Filling mode feature prediction method and device, computer equipment and storage medium

Abstract

The invention provides a prediction method and device for filling mode features, computer equipment and a storage medium. The method comprises the steps of obtaining cave filling feature values; constructing a carbonate rock cave type reservoir model by using the cave filling characteristic values; carrying out forward modeling simulation on the carbonate rock cave type reservoir model, and determining waveform characteristic parameters; the waveform characteristic parameters are optimized through an automatic waveform extraction method, and sensitive parameters are obtained; constructing a prediction model based on the sensitive parameters and the cave filling characteristic values; and utilizing the prediction model to obtain predicted filling mode characteristics. According to the method, the prediction model is established based on the relationship between the sensitive parameters and the cave filling characteristic values, and the sensitive parameters are obtained by optimizing the waveform characteristic parameters, that is, the prediction model is obtained according to the waveform characteristic parameters after datamation optimization, so that the accuracy of the prediction model is improved, and the prediction accuracy is improved. Therefore, the analysis precision of the filling mode is improved, and the drilling rate is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of seismic data interpretation, and particularly to a method, device, computer device, and storage medium for predicting filling pattern features. Background Art

[0002] Vuggy carbonate reservoirs are an important type of oil reservoir, and the determination of their filling patterns is of great significance for oil and gas field exploration and development. At present, the research on the filling patterns of vuggy carbonate reservoirs still stays at the traditional waveform similarity comparison prediction method, and the analysis of filling patterns is relatively rough, thus reducing the drilling encounter rate.

[0003] In response to this, a method based on inductive learning is proposed to classify various waveform features that can reflect filling characteristics and determine the threshold for judging whether a cave is filled. The proposal of this technology can achieve a breakthrough in the research on the filling patterns of carbonate cave reservoirs, greatly improve the drilling encounter rate of carbonate reservoirs, and enhance the exploration and development efficiency of oil and gas fields. Summary of the Invention

[0004] Based on this, in view of the above technical problems, it is necessary to provide a method, device, computer device, and storage medium for predicting filling pattern features.

[0005] A method for predicting filling pattern features includes:

[0006] Obtain cave filling characteristic values;

[0007] Use the cave filling characteristic values to construct a carbonate cave reservoir model;

[0008] Perform forward modeling on the carbonate cave reservoir model to determine waveform characteristic parameters;

[0009] Through a waveform automatic extraction method, optimize the waveform characteristic parameters to obtain sensitive parameters;

[0010] Based on the sensitive parameters and the cave filling characteristic values, construct a prediction model;

[0011] Use the prediction model to perform waveform analysis to obtain predicted filling pattern features.

[0012] In one embodiment, the sensitive parameters include peak amplitude value, trough amplitude value, and amplitude ratio;

[0013] The step of constructing a prediction model based on the sensitive parameters and the cave filling characteristic values includes:

[0014] Based on the peak amplitude value, the trough amplitude value, the amplitude ratio, and the cave filling characteristic values, construct a prediction model.

[0015] In one embodiment, the step of obtaining a prediction model based on the sensitive parameter and the cave filling eigenvalue includes:

[0016] Using an inductive learning method, training the sensitive parameter to establish a correspondence between the sensitive parameter and the cave filling eigenvalue;

[0017] Using the correspondence to construct the prediction model.

[0018] In one embodiment, the cave filling eigenvalue includes the sandstone-mudstone filling speed, the cave reservoir speed, and the surrounding rock speed;

[0019] The step of obtaining the cave filling eigenvalue includes:

[0020] Obtaining the sandstone-mudstone filling speed, the cave reservoir speed, and the surrounding rock speed;

[0021] The step of constructing a carbonate rock cave type reservoir model using the cave filling eigenvalue includes:

[0022] Using the sandstone-mudstone filling speed, the cave reservoir speed, and the surrounding rock speed to construct the carbonate rock cave type reservoir model.

[0023] In one embodiment, the step of performing waveform analysis using the prediction model to obtain predicted filling pattern characteristics includes:

[0024] Obtaining actual waveform characteristic values;

[0025] Using the prediction model to analyze the actual waveform characteristic values to obtain predicted filling pattern characteristics.

[0026] In one embodiment, the step of obtaining actual waveform characteristic values includes:

[0027] Obtaining original seismic data;

[0028] Performing waveform extraction on the original seismic data to obtain an actual waveform;

[0029] Performing feature extraction on the actual waveform to obtain actual waveform characteristic values.

[0030] In one embodiment, the step of performing forward modeling on the carbonate rock cave type reservoir model to determine waveform characteristic parameters includes:

[0031] Using a Ricker wavelet to perform forward modeling on the carbonate rock cave type reservoir model to determine waveform characteristic parameters.

[0032] A prediction device for filling pattern characteristics includes:

[0033] A filling feature value acquisition module for acquiring the cave filling feature value;

[0034] A reservoir model construction module for constructing a carbonate rock cave type reservoir model by using the cave filling feature value;

[0035] A waveform feature parameter determination module for performing forward modeling on the carbonate rock cave type reservoir model to determine waveform feature parameters;

[0036] A sensitive parameter determination module for optimizing the waveform feature parameters by a waveform automatic extraction method to obtain sensitive parameters;

[0037] A prediction model construction module for constructing a prediction model based on the sensitive parameters and the cave filling feature value;

[0038] A prediction feature determination module for performing waveform analysis by using the prediction model to obtain predicted filling pattern features.

[0039] In one embodiment, the prediction model construction module includes:

[0040] A correspondence relationship establishment unit for training the sensitive parameters by using an inductive learning method to establish a correspondence relationship between the sensitive parameters and the cave filling feature value;

[0041] A prediction model construction unit for constructing the prediction model by using the correspondence relationship.

