Quantitative characterization method for carbonate rock fault karst reservoir parameters based on extended elastic impedance inversion

Abstract

The invention provides a carbonate rock fault-karst reservoir parameter quantitative characterization method based on extended elastic impedance inversion, which comprises the following steps: determining a Chi projection angle which is highly related to reservoir parameters through logging rock physical analysis, AVO attribute analysis, elastic parameter estimation, extended elastic impedance curve drawing, curve intersection analysis and Chi projection angle scanning; through combination of prestack seismic synchronous inversion and random inversion, longitudinal wave acoustic impedance, transverse wave acoustic impedance and a density attribute body with high resolution are obtained; carrying out neural network clustering analysis by utilizing longitudinal wave acoustic impedance, transverse wave acoustic impedance and a density attribute body generated by acoustic impedance inversion, and judging lithology; reservoir parameters and the extended elastic impedance data body are subjected to intersection analysis, finally, reservoir parameters such as reservoir porosity and water saturation are quantitatively predicted, and the method has important practical significance for revealing favorable reservoir spatial distribution characteristics and oil reservoir enrichment rules, further optimizing exploration targets for fault-solution oil reservoirs and reasonably formulating development schemes.

Description

Technical Field

[0001] The present invention relates to the field of oil exploration and development, and particularly relates to a method for quantitatively characterizing reservoir parameters of carbonate fracture-vug bodies based on extended elastic impedance inversion. Background Art

[0002] Fracture-vug bodies are a special type of carbonate fracture-cavity reservoir formed by fault-controlled karstification, and are important targets for increasing reserves and production in the Tahe Oilfield in recent years. Such reservoirs are buried deep, have experienced multiple tectonic movements and karstification, and have diverse reservoir space types, including caves, pores, fractures and their combinations. They have large scale differences, extremely irregular shapes, and extremely strong heterogeneity. Affected by complex geological conditions, during the production process of fracture-vug body oil reservoirs at present, reservoir flushing and leakage are serious, and effective coring and geophysical logging cannot be carried out in most reservoir intervals. It is difficult to accurately characterize reservoir parameters through laboratory measurements, conventional logging data interpretation, and post-stack seismic data means, resulting in large deviations in reserve calculation results, which seriously restricts the further selection of exploration targets and the rational formulation of development plans. Therefore, for such oil reservoirs, inventing a reliable and accurate method to establish reservoir parameters and quantitatively describe the spatial distribution of the oil reservoir, the physical properties of the reservoir, and the variation law of oil-bearing properties has become an urgent problem to be solved for the effective development of fracture-vug body oil reservoirs.

[0003] The Chinese patent "A Characterization Method and System for Fracture-Vug Type Carbonate Reservoirs" with the patent number CN202010860655.1 proposes a characterization method for fracture-vug type carbonate reservoirs, including: determining the configuration of the carbonate reservoir of the fracture-vug body to be evaluated based on different stages of karstification by analyzing the geological origin process of the fracture-vug body and combining the fracture development scale and the formation mechanism of dissolution filling; based on the three-dimensional seismic model and / or logging interpretation data, carrying out fracture-vug body characterization analysis step by step according to the stage of the configuration to obtain the characterization model of each configuration; and performing fusion processing on all characterization models to form the corresponding three-dimensional spatial genetic body interpretation result. The present invention realizes the three-dimensional spatial genetic body interpretation of fracture-vug geological bodies and conducts targeted characterization according to the scale and origin of different configuration units, and has strong applicability for the fine description of fracture-vug bodies.

