lithofacies data in the seismic data by SIS sequence indication simulation, and continuous attribute simulation test is performed on the elastic parameter data in the seismic data by SGS sequence Gaussian simulation. The simulation test results are spatially distributed and quality controlled according to geological information to improve the spatial definition accuracy of statistical inversion parameters.

[0014] The present invention is further configured such that: the seismic physical exploration data in S1 includes calibrated well logging data, synthesized shear wave data, stratigraphic data of the target segment and seismic data, wherein the seismic data is either fully stacked seismic data and co-angularly stacked seismic data or fully stacked seismic data and co-offset stacked seismic data.

[0015] The present invention is further configured such that: before the compressed sensing algorithm is frequency-spreading, it is necessary to integrate and monitor the quality of seismic data, well logging data and stratigraphic data in accordance with the constrained sparse pulse inversion workflow.

[0016] The present invention is further configured as follows: pre-stack simultaneous inversion requires inversion of co-angle stacked seismic data or co-offset stacked seismic data based on the Zoppritz equation and its approximation within a multi-angle range to generate P-wave and S-wave impedance data volumes; based on the P-wave and S-wave impedance data volumes, Vp, Vs, Pimp, Simp, AI, Si, Pr, Vp / Vs, λρ, uρ, K, and G are obtained.

[0017] The present invention is further configured such that: before establishing the rock physics model, the well logging data needs to be quality controlled through single-well curve quality control, environmental correction and multi-well consistency processing.

[0018] The present invention is further configured such that: after the quality control of well logging data is completed, reservoir parameters need to be evaluated through the reservoir parameter interpretation model, the content of each mineral is calculated, and the content is input into the optimization interpretation model for optimization solution, so as to provide high-precision lithological and fluid characteristics for the rock physics model.

[0019] The present invention is further configured such that, for Carboniferous reservoirs, before performing S1 frequency conversion processing on seismic physical exploration data, the Xu and White model is used to predict shear waves and fit shear waves from multiple wells.

[0020] The present invention is further configured such that: the elastic parameter in S2 that is sensitive to sand and mudstone lithology is Vp / Vs, and the elastic parameter that is sensitive to fluid type is Pimp.

[0021] The present invention is further configured such that: the sandstone has a Vp / Vs ratio < 1.75; and the reservoir has a porosity greater than 8% and a Pimp ratio < 1.05 * 10⁻⁶. 7 The Pimp of the oil and gas reservoir is 8.8*10. 6 <Pimp<1.1*10 7 The Vp / Vs ratio of the oil and gas reservoir is 1.5. <Vp / Vs<1.75。

[0022] In summary, this invention offers the following advantages over existing technologies: Based on improved resolution processing of seismic data, it conducts reservoir petrophysical analysis using measured shear wave data, constructs a petrophysical model, clarifies sensitive parameters for thin reservoir identification, and bridges the gap between well logging and seismic analysis. For seismic data, it performs pre-stack seismic deterministic inversion of sand group thickness resolution and pre-stack statistical high-resolution inversion of single sand body thickness, calculating sensitive parameter data volumes. Based on petrophysical metrics, it effectively improves the spatial characterization accuracy of thin sand bodies. This invention is highly operable and easy to apply, showing promising prospects for ultra-deep and ultra-thin sand bodies similar to those in the Cretaceous and Triassic strata of the Tarim Basin's Tahe Oilfield. It is also suitable for identifying sweet spots and finely characterizing reservoir space in ultra-deep and thin reservoirs in central and western basins, significantly contributing to improved drilling success rates and production capacity development. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating an embodiment;

[0024] Figure 2 A flowchart illustrating the compressed sensing topology processing.

[0025] Figure 3 Calibrate the original PSTM (pre-stack time-shifted imaging);

[0026] Figure 4 For PSTM calibration after compressed sensing frequency extension;

[0027] Figure 5 A schematic diagram of the rock physics modeling process;

[0028] Figure 6 This is a schematic diagram of the pre-stack deterministic inversion process;

[0029] Figure 7 Pimp and Vp / Vs intersection plot;

[0030] Figure 8 This is a schematic diagram of the statistical inversion process for high-scoring pre-stack data.