[0042] In one embodiment, the filling feature value acquisition module is further configured to:

[0043] Acquire the sand-shale filling speed, the cave reservoir speed, and the surrounding rock speed;

[0044] The reservoir model construction module is further configured to:

[0045] Construct the carbonate rock cave type reservoir model by using the sand-shale filling speed, the cave reservoir speed, and the surrounding rock speed.

[0046] In one embodiment, the prediction feature determination module includes:

[0047] A waveform feature value acquisition unit for acquiring actual waveform feature values;

[0048] A prediction feature determination unit for analyzing the actual waveform feature values by using the prediction model to obtain predicted filling pattern features.

[0049] In one embodiment, the waveform feature value acquisition unit includes:

[0050] A data acquisition subunit, configured to acquire original seismic data;

[0051] An actual waveform acquisition subunit, configured to perform waveform extraction on the original seismic data to obtain an actual waveform;

[0052] A waveform eigenvalue acquisition subunit, configured to perform feature extraction on the actual waveform to obtain actual waveform eigenvalues.

[0053] In one embodiment, the waveform feature parameter determination module is further configured to:

[0054] Use a Ricker wavelet to perform forward modeling on a carbonate rock cave reservoir model to determine waveform feature parameters.

[0055] A computer device, including a memory and a processor, the memory stores a computer program, and is characterized in that when the processor executes the computer program, the following steps are implemented:

[0056] Obtain cave filling eigenvalues;

[0057] Use the cave filling eigenvalues to construct a carbonate rock cave reservoir model;

[0058] Perform forward modeling on the carbonate rock cave reservoir model to determine waveform feature parameters;

[0059] Through a waveform automatic extraction method, optimize the waveform feature parameters to obtain sensitive parameters;

[0060] Based on the sensitive parameters and the cave filling eigenvalues, construct a prediction model;

[0061] Use the prediction model to perform waveform analysis to obtain predicted filling pattern features.

[0062] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0063] Based on the wave peak amplitude value, the wave trough amplitude value, the amplitude ratio, and the cave filling eigenvalues, construct a prediction model.

[0064] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0065] Use an inductive learning method to train the sensitive parameters to establish a correspondence between the sensitive parameters and the cave filling eigenvalues;

[0066] Use the correspondence to construct the prediction model.

[0067] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0068] Obtain the sandstone-mudstone filling velocity, the cave reservoir velocity, and the surrounding rock velocity;

[0069] Use the sandstone-mudstone filling velocity, the cave reservoir velocity, and the surrounding rock velocity to construct the carbonate cave reservoir model.

[0070] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0071] Obtain the actual waveform characteristic value;

[0072] Use the prediction model to analyze the actual waveform characteristic value to obtain the predicted filling pattern characteristic.

[0073] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0074] Obtain the original seismic data;

[0075] Perform waveform extraction on the original seismic data to obtain the actual waveform;

[0076] Perform feature extraction on the actual waveform to obtain the actual waveform characteristic value.

[0077] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0078] Use the Ricker wavelet to perform forward modeling on the carbonate cave reservoir model to determine the waveform characteristic parameters.

[0079] A computer-readable storage medium stores a computer program thereon, and when the computer program is executed by a processor, the following steps are implemented:

[0080] Obtain the cave filling characteristic value;

[0081] Use the cave filling characteristic value to construct the carbonate cave reservoir model;

[0082] Perform forward modeling on the carbonate cave reservoir model to determine the waveform characteristic parameters;

[0083] Through the waveform automatic extraction method, optimize the waveform characteristic parameters to obtain the sensitive parameters;

[0084] Based on the sensitive parameters and the cave filling characteristic value, construct a prediction model;

[0085] Use the prediction model to perform waveform analysis to obtain the predicted filling pattern characteristic.

[0086] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0087] Based on the peak amplitude value, the trough amplitude value, the amplitude ratio, and the cave filling feature value, construct a prediction model.

[0088] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0089] Using an inductive learning method, train the sensitive parameters to establish the corresponding relationship between the sensitive parameters and the cave filling feature value;

[0090] Using the corresponding relationship, construct the prediction model.

[0091] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0092] Obtain the sandstone-mudstone filling velocity, the cave reservoir velocity, and the surrounding rock velocity;

[0093] Using the sandstone-mudstone filling velocity, the cave reservoir velocity, and the surrounding rock velocity, construct the carbonate rock cave reservoir model.

[0094] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0095] Obtain the actual waveform feature value;

[0096] Using the prediction model, analyze the actual waveform feature value to obtain the predicted filling pattern feature.

[0097] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0098] Obtain the original seismic data;

[0099] Extract the waveform from the original seismic data to obtain the actual waveform;

[0100] Extract the features of the actual waveform to obtain the actual waveform feature value.

[0101] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0102] Use a Ricker wavelet to perform forward modeling on the carbonate rock cave reservoir model to determine the waveform characteristic parameters.

[0103] Traditional prediction models are established based on waveform similarity and do not express waveform features numerically, resulting in low analysis accuracy of filling patterns. However, for the prediction method, device, computer equipment, and storage medium of filling pattern features in this application, the prediction model is established based on the relationship between sensitive parameters and cave filling feature values, and the sensitive parameters are preferably obtained from waveform feature parameters. That is to say, the prediction model of this application is obtained based on the optimized waveform feature parameters in a digital form, which improves the accuracy of the prediction model of this application, thereby improving the analysis accuracy of filling patterns and further increasing the drilling encounter rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0104] Figure 1 It is a schematic flowchart of the prediction method for filling pattern features in an embodiment;

[0105] Figure 2 It is a schematic flowchart of the method for predicting cave filling patterns based on inductive learning in an embodiment;

[0106] Figure 3 It is a structural block diagram of the prediction device for filling pattern features in an embodiment;