[0004] The Chinese patent "Method and Device for Analyzing Physical Property Parameters of Beaded Fracture-Vuggy Carbonate Reservoirs" with the patent number CN202011376605.2 discloses a method and device for analyzing physical property parameters of beaded fracture-vuggy carbonate reservoirs. It details the numerical inversion after establishing a series-connected mathematical model of one hole and one fracture in the beaded fracture-vuggy reservoir and obtaining the Laplace solution of the dimensionless production rate, thereby obtaining the dimensionless production rate, dimensionless production rate integral, etc., and plotting the curve chart for analyzing the instability of well production. By fitting the curve chart for analyzing the instability of well production and the normalized curve of well production, the actual physical property parameters of the target beaded fracture-vuggy carbonate reservoir are determined. This patent does not mention the method for determining the pseudo relative permeability of fracture-vuggy reservoirs. In oil and gas exploration, reservoir description, and comprehensive research of oil and gas reservoirs, accurately obtaining reservoir parameters is a key technology. Since the 1960s, with the improvement of understanding and the update of technology, a series of methods for obtaining reservoir parameters have been developed. These methods mainly include: laboratory measurement, well logging data interpretation, and combined calculation of well logging and seismic data. However, in the production process of the fault-karst reservoirs in the Tahe Oilfield at present, the reservoir is severely depleted and lost, and it is difficult to effectively core and conduct geophysical logging in most reservoir intervals. It is difficult to accurately characterize reservoir parameters through laboratory measurement, conventional well logging data interpretation, and post-stack seismic data, resulting in a large deviation in the calculated reserves, which seriously restricts the further optimization of exploration targets and the rational formulation of development plans. Therefore, for such reservoirs, inventing a reliable and accurate method to determine reservoir parameters and quantitatively describe the spatial distribution of the reservoir, the physical properties of the reservoir, and the variation law of oil-bearing properties has become an urgent problem to be solved for the effective development of fault-karst reservoirs. Summary of the Invention

[0005] In order to solve the problem of how to obtain reliable and accurate reservoir parameters, the present invention proposes a method for quantitatively characterizing carbonate fault-karst reservoir parameters based on extended elastic impedance inversion, revealing the favorable reservoir spatial distribution characteristics and reservoir enrichment laws, and providing a decision-making basis for further optimizing exploration targets and rationally formulating development plans for fault-karst reservoirs.

[0006] The specific solution provided by the present invention is as follows:

[0007] A method for quantitatively characterizing carbonate fault-karst reservoir parameters based on extended elastic impedance inversion, comprising:

[0008] S1. Rock physics analysis: Estimate elastic parameters, plot extended elastic impedance curves, conduct curve crossplot analysis and Chi projection angle scanning, and determine the Chi projection angle that meets the reservoir parameter correlation threshold;

[0009] S2. Acoustic impedance inversion: Based on the extracted seismic wavelet, using the drawn extended elastic impedance curve, through the combination of acoustic impedance synchronous inversion and stochastic inversion, high-resolution longitudinal wave acoustic impedance, shear wave acoustic impedance, and density attribute volumes are obtained;

[0010] S3. Gradient impedance inversion: First, through AVO attribute analysis technology, intercept and gradient attribute volumes are extracted; then an extended elastic impedance attribute volume with an incident angle of 90° is generated; finally, through the combination of synchronous inversion and stochastic inversion, a high-resolution gradient impedance inversion attribute volume is obtained;

[0011] S4. Extended elastic impedance inversion: According to the Chi projection angle scanning analysis results of the reservoir parameters to be predicted, an extended elastic impedance data volume is calculated;

[0012] S5. Reservoir quantitative prediction: Using the longitudinal wave acoustic impedance, shear wave acoustic impedance, and density attribute volumes generated by acoustic impedance inversion, neural network clustering analysis is carried out to identify lithology; crossplot analysis is performed between the reservoir parameters and the extended elastic impedance data volume, and relationship expressions are respectively fitted according to different lithologies, and finally the reservoir parameters are calculated and error correction is carried out.

[0013] Preferably, the steps of performing rock physics analysis in S1 are as follows:

[0014] S11: Estimate elastic parameters using a pre-set calculation model;

[0015] S12: Draw an extended elastic impedance curve according to the estimated elastic parameters and the selected formation model;

[0016] S13: Use data analysis technology to perform curve crossplot analysis, and combine this curve crossplot analysis to complete Chi projection angle scanning of the seismic source record;

[0017] S14: According to the curve crossplot analysis and Chi projection angle scanning results, determine the Chi projection angle that meets the reservoir parameter correlation threshold, and record this Chi projection angle.

[0018] Preferably, the steps of performing acoustic impedance inversion in S2 are as follows:

[0019] S21: Extract the seismic wavelet, and analyze the extended elastic impedance curve described in S1 according to the amplitude, frequency, and period attributes of the seismic wavelet;

[0020] S22: Combine acoustic impedance synchronous inversion and stochastic inversion technologies to obtain high-resolution longitudinal wave acoustic impedance, shear wave acoustic impedance, and density attribute volumes.

