A method, device, equipment and medium for predicting sweet spots of a tight sandstone reservoir
By combining pre-stack inversion and well data-driven sensitivity analysis with intelligent algorithms, a sweet spot prediction model for tight sandstone reservoirs is established, solving the problem of identifying sweet spot areas in traditional methods and achieving high accuracy and reliability in sweet spot prediction.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHANGHAI BRANCH CHINA OILFIELD SERVICES
- Filing Date
- 2026-06-01
- Publication Date
- 2026-07-24
AI Technical Summary
In tight sandstone reservoirs, existing technologies struggle to effectively identify and differentiate sweet spots, traditional post-stack impedance inversion methods fail to identify subtle impedance differences, and multi-seismic attribute analysis lacks physical meaning, leading to unstable prediction results.
By obtaining multi-dimensional elastic parameter data volumes through pre-stack inversion, and combining well data-driven sensitivity analysis and intelligent algorithms, a sweet spot prediction model with clear physical meaning is established using deep learning networks and iterative optimization techniques, and sensitive elastic parameters are selected for prediction.
It improves the accuracy and reliability of sweet spot prediction in tight sandstone reservoirs, ensuring that the model has a rock physics basis and can accurately identify sweet spot zones in reservoirs with high porosity and high permeability.
Smart Images
Figure CN122447069A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of oil and gas exploration and development technology, and in particular to a method, apparatus, equipment and medium for predicting sweet spots in tight sandstone reservoirs. Background Technology
[0002] As oil and gas exploration and development progresses to medium and deep formations, reservoirs have undergone long-term and intense diagenesis and compaction, generally exhibiting low porosity and low permeability characteristics. Under this geological context, identifying sweet spots with relatively high porosity and high permeability is crucial for the economical and effective development of tight sandstone oil and gas.
[0003] Currently, prediction of sweet spots in tight sandstone reservoirs typically relies on post-stack impedance inversion and multi-seismic attribute fusion analysis. However, due to the extremely weak impedance differences between the reservoir sweet spots and the surrounding rocks, and the high degree of overlap in characteristics, traditional post-stack impedance inversion methods struggle to effectively identify and distinguish sweet spots. In areas with complex seismic responses, while using multiple seismic attributes may increase the number of features in reservoir sweet spots, the relationships between these features and reservoir physical parameters are complex and lack physical meaning. Simple linear fitting and other mathematical statistical methods are insufficient to uncover patterns reflecting geological characteristics, resulting in models with poor generalization ability and unstable prediction results. Summary of the Invention
[0004] This application provides a method, apparatus, equipment, and medium for predicting sweet spots in tight sandstone reservoirs. The method obtains a multi-dimensional elastic parameter data volume through pre-stack inversion and combines it with well data-driven sensitivity analysis and a sweet spot prediction model. It combines elastic parameters with clear physical meaning with intelligent algorithms that can characterize complex nonlinear relationships, ensuring that the sweet spot prediction model has a rock physics basis and improving the accuracy and reliability of sweet spot prediction.
[0005] According to one aspect of this application, a method for predicting sweet spots in tight sandstone reservoirs is provided, the method comprising: Obtain CRP gathers for the target exploration area, as well as logging curves of actual drilled wells within the target exploration area; Based on the CRP gather and the logging curves, pre-stack simultaneous inversion is performed to determine the basic elastic parameter data volume of the target exploration area. Based on the aforementioned basic elastic parameter data volume, a sensitivity analysis of the sweet spot area is performed to determine the sensitive elastic parameter data volume; The sensitive elasticity parameter data volume is input into a pre-trained sweet spot prediction model to obtain the sweet spot area of the target exploration area.
[0006] According to another aspect of this application, a sweet spot prediction device for tight sandstone reservoirs is provided, characterized in that the device comprises: The data acquisition module is used to acquire CRP gathers of the target exploration area and logging curves of actual drilled wells in the target exploration area; The data volume calculation module is used to perform pre-stack simultaneous inversion based on the CRP gather and the logging curves to determine the basic elastic parameter data volume of the target exploration area; The sensitivity analysis module is used to perform sweet spot sensitivity analysis based on the basic elastic parameter data body to determine the sensitive elastic parameter data body; The sweet spot prediction module is used to input the sensitive elasticity parameter data into a pre-trained sweet spot prediction model to obtain the sweet spot of the target exploration area.
[0007] According to another aspect of this application, an electronic device is provided, the device comprising: at least one processor; and a memory communicatively connected to said at least one processor; wherein the memory stores a computer program executable by said at least one processor, said computer program being executed by said at least one processor to enable said at least one processor to perform the sweet spot prediction method for tight sandstone reservoirs according to any embodiment of this application.
[0008] According to another aspect of this application, a computer-readable storage medium is provided that stores computer instructions for causing a processor to execute and implement the sweet spot prediction method for tight sandstone reservoirs according to any embodiment of this application.
[0009] According to another aspect of this application, a computer program product is provided, the computer program product comprising a computer program that, when executed by a processor, implements the sweet spot prediction method for tight sandstone reservoirs according to any embodiment of this application.