[0031] Figure 9 For CK1, the probability density function is given by different lithological wave impedances, Vp / Vs, and densities.

[0032] Figure 10 The graph shows the variation function analysis of the target fault.

[0033] Figure 11 This is a comprehensive evaluation chart of statistical inversion profile (middle) + TKC1-4H logging.

[0034] Figure 12 This is the daily production curve for a single well, TKC1-4H. Detailed Implementation

[0035] The technical solution of the present invention will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are not all embodiments of the present invention. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the invention.

[0036] Example

[0037] like Figure 1 The diagram shown is a flowchart of a preferred embodiment of the present invention. A thin reservoir prediction method based on pre-stack frequency extension processing includes the following steps:

[0038] S1. Based on compressed sensing algorithm, frequency conversion processing of seismic physical exploration data is used to improve the resolution of geophysical exploration data.

[0039] The seismic physical exploration data includes calibrated well logging data, synthesized shear wave data, stratigraphic data of the target segment, and seismic data. The seismic data consists of either fully stacked seismic data and co-angularly stacked seismic data, or fully stacked seismic data and co-offset stacked seismic data. Before frequency overlay, the seismic data, well logging data, and stratigraphic data need to be integrated and their quality monitored according to the constrained sparse pulse inversion workflow.

[0040] like Figure 2-4 As shown, when the compressed sensing algorithm performs frequency upscaling, it first needs to extract wavelets from the original seismic record. The extracted wavelets and the original seismic record are then subjected to Fourier transforms to obtain the corresponding frequency domain data. Then, the reflection coefficient in the frequency domain is calculated using a seismic convolution model. The reflection coefficient is decomposed into even and odd components, and different weights are assigned to the even and odd components. A sparse constraint on the reflection coefficient is then added, and the objective function of the reflection coefficient is solved to obtain the even and odd components of the time domain reflection coefficient. Finally, the time domain reflection coefficient is obtained by reconstructing the even and odd components. The reflection coefficient is then convolved with the wavelet to obtain a high-resolution seismic record.

[0041] S2, Rock Physical Modeling:

[0042] like Figure 5 As shown, based on the S1 processing, well logging data undergoes quality control through single-well curve quality control, environmental correction, and multi-well consistency processing to provide highly consistent and relatively complete well logging data. Then, reservoir parameters are evaluated using a reservoir parameter interpretation model, calculating the content of each mineral and inputting it into the optimization interpretation model for optimization, providing high-precision lithological and fluid characteristics for the rock physical model. Finally, a rock physical model is established and assigned rock physical parameters.

[0043] Based on the rock physics model, reservoir physics and rock physics analysis are performed across the entire well section. According to the analysis results, several elastic parameters sensitive to lithology and fluid type are determined. These elastic parameters include P-wave velocity (Vp), S-wave velocity (Vs), P-wave impedance (Pimp), S-wave impedance (Simp), acoustic resistance (AI), shear stiffness (Si), Poisson's ratio (Pr), P-wave / S-wave velocity ratio (Vp / Vs), bulk modulus (λρ), shear modulus (uρ), bulk elastic coefficient (K), and shear modulus (G). Rock physics analysis serves as a bridge between well logging and seismic analysis. Based on rock physics analysis, rock physics elastic parameters sensitive to the rock skeleton and fluids within the rock can be determined, establishing the relationship between reservoir properties and seismic properties (well-seismic relationship). By perturbing corresponding reservoir parameters, the variation patterns of elastic parameters are analyzed, enabling the prediction of reservoir elastic parameter characteristics from a point-to-surface perspective.

