Corrosion valley undercut depth prediction model construction method, prediction method and device

Abstract

The invention relates to the field of oil-gas exploration, in particular to a corrosion valley undercut depth prediction model construction method, a prediction method and a prediction device. Comprising the following steps: determining a first sample mapping group according to the undercutting depth of a sample corrosion valley and seismic attribute values of different types of seismic attributes at different undercutting depths; according to the correlation between all seismic attributes in the first sample mapping group and the undercut depth of the corrosion valley, selecting the seismic attribute with the correlation greater than a preset correlation threshold as a first seismic attribute, and forming a second sample mapping group by the seismic attribute value of the first seismic attribute and different undercut depths; taking the second sample mapping group as a training sample data set; and training the training sample data set by using a random forest regression algorithm, and establishing a corrosion valley undercut depth prediction model. According to the method, the prediction mapping of the seismic attributes and the undercut depth of the corrosion gully is established by utilizing the geological, characteristic drilling and seismic data of the corrosion gully, and the fine prediction of the undercut depth of the corrosion gully in the three-dimensional seismic work area is completed.

Description

Technical Field

[0001] This article relates to the field of oil and gas exploration, especially to the method for constructing a prediction model for the incision depth of a karst gully, and to a prediction method and device. Background Art

[0002] Drilling has confirmed that the karst valley mudstone-filled traps at the top of Natih have great exploration potential. The key factor affecting the effectiveness of the trap is the degree of closure of the cap rock to the trap, and the depth of the karst valley directly determines the degree of closure of the cap rock. However, due to the influence of the complex faults of the Natih Formation, the depth of the karst valley is difficult to accurately predict, which restricts the exploration process of the karst valley mudstone-filled traps.

[0003] At present, the research on the prediction of karst gullies mainly includes: using paleo-geomorphological research methods to identify and predict karst gullies; using seismic attributes that are sensitive to karst gullies (for example, coherence attributes, amplitude, frequency division and other attributes) to predict the distribution characteristics of karst gullies; using a combination of model forward modeling and geomorphological analysis to predict the distribution of karst gullies. These research methods are carried out in a tectonic background with relatively gentle structures and not particularly developed faults, and are based on the prediction of the planar distribution characteristics of karst gullies, without involving the incision depth of karst gullies.

[0004] Currently, there is no public method for predicting the incision depth of karst gullies in complex fault settings. Summary of the invention

[0005] In order to solve the above-mentioned problems in the prior art, the embodiments of this article provide a method for constructing a model for predicting the incision depth of a karst gully, a prediction method and a device.

[0006] According to some embodiments of the present specification, a first mapping group of samples is determined based on the incision depth of the sample dissolution valley and the seismic attribute values ​​of different types of seismic attributes at different incision depths; based on the correlation between all seismic attributes in the first sample mapping group and the incision depth of the dissolution valley, a seismic attribute with a correlation greater than a preset correlation threshold is selected as the first seismic attribute, and the seismic attribute values ​​of the first seismic attribute and different incision depths form a second sample mapping group; the second sample mapping group is used as a training sample data set; the training sample set is trained using a random forest regression algorithm, and a dissolution valley incision depth prediction model that reflects the mapping relationship between the first seismic attribute and the incision depth of the sample dissolution valley is established.

[0007] According to one aspect of the embodiments of the present invention, determining the first mapping group of samples according to the incision depth of the sample dissolution valley and the seismic attributes of different types of seismic attributes at different incision depths includes: acquiring a three-dimensional sample seismic data body, logging sample data and a sample seismic interpretation horizon of the sample dissolution valley to determine the forward modeled sample seismic profile of the sample dissolution valley; reading the seismic attribute values ​​of each seismic attribute at different incision depths of the sample dissolution valley on the forward modeled sample seismic profile in the time domain according to the sample seismic interpretation horizon of the sample dissolution valley; reading the depth values ​​at different incision depths of the sample dissolution valley on the forward model in the depth domain according to the sample seismic interpretation horizon of the sample dissolution valley, the depth values ​​and the seismic attribute values ​​of each seismic attribute forming the first sample mapping group.

[0008] According to one aspect of the embodiments of the present invention, determining the correlation between seismic attributes and the incision depth of the karst valley includes: cross-analyzing the first sample mapping group, determining the fitting coefficients of the different incision depths of the sample karst valleys in each sample mapping group and the seismic attributes, and obtaining the correlation between different types of seismic attributes and the incision depth of the karst valley.