[0010] The technical solution provided in this application involves acquiring CRP gathers and logging curves from actual drilled wells in the target exploration area; performing pre-stack simultaneous inversion based on the CRP gathers and logging curves to determine the basic elastic parameter data volume of the target exploration area; conducting sweet spot sensitivity analysis based on the basic elastic parameter data volume to determine the sensitive elastic parameter data volume; and inputting the sensitive elastic parameter data volume into the sweet spot prediction model to obtain the sweet spot of the target exploration area. This technical solution obtains a multi-dimensional elastic parameter data volume through pre-stack inversion, combines well data-driven sensitivity analysis and a sweet spot prediction model, and integrates elastic parameters with clear physical meaning with intelligent algorithms capable of characterizing complex nonlinear relationships. This ensures that the sweet spot prediction model has a rock physics basis, improving the accuracy and reliability of sweet spot prediction.
[0011] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent from the following description. Attached Figure Description
[0012] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0013] Figure 1 This is a flowchart illustrating a method for predicting sweet spots in tight sandstone reservoirs, as provided in Embodiment 1 of this application.
[0014] Figure 2 This is a schematic diagram of the structure of a dessert prediction model provided in Embodiment 1 of this application.
[0015] Figure 3 This is a flowchart illustrating a method for predicting sweet spots in tight sandstone reservoirs, as provided in Embodiment 2 of this application.
[0016] Figure 4 This is a cross-plot of permeability and various candidate elastic parameters provided in Embodiment 2 of this application.
[0017] Figure 5 This is a comparison chart of well bypass prediction for dessert prediction provided in Embodiment 2 of this application.
[0018] Figure 6 This is a three-dimensional intersection map for dessert prediction provided in Embodiment 2 of this application.
[0019] Figure 7a is an impedance inversion cross-sectional view provided in Embodiment 2 of this application.
[0020] Figure 7b This is a pre-stack density inversion profile provided in Embodiment 2 of this application.
[0021] Figure 7c This is a pre-stack LR coefficient inversion profile provided in Embodiment 2 of this application.
[0022] Figure 8 This is a cross-sectional view of a dessert prediction model provided in Embodiment 2 of this application.
[0023] Figure 9 This is a schematic diagram of a sweet spot prediction device for tight sandstone reservoirs provided in Embodiment 3 of this application.
[0024] Figure 10This is a schematic diagram of the apparatus for implementing a sweet spot prediction method for tight sandstone reservoirs according to an embodiment of this application. Detailed Implementation
[0025] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.
[0026] It should be noted that the terms "candidate," "sensitive," "basic," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0027] It should also be noted that the acquisition, storage, use, and processing of data in the technical solution of this application all comply with the relevant provisions of national laws and regulations.
[0028] Example 1 Figure 1 This is a flowchart of a sweet spot prediction method for tight sandstone reservoirs provided in Embodiment 1 of this application. This embodiment is applicable to the situation of sweet spot prediction for tight sandstone reservoirs. The method can be executed by a sweet spot prediction device for tight sandstone reservoirs. The sweet spot prediction device for tight sandstone reservoirs can be implemented in hardware and / or software form. The sweet spot prediction device for tight sandstone reservoirs can be configured in a device with data processing capabilities.
[0029] like Figure 1 As shown, the method includes the following steps.
[0030] S110. Obtain the CRP gather of the target exploration area and the logging curves of the actual drilled wells in the target exploration area.
[0031] CRP gathers, or common reflection point gathers, consist of pre-stack seismic traces that have undergone dynamic correction and partial migration processing. These traces originate from the same underground reflection point and can reflect the relationship between seismic wave amplitude and incident angle. They are the basic data for pre-stack elastic parameter inversion.
[0032] Specifically, raw seismic single-shot records can be obtained using a source and a detector array, and then generated through a series of high-fidelity data processing procedures, including static correction, dynamic correction, pre-stack time migration, or depth migration.
[0033] Well logging curves refer to a series of curves continuously measured and recorded in a drilled wellbore using instruments such as wireline logging and logging-while-drilling (LMD) to reflect the physical properties of the formation rocks. Commonly used curves include sonic logging, density logging, gamma ray logging, resistivity logging, and neutron porosity logging, which can reflect the elastic and physical parameters of the formation at the well point.
[0034] Specifically, after drilling is completed or during drilling, a set of professional logging instruments can be lowered into the well. During the lifting or lowering process, the instrument sensors measure the formation around the well wall and transmit the data back to the surface system for recording and preliminary processing, ultimately generating multiple logging curves with depth alignment.
[0035] In some embodiments, optionally, after obtaining the CRP gather for the target exploration area, the method further includes: performing noise suppression, amplitude compensation, and amplitude preservation processing on the CRP gather to obtain a high-quality gather.
[0036] To ensure the reliability of subsequent pre-stack inversion results, this application also performs a series of preprocessing steps on the CRP gather to obtain high-quality gathers.