[0044] Cross-analysis of different elastic parameters shows that the combined pre-stack inversion of Vp / Vs and Pimp can distinguish the lithology, physical properties, and oil-bearing reservoirs of the target layer in the study area. The smaller the overlap between elastic parameters, the higher the reliability of the prediction, the more concentrated the data, and the easier it is to determine the lithological or physical property boundaries.

[0045] S3, Pre-stack Deterministic Inversion:

[0046] like Figure 6 As shown, pre-stack simultaneous inversion is performed based on rock physics modeling and data processed in S1. Pre-stack simultaneous inversion requires inverting co-angular stacked or co-offset stacked seismic data within a multi-angle range using the Zoppritz equation and its approximations to generate P-wave and S-wave impedance data volumes. Based on these data volumes, Vp, Vs, Pimp, Simp, AI, Si, Pr, Vp / Vs, λρ, uρ, K, and G are calculated to quantitatively describe the distribution of oil and gas-bearing sand bodies in the superimposed geological bodies, laying the foundation for the inversion in S5. Once the simultaneous inversion is complete, all elastic parameter data volumes can be verified using actual well logging data.

[0047] Specifically, density inversion requires the use of at least five angles of co-angle superposition data volumes. That is, a relatively reliable density data volume can only be generated when the far incident angle of the reservoir section is at least 40 degrees.

[0048] Based on the above inversion results, the quantitative cutoff values ​​of each elastic parameter are interpreted according to the rock physics interpretation scale to directly identify the reservoir structure and possible fluid interfaces; such as Figure 7 As shown, the quantitative cutoff values ​​for some elastic parameters in this embodiment are interpreted as follows: sandstone Vp / Vs < 1.75; reservoirs with porosity greater than 8% Pimp < 1.05 * 10⁻⁶. 7 The Pimp of the oil and gas reservoir is,

[0049] 8.8*10 6 <Pimp<1.1*10 7 The Vp / Vs ratio of the oil and gas reservoir is 1.5. <Vp / Vs<1.75。

[0050] S4. Based on the explanation in S3, a quantitative description is given for reservoirs with a thickness greater than or equal to 10m, and S5 is given for reservoirs with a thickness less than 10m.

[0051] S5. Pre-stack high-score statistical inversion:

[0052] like Figure 8-10 As shown in the figure, this embodiment analyzes the lithology (discrete properties), Pimp, probability density function and variogram of Vp / Vs of the target layer CK1-CK5 in the geological model.

[0053] Based on the S1 data processing, the lithological proportions and vertical and lateral variability functions of each sublayer in the target interval were first determined using SIS sequence random indicator simulation. Then, Gaussian simulation using SGS sequence was used to obtain the probability density functions and spatial variability functions of the lithofacies elastic parameters, providing reliable probability density functions and variability functions for subsequent lithological and elastic parameters. This embodiment employs a method combining Markov chain Monte Carlo algorithm inversion. By establishing a probabilistic model of reservoir attributes and outputting simulation results, multiple implementations are generated, realizing the output of a thin sandstone body model. Furthermore, combined with S3 interpretation, the already defined sandstone body model is optimized to solve the problem of quantitative prediction of thin reservoirs <10m.

[0054] Specifically, in this embodiment, the data processed by S1 was integrated and its quality monitored according to the constrained sparse pulse inversion workflow. The constrained sparse pulse inversion was used as the prior distribution for the Markov chain Monte Carlo algorithm inversion. By using the constrained sparse pulse inversion, the high horizontal resolution and predictability of the inversion results were effectively controlled by the seismic data.

[0055] Specifically, for Carboniferous reservoirs, since there is only one shear wave data point, TK157, the Xu and White model needs to be used to predict shear waves before S1 frequency extension, and fit shear waves from multiple wells for subsequent inversion constraints.