[0107] Figure 4 It is an internal structure diagram of a computer device in an embodiment;

[0108] Figure 5 For Figure 2 the geological model of changing the thickness of sandstone and mudstone filling in the method for predicting cave filling patterns based on inductive learning shown;

[0109] Figure 6 For Figure 2 the geological model of changing the physical properties of sandstone and mudstone filling in the method for predicting cave filling patterns based on inductive learning shown;

[0110] Figure 7a From Figure 5 the absolute value of the amplitude obtained from the geological model of changing the thickness of sandstone and mudstone filling shown;

[0111] Figure 7b From Figure 5 the amplitude ratio obtained from the geological model of changing the thickness of sandstone and mudstone filling shown;

[0112] Figure 8a From Figure 6 the absolute value of the amplitude obtained from the geological model of changing the physical properties of sandstone and mudstone filling shown;

[0113] Figure 8b From Figure 6 the amplitude ratio obtained from the geological model of changing the physical properties of sandstone and mudstone filling shown;

[0114] Figure 9 For Figure 2 The induction learning flowchart adopted by the method for predicting the cave filling pattern based on induction learning shown in the figure. Detailed implementation manners

[0115] In order to make the objectives, technical solutions and advantages of the present application more clear and understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0116] Embodiment 1

[0117] In this embodiment, as Figure 1 shown, a method for predicting filling pattern features is provided, which includes:

[0118] Step 110, obtaining cave filling feature values;

[0119] Step 120, using the cave filling feature values to construct a carbonate rock cave reservoir model;

[0120] Step 130, performing forward modeling on the carbonate rock cave reservoir model to determine waveform feature parameters;

[0121] Step 140, optimizing the waveform feature parameters through a waveform automatic extraction method to obtain sensitive parameters;

[0122] Step 150, constructing a prediction model based on the sensitive parameters and the cave filling feature values;

[0123] Step 160, performing waveform analysis using the prediction model to obtain predicted filling pattern features.

[0124] Traditional analysis only looks at the waveform similarity and does not express the waveform features in a numerical way. The present application is different from the traditional method in this regard.

[0125] The traditional prediction model is established based on waveform similarity and does not express the waveform features in a numerical way, resulting in a low analysis accuracy of the filling pattern. However, for the method for predicting the filling pattern features of the present application, its prediction model is established based on the relationship between the sensitive parameters and the cave filling feature values, and the sensitive parameters are obtained by optimizing the waveform feature parameters. That is to say, the prediction model of the present application is obtained based on the digital and optimized waveform feature parameters, which improves the accuracy of the prediction model of the present application, thereby improving the analysis accuracy of the filling pattern and further increasing the drilling encounter rate.

[0126] In one embodiment, the sensitive parameters include peak amplitude value, trough amplitude value and amplitude ratio;

[0127] The steps of constructing a prediction model based on the sensitive parameters and the cave filling characteristic values include:

[0128] Construct a prediction model based on the peak amplitude value, the trough amplitude value, the amplitude ratio, and the cave filling characteristic value.

[0129] In this embodiment, when the cave is not filled, the peak amplitude value and the trough amplitude value are basically equivalent. When the filling increases, the difference between the peak amplitude value and the trough amplitude value increases, and the degree of cave filling can be roughly characterized by the difference between the maximum peak amplitude value and the maximum trough amplitude value.

[0130] In one embodiment, the steps of obtaining a prediction model based on the sensitive parameters and the cave filling characteristic values include:

[0131] Use the inductive learning method to train the sensitive parameters and establish the corresponding relationship between the sensitive parameters and the cave filling characteristic values;

[0132] Use the corresponding relationship to construct the prediction model.

[0133] In this embodiment, as Figure 9 shown, inductive learning is carried out using an inductive learning flowchart.

[0134] In this embodiment, the sensitive parameters are the optimized waveform characteristic parameters, which improves the accuracy of the prediction model constructed from the sensitive parameters.

[0135] In one embodiment, the cave filling characteristic values include the sandstone-mudstone filling speed, the cave reservoir speed, and the surrounding rock speed;

[0136] The steps of obtaining the cave filling characteristic values include:

[0137] Obtain the sandstone-mudstone filling speed, the cave reservoir speed, and the surrounding rock speed;

[0138] The steps of constructing a carbonate rock cave type reservoir model using the cave filling characteristic values include:

[0139] Use the sandstone-mudstone filling speed, the cave reservoir speed, and the surrounding rock speed to construct the carbonate rock cave type reservoir model.

[0140] In this embodiment, the carbonate rock cave type reservoir model constructed using the sandstone-mudstone filling speed, the cave reservoir speed, and the surrounding rock speed has a relatively high correlation between the waveform characteristic parameters determined in subsequent forward modeling and the cave filling characteristic values.

[0141] In one embodiment, the step of performing waveform analysis using the prediction model to obtain the predicted filling pattern features includes:

[0142] Obtain the actual waveform feature values;

[0143] Use the prediction model to analyze the actual waveform feature values to obtain the predicted filling pattern features.

[0144] In this embodiment, the prediction model performs waveform analysis on the actual waveform to obtain the actual waveform feature values, and then analyzes the actual waveform feature values to obtain the predicted filling pattern features.

[0145] In one embodiment, the step of obtaining the actual waveform feature values includes:

[0146] Obtain the original seismic data;

[0147] Perform waveform extraction on the original seismic data to obtain the actual waveform;

[0148] Perform feature extraction on the actual waveform to obtain the actual waveform feature values.

[0149] In this embodiment, first perform waveform extraction on the original seismic data to obtain the actual waveform, which is convenient for observing the actual waveform for feature extraction, and then obtain the actual waveform feature values according to the extracted features, that is, dataize the features.