[0021] Preferably, the steps of acoustic impedance synchronous inversion and stochastic inversion technologies include:

[0022] S221: Data acquisition and preprocessing: Measure the behavior of seismic waves through seismic observations and perform preprocessing such as noise reduction, seismic enhancement, and amplitude correction to obtain processed seismic data;

[0023] S222: Use the acoustic impedance simultaneous inversion technique to input the preprocessed seismic data described in S221 into the acoustic impedance simultaneous inversion algorithm. The acoustic impedance simultaneous inversion algorithm uses wave reflection and refraction information to estimate changes in subsurface coefficients and capture longitudinal wave acoustic impedance, shear wave acoustic impedance, and density information;

[0024] S223: Use the stochastic inversion technique. Use random parameters as inputs, combine with the stochastic inversion mathematical model to generate pseudo data, and compare and match it with actual observed data to update the unknown parameters in the inversion;

[0025] S224: Iteration and optimization. Adjust the parameters and optimize the model according to the quality of the inversion results until the set target criteria are met. The target criteria include that the data fitting degree is better than the set threshold or the number of iterations reaches the set number.

[0026] Preferably, the gradient impedance inversion step described in S3 includes:

[0027] S31: Extract intercept and gradient attribute volumes through AVO amplitude and offset variation attribute analysis technology;

[0028] S31A: Collect multi-angle seismic data: Before starting the analysis, it is necessary to collect seismic data obtained from different offsets;

[0029] S31B: Conduct AVO analysis: Use the amplitude-versus-offset variation analysis technology to study how the seismic reflection amplitude changes with the offset and angle;

[0030] S31C: Extract intercept (A) and gradient (B) parameters: Through AVO analysis, extract intercept (A) and gradient (B) parameters to identify rock physical properties and fluid content;

[0031] S32: Use software or other tools to generate an extended elastic impedance attribute volume with an incident angle of 90°;

[0032] S32A: Calculate the extended elastic impedance (EEI): Use software tools to calculate the extended elastic impedance (EEI). The extended elastic impedance is an attribute that combines seismic data with different incident angles;

[0033] S32B: Set the incident angle to 90°: Pay special attention to the 90° incident angle situation that can provide key information on formation physical properties in the EEI calculation;

[0034] S33: Use the simultaneous inversion and stochastic inversion techniques to generate a higher-resolution gradient impedance inversion attribute volume;

[0035] S33A: Synchronous inversion: Use synchronous inversion technology to combine seismic data and other geological information to improve the accuracy of the impedance model;

[0036] S33B: Stochastic inversion: Apply stochastic inversion methods to explore multiple possible stratigraphic models by introducing randomness and find the optimal solution;

[0037] S33C: Resolution improvement: Through the above steps, generate a gradient impedance inversion attribute volume with higher resolution and more accurate formation physical property information.

[0038] Preferably, the steps of calculating and generating the extended elastic impedance data volume in S4 are as follows:

[0039] Using an inversion algorithm, compare the observed seismic data with the calculated extended elastic impedance and continuously adjust the reservoir parameter model to make the simulated data fit the observed data better; the inversion methods include: full waveform inversion algorithm, leapfrog algorithm, simulated annealing algorithm.

[0040] Preferably, the steps of the full waveform inversion algorithm are as follows:

[0041] S41: Define the initial model: The initial model is a guess of the subsurface structure and is an initialization model based on geological information;

[0042] S42: Generate synthetic seismic data: Using the initial model, generate synthetic seismic data by solving the elastic wave equation;

[0043] S43: Calculate the residual: Compare the synthetic seismic data with the observed data and calculate the residual between the model prediction and the actual observation;

[0044] S44: Construct the objective function: Take the residual as the objective function and also consider the additional condition of the smoothness of the model parameters;

[0045] S45: Gradient calculation: Calculate the gradient of the objective function with respect to the model parameters, i.e., the sensitivity;

[0046] S46: Update the model: Use the calculated gradient information to update the model through optimization algorithms such as the gradient descent method or quasi-Newton method;

[0047] S47: Iteration: Repeat S42 to S46 until the model converges to a state that satisfies certain convergence conditions.