[0037] Noise suppression involves using targeted filtering methods to suppress or eliminate various types of noise in the channel set, including random noise, coherent noise, and anomalous amplitude interference. Specifically, for random noise, denoising algorithms based on wavelet transform, curvelet transform, or singular value decomposition can be used to suppress it in the mathematical domain where the effective signal and random noise are most easily separated. For coherent noise, such as multiple waves and surface waves, methods such as Radon transform (τ-p domain), FK filtering, or predictive deconvolution are used for separation and elimination. For anomalous amplitude, sudden strong amplitude noise caused by instrument malfunction or near-surface interference can be identified and suppressed.
[0038] Amplitude compensation mainly involves two aspects: first, spherical diffusion compensation, which compensates for the geometric energy attenuation caused by wavefront expansion during seismic wave propagation. This can be based on the underground velocity model to calculate the geometric energy attenuation caused by the expansion of the seismic wavefront with the propagation distance and perform inverse proportional compensation; second, absorption attenuation compensation, which compensates for the energy loss and wavelet distortion caused by the absorption of high-frequency components of seismic waves by the strata. This can be based on the absorption characteristics of the strata (such as the quality factor Q) and techniques such as inverse Q filtering can be used to compensate for the high-frequency energy loss and wavelet phase distortion caused by strata absorption, thereby broadening the effective frequency band.
[0039] Amplitude preservation processing is used to ensure that the noise suppression and amplitude compensation processes are relatively amplitude-preserving. That is, the relative amplitude relationship between channels at different locations and with different incident angles within the gathered after processing can truly reflect the changes in the reflection coefficient of the underground interface, especially the AVO / AVA characteristics related to fluids and lithology.
[0040] S120. Based on the CRP gather and the logging curves, perform pre-stack simultaneous inversion to determine the basic elastic parameter data volume of the target exploration area.
[0041] The basic elastic parameter data volume refers to a three-dimensional data set obtained through pre-stack seismic inversion, containing physical parameters describing the elastic properties of subsurface rocks. Typically, the basic elastic parameter data volume includes P-wave velocity data, S-wave data, and density data.
[0042] In this application, a deep learning network can be used to learn the nonlinear mapping of basic elastic parameters from seismic data. Specifically, CRP gathers are used as input features to the network, and the corresponding P-wave velocity curves, S-wave velocity curves, and density curves are used as training labels to train the deep learning network. The CRP gathers of the target exploration area are then input into the trained network to quickly output the basic elastic parameter data volume of the target exploration area.
[0043] In this application, the basic elastic parameter data volume of the target exploration area can also be obtained through iterative optimization based on a physical model. Specifically, firstly, using the P-wave velocity curve, S-wave velocity curve, and density curve from the well logging curves, three-dimensional spatial interpolation and extrapolation are performed under the constraints of the interpretation horizon and structural framework to establish a low-frequency model reflecting the background trend; then, seismic wavelets related to the incident angle are extracted from the CRP gather near the well, and a forward modeling operator is established from the forward modeling of the elastic parameter synthesized angle gather; finally, starting from the low-frequency model, iterative calculation is performed using a Bayesian inversion framework or a least squares optimization algorithm. In each iteration, the residuals of the synthesized gather and the actual CRP gather are compared, and the elastic parameter model is updated through the optimization algorithm until the residuals are minimized and converged, outputting the P-wave velocity data volume, S-wave data volume, and density data volume.
[0044] In some embodiments, optionally, the step of performing pre-stack simultaneous inversion based on the CRP gather and the logging curves to determine the basic elastic parameter data volume of the target exploration area includes: performing pre-stack simultaneous inversion based on the CRP gather using the AVO approximation formula to determine the basic elastic parameter reflectivity data volume of the target exploration area; and performing integral recursion on the basic elastic parameter reflectivity data volume based on the logging curves to obtain the basic elastic parameter data volume of the target exploration area.
[0045] The AVO approximation formula, also known as the Aki-Richards equation, is used for each time sampling point. Select ( With different incident angles and amplitude values, a system of linear equations is constructed to solve for the basic elastic parameter data volume.
[0046] The linearized expression of the Aki-Richards equation is: ; In the formula, , , ; The ratio representing the average values of P-wave and S-wave velocities. ; The relative reflectivity, representing the longitudinal wave velocity, ; The relative reflectivity, representing the transverse wave velocity, ; Relative reflectivity, representing density .
[0047] In this application, for each time sampling point Select 3 different incident angles , and Substituting these values into the linear formula above, corresponding to the near, middle, and far angles respectively, yields a system of three linear equations in three variables: ; In the formula, , and The angle of incidence is The result can be obtained by calculating the coefficients using the formulas described above. , and The angle of incidence is The result can be obtained by calculating the coefficients using the formulas described above. , and The angle of incidence is The result can be obtained by calculating the coefficients using the formulas described above. , and yes The amplitude values at different incident angles corresponding to the sampling points at different times are known numbers; , and The unknown is to be solved.
[0048] Of course, more than three incident angles can be selected, and the solution can be obtained by the least squares method.