[0150] In one embodiment, the step of performing forward modeling on the carbonate rock cave reservoir model to determine the waveform feature parameters includes:

[0151] Use a Ricker wavelet to perform forward modeling on the carbonate rock cave reservoir model to determine the waveform feature parameters.

[0152] In this embodiment, using a Ricker wavelet to perform forward modeling on the carbonate rock cave reservoir model can accurately simulate the propagation and scattering process of seismic signals, and has high flexibility. The Ricker wavelet can be adjusted as needed, including amplitude, frequency, phase, etc., to adapt to different seismic exploration requirements. And since the Ricker wavelet is a mathematical function, it can be reused in different seismic exploration projects to improve the simulation efficiency, with high repeatability, and the Ricker wavelet has wide applicability and high efficiency.

[0153] It should be understood that although Figure 1 the steps in the flowchart are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, Figure 1At least a part of the steps may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed and completed at the same moment, but can be executed at different moments, and the execution order of these sub-steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or sub-steps or stages of other steps.

[0154] Embodiment 2

[0155] In this embodiment, a method for predicting a cave filling pattern based on inductive learning is provided, which includes:

[0156] First step, for the target to be studied, establish different models that can reflect the filling characteristics to obtain a theoretical model that conforms to the actual situation;

[0157] Second step, for the waveform reflection characteristics of the theoretical model, extract key parameters, and evaluate and optimize whether the key parameters can reflect the filling characteristics.

[0158] Third step, as Figure 9 shown, use the multi-concept inductive learning method to train the selected waveform feature samples, and establish the corresponding relationship between the waveform feature parameters and the filling pattern in the work area.

[0159] Fourth step, use the prediction model to analyze the filling patterns of other carbonate rock cave reservoirs in the work area, predict the filling methods, and improve the reservoir drilling encounter rate.

[0160] The method for predicting the cave filling pattern based on inductive learning in this embodiment includes four parts: forward model establishment, key parameter optimization and extraction, application of the inductive learning method for actual data, and prediction of the cave filling degree of actual data. The forward model in this embodiment is the carbonate rock cave reservoir model in the above embodiment. The key parameters in this embodiment are the waveform feature parameters in the above embodiment.

[0161] In this embodiment, for carbonate rock cave reservoirs, extract waveform feature parameters that can characterize the cave filling pattern; use the inductive learning method for training to establish the relationship between the filling position of carbonate rock cave reservoirs and the feature parameters for the prediction of actual data. The filling position of carbonate rock cave reservoirs and the feature parameters in this embodiment are the cave filling characteristic values in the above embodiment.

[0162] Traditional analysis only looks at the waveform similarity and does not express the waveform features in a numerical way. This application is different from the traditional method in this regard.

[0163] Embodiment 3

[0164] In this embodiment, another method for predicting the cave filling pattern based on inductive learning is provided, which includes:

[0165] 1) Establish a model that conforms to the situation of a cave-type reservoir filled with carbonate rocks;

[0166] 2) Conduct forward modeling using a Ricker wavelet;

[0167] 3) Extract waveform characteristic parameters, such as peak amplitude value, trough amplitude value, and relative amplitude value, etc.;

[0168] 4) Use an inductive method to relate the waveform characteristic parameters to the marked filling development characteristics and establish a relationship model;

[0169] 5) For actual seismic data, process the original seismic data and extract waveform characteristic values;

[0170] 6) Use the established relationship model to analyze the waveforms extracted from the actual data, predict the filling pattern characteristics of the actual work area, and improve the drilling encounter rate of carbonate cave-type reservoirs.

[0171] The waveform characteristic parameters in this embodiment are preferred waveform characteristic parameters, that is, the sensitive parameters in the above embodiment. The model that conforms to the situation of a cave-type reservoir filled with carbonate rocks in this embodiment is the carbonate cave-type reservoir model in the above embodiment. The relative amplitude value in this embodiment is the amplitude ratio in the above embodiment. The waveform characteristic value in this embodiment is the waveform characteristic value in the above embodiment. The relationship model in this embodiment is the prediction model in the above embodiment. The filling pattern characteristics in this embodiment are the predicted filling pattern characteristics in the above embodiment. The marked filling development characteristics in this embodiment are the cave filling characteristic values in the above embodiment.

[0172] The innovation of this application lies in the first application of the inductive learning method to the identification and analysis of carbonate cave filling characteristics, laying a foundation for subsequent research. This application uses the inductive learning method to train the waveform characteristic samples obtained from the forward model, thereby obtaining a prediction model, and then analyzing the actual data to improve the accuracy and reliability of predicting the filling degree of carbonate cave reservoirs.

[0173] In the application in a certain exploration area in the northwest, the process designed by the present invention can relatively accurately predict the filling degree of carbonate cave reservoirs, with a high coincidence rate with the logging results, providing valuable research ideas and results for the fine description of carbonate cave reservoirs in this exploration area.

[0174] In this embodiment, for carbonate rock cavernous reservoirs, waveform characteristic parameters capable of characterizing the cavern filling pattern are extracted; an inductive learning method is used for training to establish the relationship between the filling position of carbonate rock cavernous reservoirs and the characteristic parameters for predicting actual data.

[0175] Traditional analysis only looks at the similarity of waveforms and does not express waveform characteristics in numerical form. This application is different from traditional methods in this regard.

[0176] The filling position of the carbonate rock cavernous reservoir and the characteristic parameters in this embodiment are the cavern filling characteristic values in the above embodiment. The forward modeling in this embodiment is the carbonate rock cavernous reservoir model in the above embodiment.