[0056] The Xu and White model is based on A mixed sandstone-mudstone model was proposed based on theory, the Gassmann equation, and the effective differential medium (DEM). This model attributes the variation of P-wave velocity with clay content to the differences in pore geometry and pore flattening between mudstone and sandstone, and can better reflect the change in P-wave velocity with increasing clay content. Based on this model, S-wave calculations first estimate the pore volume of sandstone and mudstone. Using the Wyllie time-averaging equation, the P-wave and S-wave transit times of the rock matrix in the mixed model are calculated, and then the bulk modulus and shear modulus of the rock matrix are calculated. The equations can be used to calculate the bulk modulus and shear modulus of dry rock skeletons; the Gassmann equations can be used to calculate the bulk modulus and shear modulus of fluid-saturated rocks, and the P-wave and S-wave velocities can be obtained from these modulus parameters.

[0057] Taking the TK157 well area as an example, this well area is located in the central part of the Tarim Basin 1. The target layer is the Carboniferous Karasai Formation CK1-3# single sand body with an oil layer thickness of 6m and a burial depth of >5000m, which is a structural-lithological composite edge water reservoir. Due to the impact of reservoir prediction accuracy, the reservoir encounter rate of the previously deployed TK151X development well was only 30%.

[0058] like Figure 11-12 As shown, in 2021, this method achieved excellent results in reservoir prediction and horizontal well deployment in the Tarim Basin Carboniferous system. It effectively identified 5-10m sand bodies, improving the sand body matching rate by 10%. Through stereoscopic evaluation, it optimized sweet spots for reservoir inversion and deployed five long horizontal wells: TKC1-4H to TKC1-8H and TKC1-10H. The TKC1-4 well, already in production, has a horizontal section drilled to a length of 539m, encountering 378m of sand layers and 309m of oil and gas layers. The long horizontal well, guided by reservoir prediction, achieved a sand drilling rate of 70%, with a perforation depth of 8m, and maintained an average daily production of 25 tons of oil and 47,000 cubic meters of gas.

[0059] In summary, this embodiment, based on improved resolution processing of seismic data, conducts reservoir petrophysical analysis using measured shear wave data, constructs a petrophysical model, clarifies sensitive parameters for thin reservoir identification, and bridges the gap between well logging and seismic analysis. For the seismic data, pre-stack seismic deterministic inversion of sand group thickness resolution and pre-stack statistical high-resolution inversion of single sand body thickness are performed to calculate sensitive parameter data volumes. Based on petrophysical metrics, the spatial characterization accuracy of thin sand bodies is effectively improved. This method is highly operable and easy to apply, showing promising prospects for ultra-deep and ultra-thin sand bodies similar to those in the Cretaceous and Triassic strata of the Tarim Basin's Tahe Oilfield. It is also suitable for identifying sweet spots and finely characterizing reservoir space in ultra-deep and thin reservoirs in central and western basins, which is of great significance for improving drilling success rates and production capacity development.

[0060] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for predicting thin reservoirs based on pre-stack frequency extension processing, characterized in that: Includes the following steps: S1. Based on compressed sensing algorithm, the seismic physical exploration data is processed by frequency extension to improve the resolution of the geophysical exploration data. The seismic physical exploration data in S1 includes calibrated well logging data, synthetic shear wave data, target layer stratigraphic data and seismic data, of which the seismic data are either fully stacked seismic data and co-angular stacked seismic data or fully stacked seismic data and co-offset stacked seismic data. S2. Rock Physics Modeling: Based on the data processed in S1, a rock physics model is established and assigned rock physics parameters; reservoir physics and rock physics analysis are performed throughout the well based on the rock physics model; based on the analysis results, several elastic parameters sensitive to lithology and fluid type are determined; the elastic parameters include P-wave velocity Vp, S-wave velocity Vs, P-wave impedance Pimp, S-wave impedance Simp, acoustic resistance AI, shear stiffness Si, Poisson's ratio Pr, P-wave / S-wave velocity ratio Vp / Vs, bulk modulus λρ, shear modulus uρ, bulk elastic coefficient K, and shear modulus G; S3. Pre-stack deterministic inversion: Based on the rock physics modeling and the data processed in S1, pre-stack simultaneous inversion is performed to solve for each elastic parameter; based on the inversion results, the quantitative cutoff values ​​of each elastic parameter are interpreted according to the rock physics interpretation scale to directly identify the reservoir structure and possible fluid interfaces. S4. For reservoirs with a thickness greater than or equal to 10m, quantitative description shall be performed based on the interpretation in S3; for reservoirs with a thickness less than 10m, S5 shall be performed. S5. Pre-stack high-resolution statistical inversion: Using the data processed in S1, Markov chain Monte Carlo algorithm inversion is performed on each elastic parameter; following the constrained sparse pulse inversion workflow, the data processed in S1 is integrated and quality monitored, and the constrained sparse pulse inversion is used as the prior distribution for Markov chain Monte Carlo algorithm inversion. Based on the Markov chain Monte Carlo algorithm inversion results, combined with S3, the reservoir with a thickness of less than 10m is interpreted and quantitatively described.