[0049] Sampling points at each time point , and After solving, the relative reflectivity profiles of P-wave velocity, S-wave velocity, and density in the entire time domain can be obtained.
[0050] Subsequently, through integral recursion and by combining well logging curves and stratigraphics, an initial model was established to obtain P-wave velocity data volume, S-wave velocity data volume, and density data volume.
[0051] S130. Based on the basic elastic parameter data body, perform a sensitivity analysis of the sweet spot area to determine the sensitive elastic parameter data body.
[0052] Sensitive elastic parameter data sets refer to elastic parameter data sets that have a strong ability to distinguish sweet spots (i.e., high-porosity, high-permeability, and high-yield high-quality reservoirs). The purpose is to eliminate insensitive or redundant elastic parameters to avoid interfering with subsequent sweet spot prediction models, thereby improving the efficiency and generalization ability of the sweet spot prediction models.
[0053] In this application, the Spearman rank correlation coefficient between each basic elasticity parameter and the dessert label can be calculated. A larger absolute value of the correlation coefficient indicates a stronger monotonic correlation between the elasticity parameter and the dessert. Alternatively, an analysis of variance (ANOVA) can be performed on the basic elasticity parameters of different dessert labels to calculate the F-value, which is the ratio of component variance to within-group variance. A larger F-value indicates a more significant difference in the mean of the elasticity parameter among different dessert labels, i.e., better category separability. Finally, the elasticity parameters are selected as sensitive elasticity parameters by ranking them from highest to lowest according to the correlation coefficient or F-value.
[0054] S140. Input the sensitive elastic parameter data into the pre-trained sweet spot prediction model to obtain the sweet spot area of the target exploration area.
[0055] The dessert prediction model is a classification model trained using known data, based on machine learning algorithms (such as random forests, support vector machines, and neural networks). The input to the dessert prediction model is a data volume of sensitive elastic parameters, and the output is dessert labels.
[0056] For example, the sweet spot prediction model is a deep convolutional neural network. By inputting the sensitive elasticity parameter data volume into the sweet spot prediction model and performing a single forward propagation, the network can output the predicted sweet spot area of the target exploration zone in three-dimensional space, i.e., the sweet spot region.
[0057] In some embodiments, optionally, the training process of the dessert prediction model includes: acquiring a training sample set, and sampling with replacement from the training sample set using the Bootstrap sampling method to generate several training subsets; wherein the training sample set includes the sensitivity elasticity parameter curves of the learning wells and dessert type labels; inputting each of the training subsets into a random forest model for parallel training to obtain multiple decision trees; wherein, when training a single decision tree, at each node, a portion of parameters are randomly selected from all sensitivity elasticity parameters to find the optimal splitting rule; and obtaining the dessert prediction model in response to the training termination condition.
[0058] In this application, to better learn the nonlinear relationship between the sensitive elasticity parameter and dessert classification, a random forest model is chosen to construct the dessert prediction model, such as... Figure 2 As shown, Figure 2 This is a schematic diagram of the structure of a dessert prediction model provided in Embodiment 1 of this application.
[0059] The first step is to select the measured shear wave well as the learning well. Using the logging curve and CRP gather, the sensitive elastic parameter curve of the learning well is determined according to the method described in steps S120 to S130, as well as the sweet spot type label (including sweet spots of type I, II-1, II-2, III-1, and III-2) corresponding to the well section.
[0060] The second step involves using Bootstrap sampling, which randomly draws samples with replacement from the training sample set to generate multiple differentiated training subsets. This ensures that each decision tree in subsequent training learns from different data perspectives, reducing the variance of the random forest model and preventing overfitting.
[0061] The third step is to assign each training subset to different decision trees for parallel training. When training a single decision tree, another element of randomness is introduced: when searching for the optimal splitting rule at each node, not all sensitive elasticity parameters are considered, but only a subset of all parameters is randomly selected for evaluation.
[0062] Because the Bootstrap sampling method is used to randomly sample the training sample set, some original samples will be missed in the generated training subset. The samples that are not selected are used as the out-of-bag data of the decision tree. The out-of-bag error is calculated using the out-of-bag data to evaluate the performance of the decision tree.
[0063] The fourth step involves training all decision trees. Once the training termination condition is met, such as reaching a preset number of trees, all decision trees are combined to form a sweet spot prediction model. When using this model to predict new data, all decision trees vote, and the combined predictions determine the sweet spot prediction.
[0064] Since the sweet spot prediction model is trained based on the logging curve of a sampling point in a learning well, the sensitive elastic parameter data volume of the target exploration area needs to be input into the sweet spot prediction model one by one. Finally, the prediction results are combined to obtain the three-dimensional sweet spot data volume, thereby obtaining the sweet spot area of the target exploration area.