[0048] Preferably, the steps of calculating and generating the extended elastic impedance data volume in S4 also include: evaluating the inversion result and evaluating the reservoir parameters obtained by inversion, including checking the consistency with geological knowledge and the stability of the model.

[0049] Preferably, the method for performing cluster analysis in S5 is as follows:

[0050] S51: Data preparation and preprocessing: P-wave acoustic impedance, S-wave acoustic impedance, and density data. These data are usually obtained from seismic data through seismic inversion techniques.

[0051] S52: Data cleaning: Clean the data and remove outliers or inconsistent data;

[0052] S53: Standardization or normalization: Since the magnitudes of different attributes may vary, perform standardization or normalization for easier analysis;

[0053] S54: Select and train a neural network model. Input the prepared data into the neural network, set the learning rate and number of iterations of the neural network, and use the data to train the neural network until the model is stable;

[0054] S55: Cluster analysis: Use the trained neural network to cluster the data. The neural network will divide the data into different groups according to the similarity of acoustic impedance and density attributes, and each group corresponds to a different lithology type; Interpret the clustering results: Interpret the lithology represented by each group according to the output of the neural network; Verification: Use known geological information or well logging data to verify the accuracy of the clustering results.

[0055] Preferably, the reservoir parameters in S5 include: reservoir porosity and water saturation.

[0056] Advantages of the present invention:

[0057] The present invention provides a method for quantitatively characterizing reservoir parameters of carbonate fracture-vug reservoirs based on extended elastic impedance inversion. By performing well logging rock physics analysis and AVO attribute analysis, elastic parameters are estimated, extended elastic impedance curves are plotted, curve crossplot analysis and Chi projection angle scanning are carried out to determine the Chi projection angle with a very high correlation with reservoir parameters; through the combination of prestack seismic synchronous inversion and stochastic inversion, high-resolution P-wave acoustic impedance, S-wave acoustic impedance, and density attribute volumes are obtained; using the P-wave acoustic impedance, S-wave acoustic impedance, and density attribute volumes generated by acoustic impedance inversion, neural network cluster analysis is performed to identify lithology; the reservoir parameters are crossplotted with the extended elastic impedance data volume, and relationship expressions are respectively fitted according to different lithologies, and finally reservoir parameters such as reservoir porosity and water saturation are quantitatively predicted, which has important practical significance for revealing the spatial distribution characteristics of favorable reservoirs and the enrichment laws of oil reservoirs, and for further optimizing exploration targets and reasonably formulating development plans for fracture-vug oil reservoirs.

[0058] To improve the accuracy of reservoir prediction, the extended elastic impedance technique is adopted for fine quantitative prediction of reservoirs. First, well logging petrophysical analysis is carried out. Here, we specify effective porosity - extended elastic impedance for correlation analysis. When the rotation red angle is 3°, the correlation between effective porosity and extended elastic impedance is the highest. In carbonate rock formations, from the EEI curves obtained by well logging petrophysical analysis, the EEI curves have a high correlation with the Por curves, while the correlation between the AI curves and the Por length is slightly lower. By comprehensively analyzing the EEI extended elastic impedance transformation parameters of multiple wells, the optimal rotation red projection angle is selected, and the EEI extended elastic impedance data volume with the best correlation with effective porosity is processed. Description of the Drawings

[0059] Figure 1 It is a flow chart of a method for quantitatively characterizing carbonate rock fracture - dissolution reservoir parameters based on extended elastic impedance inversion provided by the present invention.

[0060] Figure 2 It is a schematic diagram of the seismic attribute, reservoir inversion, and reservoir qualitative or quantitative prediction process in the embodiment. Detailed Embodiments

[0061] The present invention will be further described below with reference to the drawings and embodiments.