[0177] Embodiment Four

[0178] In this embodiment, as Figure 2 shown, another method for predicting the cavern filling pattern based on inductive learning is provided, which includes:

[0179] Forward modeling feature analysis or bead string feature analysis;

[0180] Automatic waveform extraction;

[0181] Determination of sensitive parameters;

[0182] Establishment of an inductive learning model;

[0183] Prediction and evaluation of actual work area data;

[0184] Optimization of target points.

[0185] In this embodiment, for carbonate rock cavernous reservoirs, waveform characteristic parameters capable of characterizing the cavern filling pattern are extracted; an inductive learning method is used for training to establish the relationship between the filling position of carbonate rock cavernous reservoirs and the characteristic parameters for predicting actual data. The filling position of the carbonate rock cavernous reservoir and the characteristic parameters in this embodiment are the cavern filling characteristic values in the above embodiment. The inductive learning model in this embodiment is the prediction model in the above embodiment.

[0186] Embodiment Five

[0187] In this embodiment, a method flow for predicting the cavern filling pattern based on inductive learning is provided, which includes:

[0188] First, a model conforming to the carbonate cave filling pattern is established. The Ricker wavelet is used for forward modeling to obtain waveform characteristics, and characteristic values such as the peak amplitude value, trough amplitude value, and relative amplitude value of the waveform, which can reflect the cave filling characteristics, are extracted. By using the inductive learning method, the correlation between waveform characteristic parameters and cave filling characteristic values is obtained and applied to the actual work area, and finally the prediction of cave filling characteristics is obtained. This method can relatively accurately predict the filling degree of cavernous carbonate reservoirs, has a high coincidence rate with logging results, provides valuable research ideas and results for the fine description of cavernous carbonate reservoirs, and has good application prospects.

[0189] Aiming at the problem of unclear analysis of the characteristics of the carbonate cave filling pattern, the present invention establishes a set of waveform characteristic description processes for carbonate-filled caves based on the inductive learning method. Based on forward modeling, the present invention optimizes waveform characteristic parameters that can characterize the filling characteristics of cave-type reservoirs, and clarifies the connection between the extracted waveform response characteristics and the cave-type reservoir filling pattern. In actual seismic data, the drilled cave filling patterns with logging data are classified, and the thresholds between various types are determined by using the multi-concept inductive learning method, so as to predict the filling conditions of other caves in the same work area, laying a foundation for the subsequent research on cavernous carbonate reservoirs.

[0190] In this embodiment, for the cavernous carbonate reservoir, waveform characteristic parameters that can characterize the cave filling pattern are extracted; training is carried out by using the inductive learning method to establish the relationship between the filling position of the cavernous carbonate reservoir and the characteristic parameters for the prediction of actual data. The filling position and characteristic parameters of the cavernous carbonate reservoir in this embodiment are the cave filling characteristic values in the above embodiment. The relative amplitude value in this embodiment is the amplitude ratio in the above embodiment. The waveform characteristic value in this embodiment is the waveform characteristic value in the above embodiment. The characteristic value of the cave filling characteristic in this embodiment is the cave filling characteristic value in the above embodiment.

[0191] Embodiment Six

[0192] In a certain exploration area in the northwest, there is relatively little research on the filling pattern of carbonate karst caves, and the difficulty is relatively large. Aiming at the filling problem after the formation of carbonate karst caves, the present invention obtains the relationship between waveform characteristic parameters and the corresponding filling degree by establishing a geological model of the cave filling pattern and performing forward modeling, and uses the inductive learning method to establish the relevant connection, improving the accuracy and reliability of the prediction of cavernous carbonate reservoirs, and providing new ideas and solutions for the fine research of cavernous carbonate reservoirs.

[0193] In this embodiment, another method for predicting the cave filling pattern based on inductive learning is provided, which includes:

[0194] First, establish a corresponding geological body model according to the characteristics of cave filling. The filling speed of sandstone and mudstone is 3600 m / s, the velocity of the cave reservoir is 4700 m / s, and the velocity of the surrounding rock is 6200 m / s. Use a zero-phase Ricker wavelet for forward modeling. The wavelet frequency is 22 Hz to obtain the final forward model. According to well logging data, the greatest impact of cave filling on reservoir properties has two aspects: one is the thickness of the filling material, and the other is the physical properties of the filling. Using the method of controlling variables, two sets of geological body models are designed specifically, as Figure 5 shown. One set of models only changes the filling thickness of sandstone and mudstone, as Figure 6 shown. One set of models only changes the physical properties of the sandstone and mudstone filling. Extract the relative amplitude values of the wave peaks and wave troughs of the waveform. When the cave is not filled, the amplitudes of the wave peaks and wave troughs are basically equivalent. When the filling increases, the difference between the wave peaks and wave troughs increases. The amplitude difference between the maximum wave peak and the maximum wave trough can be used to roughly characterize the degree of cave filling. However, in specific cases, it is not possible to simply judge the cave filling mode based on the amplitude difference between the wave peaks and wave troughs because the waveform characteristics are also affected by the filling position of sandstone and mudstone, the filling combination method, the filling angle, and the physical properties of sandstone and mudstone, with a certain degree of non-uniqueness. This requires analysis relying on the method of inductive learning to study the probability of the development of cave-type reservoirs and the distribution of favorable reservoirs, etc. As Figures 7a to 8b shown, the waveform characteristics we have selected currently include wave peak amplitude, wave trough amplitude, wave peak-wave trough time difference, and amplitude ratio. According to the obtained results, the characterization effect of the ratio is better than that of the absolute value, and it can better reflect the degree of cave filling. Therefore, the wave peak-wave trough time difference and amplitude ratio are selected as the waveform characteristic sensitive parameters.

[0195] Use the inductive learning method for training to establish the relationship between the waveform ratio parameter and the filling position and physical properties of the filling material, obtain the prediction model, and thus achieve the fine characterization of the filling mode of carbonate rock cave-type reservoirs, providing a more reliable basis for target selection.