[0065] This application provides a method for predicting sweet spots in tight sandstone reservoirs. The method involves acquiring CRP gathers and logging curves from drilled wells in the target exploration area; performing pre-stack simultaneous inversion based on the CRP gathers and logging curves to determine the basic elastic parameter data volume of the target exploration area; conducting sweet spot sensitivity analysis based on the basic elastic parameter data volume to determine the sensitive elastic parameter data volume; and inputting the sensitive elastic parameter data volume into the sweet spot prediction model to obtain the sweet spot area of the target exploration area. This technical solution obtains a multi-dimensional elastic parameter data volume through pre-stack inversion, combines well data-driven sensitivity analysis and a sweet spot prediction model, and integrates elastic parameters with clear physical meaning with intelligent algorithms capable of characterizing complex nonlinear relationships. This ensures that the sweet spot prediction model has a rock physics basis, improving the accuracy and reliability of sweet spot prediction.
[0066] Example 2 Figure 3 This is a flowchart illustrating a sweet spot prediction method for tight sandstone reservoirs provided in Embodiment 2 of this application. This embodiment is an optimization based on the above embodiment, specifically optimizing the process of the equations. Figure 3 As shown, the method includes the following steps.
[0067] S210. Obtain the CRP gather of the target exploration area and the logging curves of the actual drilled wells in the target exploration area.
[0068] S220. Based on the CRP gather and the logging curves, perform pre-stack simultaneous inversion to determine the basic elastic parameter data volume of the target exploration area.
[0069] S230. Based on the basic elastic parameter data body, determine the candidate elastic parameter data body.
[0070] The candidate elastic parameter data volume is a derived elastic parameter data volume derived from the basic elastic parameter data volume through mathematical operations and rock physics formulas.
[0071] In some embodiments, the candidate elastic parameter data body may optionally include at least the longitudinal wave impedance data body, the transverse wave impedance data body, the Young's modulus data body, the Poisson's ratio data body, the Lamé constant data body, the shear modulus data body, the longitudinal and transverse wave velocity ratio data body, and the bulk modulus data body.
[0072] Specifically, the longitudinal wave impedance, transverse wave impedance, Young's modulus, Poisson's ratio, Lamé constant, shear modulus, longitudinal wave velocity ratio, and bulk modulus can be determined using the following formulas.
[0073] Longitudinal wave impedance: ; Transverse wave impedance: ; Young's modulus: ; Poisson's ratio: ; Lamé constant: ; Shear modulus: ; P-wave to S-wave velocity ratio: ; Bulk modulus: ; In the above formulas, Indicates the longitudinal wave velocity. Indicates the transverse wave velocity. Indicates density.
[0074] The candidate elastic parameter data body is obtained by calculating the basic elastic parameter data body using the above calculation formulas.
[0075] S240. Based on the interpretation results of the well logging curves, perform sweet spot sensitivity analysis on each of the candidate elastic parameter data volumes to determine the sensitive elastic parameter data volumes.
[0076] The interpretation curve of well logging curves refers to the curves of parameters such as porosity, permeability, water saturation, and clay content obtained by comprehensively interpreting well logging curves.
[0077] In this application, a cross plot can be established based on the interpretation result curve and the candidate elastic parameter data volume, such as a cross plot with the interpretation result parameters as the vertical axis and the candidate elastic parameters as the horizontal axis. By performing quantitative statistics on the well point data in the cross plot, the sensitive elastic parameter data volume that meets the sensitivity threshold can be determined.
[0078] In some embodiments, optionally, the step of performing sweet spot sensitivity analysis on each candidate elastic parameter data body based on the interpretation result curve of the well logging curve to determine the sensitive elastic parameter data body includes: establishing a sweet spot type label based on the interpretation result curve of the well logging curve; and performing correlation analysis on the interpretation result curve of the well logging curve and each candidate elastic parameter data body based on the sweet spot type label to determine the sensitive elastic parameter data body.
[0079] The reservoir's quality can be classified into four categories, from highest to lowest, according to the comprehensive quality evaluation standard: Category I, Category II-1, Category II-2, Category III-1, and Category III-2. The higher the category, the better the reservoir quality.
[0080] Specifically, the sweet spot type label for a wellpoint can be determined according to a predetermined classification standard based on its porosity, oil saturation, brittleness index, and clay content. For example, if a wellpoint has a porosity > 12%, oil saturation > 65%, brittleness index > 0.5, and clay content < 10%, it is classified as a Class I sweet spot.
[0081] In this application, cross-plots with each candidate elastic parameter data volume can be established based on the permeability curve, such as... Figure 4 As shown. In Figure 4 In this model, data points for different dessert types are represented by dots of different colors. Based on the distribution of data points for different dessert types, the sensitivity elasticity parameters that are sensitive to the dessert region can be determined by calculating correlation coefficients, F-values, or through visual analysis, thereby defining the sensitivity elasticity parameter data volume.
[0082] S250. Input the sensitive elastic parameter data into the pre-trained sweet spot prediction model to obtain the sweet spot area of the target exploration area.