2. The thin reservoir prediction method based on pre-stack frequency extension processing according to claim 1, characterized in that: Before S5 inversion, discrete attribute lithofacies simulation tests are performed on the lithofacies data in the seismic data using SIS sequence indicator simulation, and continuous attribute simulation tests are performed on the elastic parameter data in the seismic data using SGS sequence Gaussian simulation. The simulation test results are spatially distributed and quality controlled based on geological information to improve the spatial definition accuracy of statistical inversion parameters.

3. The thin reservoir prediction method based on pre-stack frequency extension processing according to claim 1, characterized in that: Before the compressed sensing algorithm is frequency-spreading, it is necessary to integrate and monitor the quality of seismic data, well logging data, and stratigraphic data according to the constrained sparse pulse inversion workflow.

4. The thin reservoir prediction method based on pre-stack frequency extension processing according to claim 1, characterized in that: Pre-stack simultaneous inversion requires inverting co-angular stacked seismic data or co-offset stacked seismic data within multiple angle ranges based on the Zoppritz equation and its approximations to generate P-wave and S-wave impedance data volumes; and then obtaining Vp, Vs, Pimp, Simp, AI, Si, Pr, Vp / Vs, λρ, uρ, K, and G based on the P-wave and S-wave impedance data volumes.

5. The thin reservoir prediction method based on pre-stack frequency extension processing according to claim 1, characterized in that: Before establishing a rock physics model, well logging data needs to be quality controlled through single-well curve quality control, environmental correction, and multi-well consistency processing.

6. The thin reservoir prediction method based on pre-stack frequency extension processing according to claim 5, characterized in that: After the well logging data quality control is completed, reservoir parameters need to be evaluated through the reservoir parameter interpretation model. The content of each mineral is calculated and then fed into the optimization interpretation model for optimization and solution, providing high-precision lithological and fluid characteristics for the rock physics model.

7. The thin reservoir prediction method based on pre-stack frequency extension processing according to claim 1, characterized in that: For Carboniferous reservoirs, before processing seismic physical exploration data using S1 topology, it is necessary to use the Xu and White model to predict shear waves and fit shear waves from multiple wells. The Xu and White model is a sandstone-mudstone hybrid model that includes Kuster-Toksöz theory, Gassmann equations, and effective differential media.

8. A thin reservoir prediction method based on pre-stack frequency extension processing according to claim 7, characterized in that: In S2, the elastic parameter sensitive to sandstone and mudstone lithology is Vp / Vs, and the elastic parameter sensitive to fluid type is Pimp.

9. A thin reservoir prediction method based on pre-stack frequency extension processing according to claim 8, characterized in that: For sandstone reservoirs, Vp / Vs < 1.75; for reservoirs with porosity greater than 8%, Pimp < 1.05 × 10⁻⁶. 7 The Pimp of the oil and gas reservoir is 8.8*10. 6 < Pimp < 1.1 * 10 7 The Vp / Vs ratio of the oil and gas reservoir is 1.

5. <Vp / Vs<1.75。

Images