[0196] In this embodiment, for carbonate rock cavernous reservoirs, waveform characteristic parameters that can characterize the cavern filling pattern are extracted; an inductive learning method is used for training to establish the relationship between the filling position of carbonate rock cavernous reservoirs and the characteristic parameters, which is used for the prediction of actual data. The filling position of carbonate rock cavernous reservoirs and the characteristic parameters in this embodiment are the cavern filling characteristic values in the above embodiment. The forward model in this embodiment is the carbonate rock cavernous reservoir model in the above embodiment. The waveform ratio parameter in this embodiment is the sensitive parameter in the above embodiment. The fine characterization of the filling pattern of carbonate rock cavernous reservoirs in this embodiment is the predicted filling pattern characteristic in the above embodiment. The characteristic values of the filling material position and physical properties in this embodiment are the cavern filling characteristic values in the above embodiment. The geological body model in this embodiment is the carbonate rock cavernous reservoir model in the above embodiment. The waveform characteristic sensitive parameter in this embodiment is the sensitive parameter in the above embodiment. The filling degree in this embodiment is the cavern filling characteristic value in the above embodiment.

[0197] Embodiment Seven

[0198] In this embodiment, as Figure 3 shown, a prediction device for filling pattern characteristics is provided, including:

[0199] A filling characteristic value acquisition module 210, configured to acquire cavern filling characteristic values;

[0200] A reservoir model construction module 220, configured to construct a carbonate rock cavernous reservoir model by using the cavern filling characteristic values;

[0201] A waveform characteristic parameter determination module 230, configured to perform forward simulation on the carbonate rock cavernous reservoir model to determine waveform characteristic parameters;

[0202] A sensitive parameter determination module 240, configured to optimize the waveform characteristic parameters through a waveform automatic extraction method to obtain sensitive parameters;

[0203] A prediction model construction module 250, configured to construct a prediction model based on the sensitive parameters and the cavern filling characteristic values;

[0204] A prediction characteristic determination module 260, configured to perform waveform analysis by using the prediction model to obtain the predicted filling pattern characteristic.

[0205] In one embodiment, the prediction model construction module is further configured to:

[0206] Construct a prediction model based on the wave peak amplitude value, the wave valley amplitude value, the amplitude ratio, and the cavern filling characteristic values.

[0207] In this embodiment, the sensitive parameters include the peak amplitude value, the trough amplitude value, and the amplitude ratio.

[0208] In one embodiment, the prediction model construction module includes:

[0209] A correspondence establishing unit, configured to use an inductive learning method to train the sensitive parameters and establish a correspondence between the sensitive parameters and the cave filling characteristic values;

[0210] A prediction model construction unit, configured to construct the prediction model by using the correspondence.

[0211] In one embodiment, the filling characteristic value acquisition module is further configured to:

[0212] Obtain the sandstone-mudstone filling velocity, the cave reservoir velocity, and the surrounding rock velocity;

[0213] The reservoir model construction module is further configured to:

[0214] Construct the carbonate rock cave type reservoir model by using the sandstone-mudstone filling velocity, the cave reservoir velocity, and the surrounding rock velocity.

[0215] In this embodiment, the cave filling characteristic values include the sandstone-mudstone filling velocity, the cave reservoir velocity, and the surrounding rock velocity.

[0216] In one embodiment, the prediction feature determination module includes:

[0217] A waveform feature value acquisition unit, configured to acquire actual waveform feature values;

[0218] A prediction feature determination unit, configured to analyze the actual waveform feature values by using the prediction model to obtain predicted filling pattern features.

[0219] In one embodiment, the waveform feature value acquisition unit includes:

[0220] A data acquisition subunit, configured to acquire original seismic data;

[0221] An actual waveform acquisition subunit, configured to perform waveform extraction on the original seismic data to obtain an actual waveform;

[0222] A waveform feature value acquisition subunit, configured to perform feature extraction on the actual waveform to obtain actual waveform feature values.

[0223] In one embodiment, the waveform feature parameter determination module is further configured to:

[0224] Perform forward modeling simulation on the carbonate rock cave type reservoir model by using a Ricker wavelet to determine waveform feature parameters.

[0225] For the specific limitations of the prediction device for filling pattern features, reference can be made to the limitations of the prediction method for filling pattern features in the above text, which will not be elaborated here. Each unit in the above prediction device for filling pattern features can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above units can be embedded in the processor of the computer device in hardware form or be independent of it, or can be stored in the memory of the computer device in software form, so as to facilitate the processor to call and execute the operations corresponding to each of the above units.

[0226] Embodiment VIII

[0227] In this embodiment, a computer device is provided. Its internal structural diagram can be as Figure 4 shown. The computer device includes a processor, a memory, a network interface, a display screen, and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program, and a database is deployed on the non-volatile storage medium, and the database is used to store cave filling feature values and original seismic data. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with other computer devices on which application software is deployed. When the computer program is executed by the processor, it implements a prediction method for filling pattern features. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0228] Those skilled in the art can understand that Figure 4 the structure shown in

[0229] is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0230] Step 110, obtain cave filling feature values;

[0231] Step 120, use the cave filling feature values to construct a carbonate rock cave reservoir model;

[0232] Step 130, perform forward modeling on the carbonate rock cave-type reservoir model to determine waveform characteristic parameters;

[0233] Step 140, optimize the waveform characteristic parameters through a waveform automatic extraction method to obtain sensitive parameters;

[0234] Step 150, construct a prediction model based on the sensitive parameters and the cave filling characteristic values;

[0235] Step 160, perform waveform analysis using the prediction model to obtain the predicted filling pattern characteristics.

[0236] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0237] Construct a prediction model based on the wave peak amplitude value, the wave trough amplitude value, the amplitude ratio, and the cave filling characteristic value.