[0062] Embodiment 1:

[0063] As Figure 1 , 2 shown, a method for quantitatively characterizing carbonate rock fracture - dissolution reservoir parameters based on extended elastic impedance inversion includes:

[0064] S1. Petrophysical analysis: Estimate elastic parameters, draw extended elastic impedance curves, conduct curve cross - plot analysis and Chi projection angle scanning, and determine the Chi projection angle that meets the reservoir parameter correlation threshold;

[0065] S2. Acoustic impedance inversion: Based on the extracted seismic wavelet, using the drawn extended elastic impedance curves, through the combination of acoustic impedance synchronous inversion and stochastic inversion, obtain high - resolution longitudinal wave acoustic impedance, shear wave acoustic impedance, and density attribute volumes;

[0066] S3. Gradient impedance inversion: First, through AVO attribute analysis technology, extract intercept and gradient attribute volumes; then generate an extended elastic impedance attribute volume with an incident angle of 90°; finally, through the combination of synchronous inversion and stochastic inversion, obtain a high - resolution gradient impedance inversion attribute volume;

[0067] S4. Extended elastic impedance inversion: According to the Chi projection angle scanning analysis results of the reservoir parameters to be predicted, calculate the extended elastic impedance data volume;

[0068] S5. Reservoir Quantitative Prediction: Using the longitudinal wave acoustic impedance, shear wave acoustic impedance, and density attribute volumes generated by acoustic impedance inversion, perform neural network clustering analysis to identify lithology; conduct crossplot analysis of reservoir parameters and the extended elastic impedance data volume, fit relationship expressions separately according to different lithologies, finally calculate reservoir parameters, and perform error correction.

[0069] Preferably, the steps of performing rock physics analysis in S1 are as follows:

[0070] S11: Estimate elastic parameters using a pre-set calculation model;

[0071] S12: Draw an extended elastic impedance curve based on the estimated elastic parameters and the selected formation model;

[0072] S13: Use data analysis techniques to perform curve crossplot analysis, and complete Chi projection angle scanning on the seismic source record in combination with this curve crossplot analysis;

[0073] S14: Determine the Chi projection angle that meets the reservoir parameter correlation threshold according to the curve crossplot analysis and Chi projection angle scanning results, and record this Chi projection angle.

[0074] Preferably, the steps of performing acoustic impedance inversion in S2 are as follows:

[0075] S21: Extract the seismic wavelet, and analyze the extended elastic impedance curve described in S1 according to the amplitude, frequency, and period attributes of the seismic wavelet;

[0076] S22: Combine acoustic impedance simultaneous inversion and stochastic inversion techniques to obtain high-resolution longitudinal wave acoustic impedance, shear wave acoustic impedance, and density attribute volumes.

[0077] Preferably, the steps of acoustic impedance simultaneous inversion and stochastic inversion techniques include:

[0078] S221: Data acquisition and preprocessing: Measure the behavior of seismic waves through seismic observations and perform preprocessing such as noise reduction, seismic enhancement, and amplitude correction to obtain processed seismic data;

[0079] S222: Use acoustic impedance simultaneous inversion technology to input the preprocessed seismic data described in S221 into the acoustic impedance simultaneous inversion algorithm. The acoustic impedance simultaneous inversion algorithm uses wave reflection and refraction information to estimate changes in underground coefficients and captures longitudinal wave acoustic impedance, shear wave acoustic impedance, and density information;

[0080] S223: Use stochastic inversion technology, use random parameters as input, combine the stochastic inversion mathematical model to generate pseudo data, compare and match it with the actual observed data to update the unknown parameters in the inversion;

[0081] S224: Iteration and optimization. Adjust parameters and optimize the model according to the quality of the inversion results until the set target criteria are met. The target criteria include that the data fitting degree is better than the set threshold or the number of iterations reaches the set number of times.

[0082] Preferably, the gradient impedance inversion step in S3 includes:

[0083] S31: Extract intercept and gradient attribute volumes through AVO amplitude and offset variation attribute analysis technology;

[0084] S31A: Collect multi - angle seismic data: Before starting the analysis, seismic data obtained from different offsets need to be collected;

[0085] S31B: Conduct AVO analysis: Use the amplitude - versus - offset variation analysis technology to study how seismic reflection amplitudes change with offsets and angles;

[0086] S31C: Extract intercept (A) and gradient (B) parameters: Through AVO analysis, extract intercept (A) and gradient (B) parameters to identify rock physical properties and fluid content;

[0087] S32: Generate an extended elastic impedance attribute volume with an incident angle of 90° using software or other tools;