[0083] This embodiment provides a method for predicting sweet spots in tight sandstone reservoirs. This method derives multiple candidate parameters from a basic elastic parameter system to construct an information-rich feature space. Then, it uses the real sweet spot labels interpreted from well logging to perform sensitivity analysis on the candidate parameters, selecting a few sensitive elastic parameters that are most strongly associated with the sweet spot category. This quantitatively correlates geological understanding with geophysical response, providing high-quality input features with clear physical meaning and strong discriminative ability for subsequent machine learning models. This significantly improves model training efficiency, prediction accuracy, and generalization ability, ultimately achieving accurate and reliable prediction of sweet spots in complex tight reservoirs.
[0084] Based on the above embodiments, an explanation will be given using a certain region as an example.
[0085] 1. Select the measured shear wave well as the learning well, and use the Vp, Vs, and Den logging curves to calculate other elastic parameter curves such as P-wave impedance, S-wave impedance, Young's modulus, Poisson's ratio, Lamé constant, shear modulus, P-wave / S-wave velocity ratio, and bulk modulus using elastic parameter calculation formulas. 2. Use the logging interpretation results curves of the learning wells, i.e., the permeability curves, to establish sweet spot type labels and further subdivide them into categories I, II-1, II-2, III-1, and III-2. 3. Correlation analysis is used to screen elastic parameters that are sensitive to reservoir sweet spots, so as to avoid interference from other elastic parameters with weak correlation to the model; 4. Use the selected elastic parameter curves that are sensitive to reservoir sweet spots as feature data and the corresponding sweet spot labels as target variables to construct a training sample set; 5. The Bootstrap method is used to sample with replacement to generate multiple training subsets. When splitting at a node, each decision tree randomly selects some features for optimal splitting to generate T decision trees, forming a random forest. The out-of-bag error is calculated using out-of-bag data to evaluate the model performance. 6. Obtain a sweet spot prediction model that can predict reservoir sweet spots; 7. Optimize the original CRP gathers in this region by performing noise suppression, amplitude compensation, and amplitude preservation processing to obtain high-quality gather data; 8. Using the Aki-Richards approximation equation and combining it with the Vp, Vs, and Den logging curves, pre-stack simultaneous inversion is performed to obtain the Vp, Vs, and Den data volumes. 9. Using the obtained Vp, Vs, and Den data, and employing the elastic parameter calculation formula and sweet spot sensitivity correlation analysis, the reservoir sweet spot sensitivity elastic parameter data body is obtained. 10. Finally, using the obtained three-dimensional pre-stack inversion elastic parameter data volume, the trained dessert prediction model is input one track at a time to generate a three-dimensional dessert data volume, thereby determining the dessert area of the region.
[0086] like Figure 5 and Figure 6 As shown, Figure 5 This is a comparison chart of well bypass prediction for dessert prediction provided in Embodiment 2 of this application. Figure 6 This is a three-dimensional intersection map for dessert prediction provided in Embodiment 2 of this application. From... Figure 5 It can be seen that, based on the logging curves of wells 1 and 2, the actual sweet spot label is almost identical to the sweet spot predicted label obtained by the sweet spot prediction model. From Figure 6As can be seen, when a three-dimensional coordinate system is established with density, gamma and P-wave velocity ratio as coordinates, the left figure is the intersection map of the true sweet spot labels and the right figure is the intersection map of the predicted sweet spot labels. The clustering areas of the true sweet spot labels and the predicted sweet spot labels are basically consistent, especially the sweet spot prediction model has a consistency rate of 85% for high-quality reservoirs.
[0087] In addition, this application also utilizes traditional inversion methods to predict the dessert area of the region, such as... Figures 7a to 7c As shown, Figure 7a This is an impedance inversion cross-sectional view provided in Embodiment 2 of this application. Figure 7b This is a pre-stack density inversion profile provided in Embodiment 2 of this application. Figure 7c This is a pre-stack LR coefficient inversion profile provided in Embodiment 2 of this application.
[0088] from Figure 7a It can be seen that while impedance inversion can delineate large stratigraphic frameworks, it easily confuses tight layers with effective reservoirs. From Figure 7b It can be seen that pre-stack density inversion provides density information of rocks and has a certain response to gas content, but the ambiguity of a single parameter is strong, and its sensitivity to lithological changes is poor. From Figure 7c It can be seen that the pre-stack LR coefficient inversion is more sensitive to lithology and fluids, but in the context of complex and dense sandstone, local artifacts or discontinuities may still occur, making it difficult to form a complete sweet spot interconnection.
[0089] Figure 8 This is a cross-sectional view based on a sweet spot prediction model provided in Embodiment 2 of this application. This cross-sectional view is obtained by predicting the sweet spot of tight sandstone reservoirs in this region using the sweet spot prediction method provided in this application; that is, it is obtained by first performing pre-stack inversion and then inputting the inversion results into the sweet spot prediction model. Figure 8 As shown, compared with traditional inversion methods, the profile obtained by the sweet spot prediction method for tight sandstone reservoirs provided in this application exhibits strong spatial continuity and stratification, identifies the lateral distribution pattern of sweet spots, and can effectively connect isolated sweet spots to guide well placement.