[0238] In this embodiment, the sensitive parameters include the wave peak amplitude value, the wave trough amplitude value, and the amplitude ratio.

[0239] In this embodiment, when the cave is not filled, the wave peak amplitude value and the wave trough amplitude value are basically equivalent. When the filling increases, the difference between the wave peak amplitude value and the wave trough amplitude value increases, and the difference between the maximum wave peak amplitude value and the maximum wave trough amplitude value can be used to roughly characterize the degree of cave filling.

[0240] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0241] Use the inductive learning method to train the sensitive parameters and establish the corresponding relationship between the sensitive parameters and the cave filling characteristic values;

[0242] Construct the prediction model using the corresponding relationship.

[0243] In this embodiment, the sensitive parameters are the optimized waveform characteristic parameters, which improves the accuracy of the prediction model constructed by the sensitive parameters.

[0244] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0245] Obtain the sand-shale filling speed, the cave reservoir speed, and the surrounding rock speed;

[0246] Construct the carbonate rock cave-type reservoir model using the sand-shale filling speed, the cave reservoir speed, and the surrounding rock speed.

[0247] In this embodiment, the cave filling characteristic values include the sand-shale filling speed, the cave reservoir speed, and the surrounding rock speed.

[0248] In this embodiment, the carbonate rock cavernous reservoir model constructed using the filling speed of sandstone and mudstone, the speed of the cavern reservoir, and the speed of the surrounding rock has a high correlation between the waveform characteristic parameters determined in subsequent forward modeling and the cavern filling characteristic values.

[0249] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0250] Obtain the actual waveform characteristic values;

[0251] Use the prediction model to analyze the actual waveform characteristic values to obtain the predicted filling pattern characteristics.

[0252] In this embodiment, the prediction model performs waveform analysis on the actual waveform to obtain the actual waveform characteristic values, and then analyzes the actual waveform characteristic values to obtain the predicted filling pattern characteristics.

[0253] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0254] Obtain the original seismic data;

[0255] Extract the waveform from the original seismic data to obtain the actual waveform;

[0256] Extract the characteristics of the actual waveform to obtain the actual waveform characteristic values.

[0257] In this embodiment, first, the waveform is extracted from the original seismic data to obtain the actual waveform, which is convenient for observing the actual waveform for feature extraction. Then, based on the extracted features, the actual waveform characteristic values are obtained, that is, the feature data is digitalized.

[0258] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0259] Perform forward modeling on the carbonate rock cavernous reservoir model using the Ricker wavelet to determine the waveform characteristic parameters.

[0260] In this embodiment, performing forward modeling on the carbonate rock cavernous reservoir model using the Ricker wavelet can accurately simulate the propagation and scattering process of seismic signals, and has high flexibility. The Ricker wavelet can be adjusted as needed, including amplitude, frequency, phase, etc., to adapt to different seismic exploration requirements. Moreover, since the Ricker wavelet is a mathematical function, it can be reused in different seismic exploration projects, improving the simulation efficiency, having high repeatability, and having wide applicability and high efficiency.

[0261] Traditional analysis only looks at the similarity of waveforms and does not express waveform features in numerical form. This application differs from traditional methods in this regard.

[0262] The traditional prediction model is established based on waveform similarity and does not express waveform features in numerical form, resulting in a low analysis accuracy of the filling pattern. However, for the prediction computer device of the filling pattern features in this application, its prediction model is established based on the relationship between sensitive parameters and cave filling feature values, and the sensitive parameters are preferably obtained from waveform feature parameters. That is to say, the prediction model of this application is obtained based on the digitalized and optimized waveform feature parameters, which improves the accuracy of the prediction model of this application, thereby improving the analysis accuracy of the filling pattern and further increasing the drilling encounter rate.

[0263] Embodiment Nine

[0264] In this embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0265] Step 110, obtaining cave filling feature values;

[0266] Step 120, using the cave filling feature values to construct a carbonate rock cave reservoir model;

[0267] Step 130, performing forward modeling on the carbonate rock cave reservoir model to determine waveform feature parameters;

[0268] Step 140, through a waveform automatic extraction method, optimizing the waveform feature parameters to obtain sensitive parameters;

[0269] Step 150, based on the sensitive parameters and the cave filling feature values, constructing a prediction model;

[0270] Step 160, using the prediction model to perform waveform analysis to obtain predicted filling pattern features.

[0271] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0272] Based on the peak amplitude value, the trough amplitude value, the amplitude ratio, and the cave filling feature values, constructing a prediction model.

[0273] In this embodiment, the sensitive parameters include the peak amplitude value, the trough amplitude value, and the amplitude ratio.

[0274] In this embodiment, when the cave is not filled, the peak amplitude value and the trough amplitude value are basically equivalent. When the filling increases, the difference between the peak amplitude value and the trough amplitude value becomes larger, and the difference between the maximum peak amplitude value and the maximum trough amplitude value can be used to roughly characterize the degree of cave filling.

[0275] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented:

[0276] Using the inductive learning method, training the sensitive parameter to establish the corresponding relationship between the sensitive parameter and the cave filling characteristic value;

[0277] Using the corresponding relationship to construct the prediction model.

[0278] In this embodiment, the sensitive parameter is the waveform characteristic parameter after optimization, which improves the accuracy of the prediction model constructed by the sensitive parameter.

[0279] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented:

[0280] Obtaining the sand-shale filling speed, the cave reservoir speed, and the surrounding rock speed;

[0281] Using the sand-shale filling speed, the cave reservoir speed, and the surrounding rock speed to construct the carbonate rock cave-type reservoir model.

[0282] In this embodiment, the cave filling characteristic value includes the sand-shale filling speed, the cave reservoir speed, and the surrounding rock speed.