[0088] S32A: Calculate the extended elastic impedance (EEI): Use software tools to calculate the extended elastic impedance (EEI), which is an attribute that combines seismic data with different incident angles;

[0089] S32B: Set the incident angle to 90°: Pay special attention to the 90° incident angle situation in the EEI calculation, which can provide key information about formation physical properties;

[0090] S33: Use simultaneous inversion and stochastic inversion techniques to generate a higher - resolution gradient impedance inversion attribute volume;

[0091] S33A: Simultaneous inversion: Use simultaneous inversion technology to combine seismic data and other geological information to improve the accuracy of the impedance model;

[0092] S33B: Stochastic inversion: Apply the stochastic inversion method, introduce randomness to explore multiple possible formation models, and find the optimal solution;

[0093] S33C: Improve resolution: Through the above steps, generate a gradient impedance inversion attribute volume with higher resolution and more accurate formation physical property information.

[0094] Preferably, the steps for calculating and generating the extended elastic impedance data volume in S4 are:

[0095] Using the inversion algorithm, compare the observed seismic data with the calculated extended elastic impedance, and continuously adjust the reservoir parameter model to make the simulated data fit the observed data better; the inversion method includes: full waveform inversion algorithm, leapfrog algorithm, simulated annealing algorithm.

[0096] Preferably, the steps of the full waveform inversion algorithm are as follows:

[0097] S41: Define the initial model: The initial model is a guess of the subsurface structure and is an initialization model based on geological information;

[0098] S42: Generate synthetic seismic data: Using the initial model, solve the elastic wave equation to generate synthetic seismic data;

[0099] S43: Calculate the residual: Compare the synthetic seismic data with the observed data and calculate the residual between the model prediction and the actual observation;

[0100] S44: Construct the objective function: Use the residual as the objective function and consider the additional condition of the smoothness of the model parameters;

[0101] S45: Gradient calculation: Calculate the gradient of the objective function with respect to the model parameters, i.e., the sensitivity;

[0102] S46: Update the model: Use the calculated gradient information to update the model through optimization algorithms such as the gradient descent method or the quasi-Newton method;

[0103] S47: Iteration: Repeat S42 to S46 until the model converges to a state that satisfies certain convergence conditions.

[0104] Preferably, the steps of calculating and generating the extended elastic impedance data volume in S4 further include: evaluating the inversion results and evaluating the reservoir parameters obtained by inversion, including checking the consistency with geological knowledge and the stability of the model.

[0105] Preferably, the method for performing cluster analysis in S5 is as follows:

[0106] S51: Data preparation and preprocessing: P-wave acoustic impedance, S-wave acoustic impedance, and density data. These data are usually obtained from seismic data through seismic inversion techniques.

[0107] S52: Data cleaning: Clean the data and remove outliers or inconsistent data;

[0108] S53: Standardization or normalization: Since the magnitudes of different attributes may be different, perform standardization or normalization for easy analysis;

[0109] S54: Select and train a neural network model. Input the prepared data into the neural network, set the learning rate and iteration number parameters of the neural network, and use the data to train the neural network until the model is stable;

[0110] S55: Cluster analysis: Use the trained neural network to cluster the data. The neural network will divide the data into different groups according to the similarity of acoustic impedance and density attributes, and each group corresponds to a different lithology type; Interpret the clustering results: Interpret the lithology represented by each group according to the output of the neural network; Verification: Use known geological information or well logging data to verify the accuracy of the clustering results.

[0111] Preferably, the reservoir parameters described in S5 include: reservoir porosity, water saturation.

[0112] It should be noted that the above specific embodiments can enable those skilled in the art to understand the present invention more comprehensively, but do not limit the present invention in any way. Therefore, although this specification has described the present invention in detail with reference to the drawings and embodiments, those skilled in the art should understand that the present invention can still be modified or equivalently replaced. In short, all technical solutions and their improvements that do not depart from the spirit and scope of the present invention should be covered by the protection scope of the present invention patent.