[0090] Example 3 Figure 9 This is a schematic diagram of a sweet spot prediction device for tight sandstone reservoirs provided in Embodiment 3 of this application. Figure 9 As shown, the device includes: The data acquisition module 310 is used to acquire the CRP gather of the target exploration area and the logging curves of the actual drilled wells in the target exploration area. The data volume calculation module 320 is used to perform pre-stack simultaneous inversion based on the CRP gather and the logging curve to determine the basic elastic parameter data volume of the target exploration area; Sensitivity analysis module 330 is used to perform sweet spot sensitivity analysis based on the basic elastic parameter data body to determine the sensitive elastic parameter data body; The sweet spot prediction module 340 is used to input the sensitive elastic parameter data into a pre-trained sweet spot prediction model to obtain the sweet spot of the target exploration area.
[0091] The sweet spot prediction device for tight sandstone reservoirs provided in this application acquires CRP gathers and logging curves from actual drilled wells in the target exploration area; performs pre-stack simultaneous inversion based on the CRP gathers and logging curves to determine the basic elastic parameter data volume of the target exploration area; conducts sweet spot sensitivity analysis based on the basic elastic parameter data volume to determine the sensitive elastic parameter data volume; and inputs the sensitive elastic parameter data volume into the sweet spot prediction model to obtain the sweet spot of the target exploration area. This technical solution obtains a multi-dimensional elastic parameter data volume through pre-stack inversion, combines well data-driven sensitivity analysis and a sweet spot prediction model, and integrates elastic parameters with clear physical meaning with intelligent algorithms capable of characterizing complex nonlinear relationships. This ensures that the sweet spot prediction model has a rock physics basis, improving the accuracy and reliability of sweet spot prediction.
[0092] Furthermore, the sensitivity analysis module 330 includes: The candidate elastic parameter calculation unit is used to determine the candidate elastic parameter data body based on the basic elastic parameter data body. The sensitive elastic parameter screening unit is used to perform sweet spot sensitivity analysis on each candidate elastic parameter data body based on the interpretation result curve of the well logging curve, and to determine the sensitive elastic parameter data body.
[0093] Furthermore, the sensitive elastic parameter filtering unit includes: A dessert type label creation sub-unit is used to create dessert type labels based on the interpretation result curve of the well logging curve; The sensitive elastic parameter screening subunit is used to perform correlation analysis on the interpretation curve of the well logging curve and each of the candidate elastic parameter data volumes based on the dessert type label, and to determine the sensitive elastic parameter data volume.
[0094] Furthermore, the candidate elastic parameter data body includes at least the longitudinal wave impedance data body, the transverse wave impedance data body, the Young's modulus data body, the Poisson's ratio data body, the Lamé constant data body, the shear modulus data body, the longitudinal and transverse wave velocity ratio data body, and the bulk modulus data body.
[0095] Furthermore, the data volume calculation module 320 includes: The reflectivity data volume determination unit is used to perform pre-stack simultaneous inversion based on the CRP gather and the AVO approximation formula to determine the basic elastic parameter reflectivity data volume of the target exploration area. The elastic parameter data volume determination unit is used to perform integral recursion on the basic elastic parameter reflectivity data volume based on the well logging curve to obtain the basic elastic parameter data volume of the target exploration area.
[0096] Furthermore, the training process of the dessert prediction model includes: A training sample set is obtained, and several training subsets are generated by sampling with replacement from the training sample set using the Bootstrap sampling method; wherein, the training sample set includes the sensitivity elasticity parameter curve of the learning well and dessert type labels; Each of the training subsets is input into the random forest model for parallel training to obtain multiple decision trees; wherein, when training a single decision tree, a portion of parameters are randomly selected from all sensitive elastic parameters at each node to find the optimal splitting rule; In response to the training termination condition, a dessert prediction model is obtained.
[0097] Furthermore, the device also includes: The gather preprocessing module is used to perform noise suppression, amplitude compensation, and amplitude preservation processing on the CRP gathers after obtaining the CRP gathers of the target exploration area, so as to obtain high-quality gathers.
[0098] The sweet spot prediction device for tight sandstone reservoirs provided in this application embodiment can execute the sweet spot prediction method for tight sandstone reservoirs provided in any embodiment of this application, and has the corresponding functional modules and beneficial effects of the method execution.
[0099] Example 4 Figure 10 A schematic diagram of the structure of a device 10 that can be used to implement embodiments of this application is shown. The device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the application described and / or claimed herein.
[0100] like Figure 10As shown, device 10 includes at least one processor 11 and a memory, such as read-only memory (ROM) 12, random access memory (RAM) 13, etc., communicatively connected to at least one processor 11. The memory stores computer programs executable by at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded into the RAM 13 from storage unit 18. The RAM 13 may also store various programs and data required for the operation of device 10. The processor 11, ROM 12, and RAM 13 are interconnected via bus 14. Input / output (I / O) interface 15 is also connected to bus 14.
[0101] Multiple components in device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of monitors, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.
[0102] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, central processing unit (CPU), graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as sweet spot prediction methods for tight sandstone reservoirs.