[0283] In this embodiment, the carbonate rock cave-type reservoir model constructed using the sand-shale filling speed, the cave reservoir speed, and the surrounding rock speed has a relatively high correlation between the waveform characteristic parameters determined in the subsequent forward modeling and the cave filling characteristic value.

[0284] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented:

[0285] Obtaining the actual waveform characteristic value;

[0286] Using the prediction model to analyze the actual waveform characteristic value to obtain the predicted filling mode characteristic.

[0287] In this embodiment, the prediction model performs waveform analysis on the actual waveform to obtain the actual waveform characteristic value, and then analyzes the actual waveform characteristic value to obtain the predicted filling mode characteristic.

[0288] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented:

[0289] Obtain the original seismic data;

[0290] Extract the waveform from the original seismic data to obtain the actual waveform;

[0291] Extract the features of the actual waveform to obtain the actual waveform feature values.

[0292] In this embodiment, first, the waveform is extracted from the original seismic data to obtain the actual waveform, which is convenient for observing the actual waveform for feature extraction. Then, according to the extracted features, the actual waveform feature values are obtained, that is, the features are digitized.

[0293] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented:

[0294] Use the Ricker wavelet to perform forward modeling on the carbonate rock cave reservoir model to determine the waveform characteristic parameters.

[0295] In this embodiment, using the Ricker wavelet to perform forward modeling on the carbonate rock cave reservoir model can accurately simulate the propagation and scattering process of seismic signals, and has high flexibility. The Ricker wavelet can be adjusted according to needs, including amplitude, frequency, phase, etc., to adapt to different seismic exploration requirements. Moreover, since the Ricker wavelet is a mathematical function, it can be reused in different seismic exploration projects, improving the simulation efficiency, with high repeatability, and the Ricker wavelet has wide applicability and high efficiency.

[0296] Traditional analysis only looks at the similarity of waveforms and does not express waveform features in a numerical way. This application is different from traditional methods in this regard.

[0297] The traditional prediction model is established based on waveform similarity and does not express waveform features in a numerical way, resulting in a low analysis accuracy of the filling pattern. However, for the prediction computer device of the filling pattern features in this application, its prediction model is established based on the relationship between sensitive parameters and cave filling feature values, and the sensitive parameters are preferably obtained from waveform characteristic parameters. That is to say, the prediction model of this application is obtained based on the digitized and optimized waveform characteristic parameters, which improves the accuracy of the prediction model of this application, thereby improving the analysis accuracy of the filling pattern and further increasing the drilling encounter rate.

[0298] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0299] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0300] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the appended claims.

Claims

1. A method for predicting filling pattern features, characterized in that, comprising: Obtaining cave filling feature values; Using the cave filling feature values to construct a carbonate cave reservoir model; Performing forward modeling on the carbonate cave reservoir model to determine waveform feature parameters; Optimizing the waveform feature parameters through a waveform automatic extraction method to obtain sensitive parameters; Constructing a prediction model based on the sensitive parameters and the cave filling feature values; Performing waveform analysis using the prediction model to obtain the predicted filling pattern features.

2. The method according to claim 1, characterized in that, the sensitive parameters include peak amplitude value, trough amplitude value and amplitude ratio; the step of constructing a prediction model based on the sensitive parameters and the cave filling feature values includes: Constructing a prediction model based on the peak amplitude value, the trough amplitude value, the amplitude ratio and the cave filling feature values.

3. The method according to claim 1, characterized in that, the step of obtaining a prediction model based on the sensitive parameters and the cave filling feature values includes: Using an inductive learning method to train the sensitive parameters to establish a correspondence between the sensitive parameters and the cave filling feature values; Using the correspondence to construct the prediction model.

4. The method according to claim 1, characterized in that, the cave filling feature values include sand-shale filling speed, cave reservoir speed and surrounding rock speed; the step of obtaining cave filling feature values includes: Obtaining the sand-shale filling speed, the cave reservoir speed and the surrounding rock speed; the step of using the cave filling feature values to construct a carbonate cave reservoir model includes: Using the sand-shale filling speed, the cave reservoir speed and the surrounding rock speed to construct the carbonate cave reservoir model.

5. The method according to claim 1, characterized in that, the step of performing waveform analysis using the prediction model to obtain the predicted filling pattern features includes: Obtaining actual waveform feature values; Using the prediction model to analyze the actual waveform feature values to obtain the predicted filling pattern features.

6. The method according to claim 5, characterized in that, the step of obtaining actual waveform feature values includes: Obtaining original seismic data; Performing waveform extraction on the original seismic data to obtain an actual waveform; Performing feature extraction on the actual waveform to obtain actual waveform feature values.

7. The method according to claim 1, characterized in that, the step of performing forward modeling on the carbonate cave reservoir model to determine waveform feature parameters includes: Performing forward modeling on the carbonate cave reservoir model using a Ricker wavelet to determine waveform feature parameters.

8. A prediction device for filling pattern features, characterized in that, comprising: A filling feature value acquisition module for obtaining cave filling feature values; A reservoir model construction module for using the cave filling feature values to construct a carbonate cave reservoir model; A waveform feature parameter determination module for performing forward modeling on the carbonate cave reservoir model to determine waveform feature parameters; A sensitive parameter determination module, configured to optimize the waveform feature parameters through a waveform automatic extraction method to obtain sensitive parameters; A prediction model construction module, configured to construct a prediction model based on the sensitive parameters and the cave filling feature values; A prediction feature determination module, configured to perform waveform analysis using the prediction model to obtain predicted filling pattern features.

9. A computer device, including a memory and a processor, where the memory stores a computer program, characterized in that, when the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, on which a computer program is stored, characterized in that, when the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images