Claims

1. A method for quantitatively characterizing reservoir parameters of carbonate fracture-vuggy bodies based on extended elastic impedance inversion, characterized in that, Including: S1. Rock physics analysis: estimating elastic parameters, plotting extended elastic impedance curves, performing crossplot analysis and Chi projection angle scanning, and determining the Chi projection angle that meets the reservoir parameter correlation threshold; S2. Acoustic impedance inversion: based on the extracted seismic wavelet, using the plotted extended elastic impedance curve, through the combination of acoustic impedance simultaneous inversion and stochastic inversion, obtaining high-resolution longitudinal wave acoustic impedance, transverse wave acoustic impedance, and density attribute volumes; S3. Gradient impedance inversion: first, through AVO attribute analysis technology, extracting intercept and gradient attribute volumes; then generating an extended elastic impedance attribute volume with an incident angle of 90°; finally, through the combination of simultaneous inversion and stochastic inversion, obtaining a high-resolution gradient impedance inversion attribute volume; S4. Extended elastic impedance inversion: calculating the extended elastic impedance data volume according to the Chi projection angle scanning analysis results of the reservoir parameters to be predicted; S5. Reservoir quantitative prediction: using the longitudinal wave acoustic impedance, transverse wave acoustic impedance, and density attribute volumes generated by acoustic impedance inversion to perform neural network clustering analysis to discriminate lithology; performing crossplot analysis on the reservoir parameters and the extended elastic impedance data volume, fitting relationship expressions respectively according to different lithologies, and finally calculating the reservoir parameters and performing error correction.

2. A method for quantitatively characterizing reservoir parameters of carbonate fracture-vuggy bodies based on extended elastic impedance inversion according to claim 1, characterized in that The steps of performing rock physics analysis in S1 are as follows: S11: Estimating elastic parameters using a pre-set calculation model; S12: Plotting an extended elastic impedance curve according to the estimated elastic parameters and the selected formation model; S13: Performing crossplot analysis using data analysis technology and completing Chi projection angle scanning on the seismic source record in combination with this crossplot analysis; S14: According to the crossplot analysis and Chi projection angle scanning results, determining the Chi projection angle that meets the reservoir parameter correlation threshold and recording this Chi projection angle.

3. A method for quantitatively characterizing carbonate fracture-vug reservoir parameters based on extended elastic impedance inversion according to claim 1, characterized in that, The steps of performing acoustic impedance inversion in S2 are as follows: S21: Extracting the seismic wavelet and analyzing the extended elastic impedance curve described in S1 according to the amplitude, frequency, and period attributes of the seismic wavelet; S22: Combining acoustic impedance simultaneous inversion and stochastic inversion technologies to obtain high-resolution longitudinal wave acoustic impedance, transverse wave acoustic impedance, and density attribute volumes.

4. A method for quantitatively characterizing reservoir parameters of carbonate fracture-vuggy bodies based on extended elastic impedance inversion according to claim 3, characterized in that The steps of acoustic impedance simultaneous inversion and stochastic inversion technologies include: S221: Data acquisition and preprocessing: measuring the behavior of seismic waves through seismic observations and performing preprocessing such as noise reduction, seismic enhancement, and amplitude correction to obtain processed seismic data; S222: Using acoustic impedance simultaneous inversion technology, inputting the preprocessed seismic data described in S221 into the acoustic impedance simultaneous inversion algorithm, and the acoustic impedance simultaneous inversion algorithm uses wave reflection and refraction information to estimate the change of underground coefficients and capture longitudinal wave acoustic impedance, transverse wave acoustic impedance, and density information; S223: Using stochastic inversion technology, using random parameters as input, combining with a stochastic inversion mathematical model to generate pseudo data, comparing and matching with actual observed data to update the unknown parameters in the inversion; S224: Iteration and optimization. Adjust parameters and optimize the model according to the quality of the inversion results until the set target criteria are met. The target criteria include that the data fitting degree is better than the set threshold or the number of iterations reaches the set number of times.