[0103] In some embodiments, the sweet spot prediction method for tight sandstone reservoirs may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the sweet spot prediction method for tight sandstone reservoirs described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the sweet spot prediction method for tight sandstone reservoirs by any other suitable means (e.g., by means of firmware).
[0104] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.
[0105] Computer programs used to implement the methods of this application may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.
[0106] In the context of this application, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium can be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.
[0107] To provide interaction with a user, the systems and techniques described herein can be implemented on a device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).
[0108] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.
[0109] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.
[0110] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication unit 19, or installed from storage unit 18, or installed from ROM 12. When the computer program is executed by processor 11, it performs the functions defined in the methods of embodiments of this application.
[0111] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this application can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this application can be achieved, and this is not limited herein.
[0112] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.
Claims
1. A method for predicting sweet spots in tight sandstone reservoirs, characterized in that, include: Obtain CRP gathers for the target exploration area, as well as logging curves of actual drilled wells within the target exploration area; Based on the CRP gather and the logging curves, pre-stack simultaneous inversion is performed to determine the basic elastic parameter data volume of the target exploration area. Based on the aforementioned basic elastic parameter data volume, a sensitivity analysis of the sweet spot area is performed to determine the sensitive elastic parameter data volume; The sensitive elasticity parameter data volume is input into a pre-trained sweet spot prediction model to obtain the sweet spot area of the target exploration area.
2. The method for predicting sweet spots in tight sandstone reservoirs according to claim 1, characterized in that, The step of performing dessert sensitivity analysis based on the basic elastic parameter data body to determine the sensitive elastic parameter data body includes: Based on the aforementioned basic elastic parameter data body, candidate elastic parameter data bodies are determined; Based on the interpretation results of the well logging curves, a sweet spot sensitivity analysis is performed on each of the candidate elastic parameter data volumes to determine the sensitive elastic parameter data volumes.
3. The method for predicting sweet spots in tight sandstone reservoirs according to claim 2, characterized in that, The step of performing sweet spot sensitivity analysis on each candidate elastic parameter data volume based on the interpretation results of the well logging curves to determine the sensitive elastic parameter data volume includes: Based on the interpretation results of the well logging curves, establish dessert type labels; Based on the dessert type tags, correlation analysis is performed on the interpretation results curves of the well logging curves and the candidate elastic parameter data volumes to determine the sensitive elastic parameter data volumes.
4. The method for predicting sweet spots in tight sandstone reservoirs according to claim 2, characterized in that, The candidate elastic parameter data volume includes at least the longitudinal wave impedance data volume, the transverse wave impedance data volume, the Young's modulus data volume, the Poisson's ratio data volume, the Lamé constant data volume, the shear modulus data volume, the longitudinal and transverse wave velocity ratio data volume, and the bulk modulus data volume.
5. The method for predicting sweet spots in tight sandstone reservoirs according to claim 1, characterized in that, The step of performing pre-stack simultaneous inversion based on the CRP gather and the logging curves to determine the basic elastic parameter data volume of the target exploration area includes: Based on the CRP gather, the AVO approximation formula is used to perform pre-stack simultaneous inversion to determine the basic elastic parameter reflectivity data volume of the target exploration area; Based on the well logging curves, the basic elastic parameter reflectivity data volume is integrally recursively calculated to obtain the basic elastic parameter data volume of the target exploration area.
6. The method for predicting sweet spots in tight sandstone reservoirs according to claim 1, characterized in that, The training process of the dessert prediction model includes: A training sample set is obtained, and several training subsets are generated by sampling with replacement from the training sample set using the Bootstrap sampling method; wherein, the training sample set includes the sensitivity elasticity parameter curve of the learning well and dessert type labels; Each of the training subsets is input into the random forest model for parallel training to obtain multiple decision trees; wherein, when training a single decision tree, a portion of parameters are randomly selected from all sensitive elastic parameters at each node to find the optimal splitting rule; In response to the training termination condition, a dessert prediction model is obtained.
7. The method for predicting sweet spots in tight sandstone reservoirs according to claim 1, characterized in that, After obtaining the CRP gather for the target exploration area, the method further includes: The CRP gather is subjected to noise suppression, amplitude compensation, and amplitude preservation processing to obtain a high-quality gather.
8. A sweet spot prediction device for tight sandstone reservoirs, characterized in that, The device includes: The data acquisition module is used to acquire CRP gathers of the target exploration area and logging curves of actual drilled wells in the target exploration area; The data volume calculation module is used to perform pre-stack simultaneous inversion based on the CRP gather and the logging curves to determine the basic elastic parameter data volume of the target exploration area; The sensitivity analysis module is used to perform sweet spot sensitivity analysis based on the basic elastic parameter data body to determine the sensitive elastic parameter data body; The sweet spot prediction module is used to input the sensitive elasticity parameter data into a pre-trained sweet spot prediction model to obtain the sweet spot of the target exploration area.
9. An electronic device, characterized in that, The device includes: At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the sweet spot prediction method for tight sandstone reservoirs according to any one of claims 1-7.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the sweet spot prediction method for tight sandstone reservoirs according to any one of claims 1-7.