5. A method for quantitatively characterizing reservoir parameters of carbonate fracture-vuggy bodies based on extended elastic impedance inversion according to claim 1, characterized in that The gradient impedance inversion steps described in S3 include: S31: Extract the intercept and gradient attribute volumes through the AVO amplitude and offset variation attribute analysis technique; S31A: Collect multi-angle seismic data: Before starting the analysis, seismic data obtained from different offsets need to be collected; S31B: Conduct AVO analysis: Use the amplitude-versus-offset analysis technique to study how the seismic reflection amplitude changes with the offset and angle; S31C: Extract the intercept (A) and gradient (B) parameters: Through AVO analysis, extract the intercept (A) and gradient (B) parameters to identify rock physical properties and fluid content; S32: Generate an extended elastic impedance attribute volume with an incident angle of 90° using software or other tools; S32A: Calculate the extended elastic impedance (EEI): Use software tools to calculate the extended elastic impedance (EEI), which is an attribute that combines seismic data at different incident angles; S32B: Set the incident angle to 90°: Pay special attention to the 90° incident angle situation that can provide key information on formation physical properties in the EEI calculation; S33: Use synchronous inversion and stochastic inversion techniques to generate a higher-resolution gradient impedance inversion attribute volume; S33A: Synchronous inversion: Use synchronous inversion techniques to combine seismic data and other geological information to improve the accuracy of the impedance model; S33B: Stochastic inversion: Apply the stochastic inversion method to explore multiple possible formation models by introducing randomness and find the optimal solution; S33C: Improve the resolution: Through the above steps, generate a gradient impedance inversion attribute volume with higher resolution and more accurate formation physical property information.

6. A method for quantitatively characterizing reservoir parameters of carbonate fracture-vuggy bodies based on extended elastic impedance inversion according to claim 1, characterized in that The steps for calculating and generating the extended elastic impedance data volume in S4 are: Using the inversion algorithm, compare the observed seismic data with the calculated extended elastic impedance, and continuously adjust the reservoir parameter model to make the simulated data fit the observed data better; The inversion methods include: full waveform inversion algorithm, leapfrog algorithm, simulated annealing algorithm.

7. A method for quantitatively characterizing reservoir parameters of carbonate fracture-vuggy bodies based on extended elastic impedance inversion according to claim 1, characterized in that, The steps of the full waveform inversion algorithm are: S41: Define the initial model: The initial model is a guess of the underground structure and is an initialized model based on geological information; S42: Generate synthetic seismic data: Using the initial model, solve the elastic wave equation to generate synthetic seismic data; S43: Calculate the residual: Compare the synthetic seismic data with the observed data and calculate the residual between the model prediction and the actual observation; S44: Construct the objective function: Use the residual as the objective function, and at the same time consider the additional condition of the smoothness of the model parameters; S45: Gradient calculation: Calculate the gradient of the objective function with respect to the model parameters, that is, the sensitivity; S46: Update the model: Use the calculated gradient information and update the model through optimization algorithms such as the gradient descent method or the quasi-Newton method; S47: Iteration: Repeat S42 to S46 until the model converges to a state that meets certain convergence conditions.

8. A method for quantitatively characterizing reservoir parameters of carbonate fracture-vuggy bodies based on extended elastic impedance inversion according to claim 1, characterized in that The steps of calculating and generating the extended elastic impedance data volume in S4 further include: evaluating the inversion results and evaluating the reservoir parameters obtained by inversion, including checking the consistency with geological knowledge and the stability of the model.

9. A method for quantitatively characterizing reservoir parameters of carbonate fracture-vuggy bodies based on extended elastic impedance inversion according to claim 1, characterized in that The method of performing cluster analysis in S5 is as follows: S51: Data preparation and preprocessing: P-wave acoustic impedance, S-wave acoustic impedance, and density data. These data are usually obtained from seismic data through seismic inversion techniques. S52: Data cleaning: Clean the data and remove outliers or inconsistent data. S53: Standardization or normalization: Since the magnitudes of different attributes may be different, perform standardization or normalization for easy analysis. S54: Select and train a neural network model. Input the prepared data into the neural network, set the learning rate and iteration number parameters of the neural network, and use the data to train the neural network until the model is stable. S55: Cluster analysis: Use the trained neural network to cluster the data. The neural network will divide the data into different groups according to the similarity of acoustic impedance and density attributes, and each group corresponds to a different lithology type; Interpret the clustering results: Interpret the lithology represented by each group according to the output of the neural network. Verification: Use known geological information or well logging data to verify the accuracy of the clustering results.

10. A method for quantitatively characterizing reservoir parameters of carbonate fracture-vuggy reservoirs based on extended elastic impedance inversion according to claim 1, characterized in that, The reservoir parameters described in S5 include: reservoir porosity, water saturation.

Images