[0009] According to one aspect of the embodiments of the present invention, a forward modeled sample seismic profile of the sample dissolution valley is determined based on a three-dimensional sample seismic data body, well logging sample data and sample seismic interpretation layer of the sample dissolution valley, and the method includes: analyzing the well logging sample data and the three-dimensional sample seismic data to determine the geological characteristics of the sample dissolution valley, wherein the geological characteristics include the width range and depth range of the sample dissolution valley; constructing forward modeling models of sample dissolution valleys with different downcutting depths and different widths according to the width range and depth range of the sample dissolution valley; and determining the forward modeling seismic profile of the sample dissolution valley corresponding to the forward modeling models of the sample dissolution valleys with different downcutting depths and different widths by using a wave equation forward simulation method.

[0010] According to one aspect of the embodiments of this document, the different undercutting depths include: a maximum undercutting depth, a half undercutting depth, and a quarter undercutting depth.

[0011] According to one aspect of the embodiments of the present invention, after constructing a forward model of a dissolution valley with different incision depths and different widths, the method further includes: determining the drilling geological stratification corresponding to the top and bottom surfaces of each layer segment in the sample dissolution valley forward model; determining the velocity and density of each layer segment according to the drilling geological stratification corresponding to the top and bottom surfaces; and determining the velocity and density of the sample dissolution valley forward model according to the density and velocity.

[0012] The embodiments of the present invention also provide a method for predicting the incision depth of a karst gully, which utilizes a karst gully incision depth prediction model, including: obtaining a three-dimensional seismic data body, well logging data, and seismic interpretation layer of the karst gully in the target work area; determining a forward seismic profile of the karst gully in the target work area based on the three-dimensional seismic data body, well logging data, and seismic interpretation layer of the karst gully in the target work area; and inputting the value of the first seismic attribute on the forward seismic profile into the karst gully incision depth prediction model to obtain a predicted incision depth of the karst gully.

[0013] According to one aspect of the embodiments of the present invention, before inputting the value of the first seismic attribute into the karst valley incision depth prediction model, the method includes: extracting the layer-by-layer slices of the first seismic attribute from the seismic interpretation layer of the karst valley in the target work area to obtain the plane prediction range of the karst valley in the target work area; creating a distribution polygon of the karst valley in the plane prediction range of the karst valley in the target work area to determine the plane distribution constraint boundary of the karst valley in the target work area; reading the seismic attribute value of the first seismic attribute at each CDP within the plane distribution constraint boundary of the karst valley in the target work area, and inputting the seismic attribute value of the first seismic attribute into the karst valley incision depth prediction model.

[0014] The embodiments of the present invention also provide a device for constructing a karst valley incision depth prediction model, the device comprising: a sample first mapping group determination unit, used to determine the sample first mapping group according to the incision depth of the sample karst valley and the seismic attributes of different types of seismic attributes at different incision depths; a second sample mapping group determination unit, used to select a seismic attribute with a correlation greater than a preset correlation threshold as a first seismic attribute according to the correlation between all seismic attributes in the first sample mapping group and the incision depth of the karst valley, and the seismic attribute value of the first seismic attribute and different incision depths form a second sample mapping group; a training sample data set determination unit, used to use the second sample mapping group as a training sample data set; a karst valley incision depth prediction model determination unit, used to train the sample data set using a random forest regression algorithm, establish a mapping relationship between the first seismic attribute and the sample karst valley incision depth, and obtain a karst valley incision depth prediction model.

[0015] The embodiment of the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the target well area multi-layer fracture interpretation method is implemented.

[0016] The embodiments of the present invention further provide a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the multi-layer fracture interpretation method of the target well area is implemented.

[0017] The embodiments of this specification start from the actual geological conditions, take the analysis of the geological characteristics of the karst valley as a prerequisite, make full use of drilling and seismic data, establish a forward model of the karst valley, and based on the wave equation forward simulation technology and the random forest multivariate regression method, establish a predictive mapping between seismic attributes and the incision depth of the karst valley, and use this predictive mapping based on the preferred three-dimensional seismic attributes to complete the detailed prediction of the incision depth of the karst valley in the three-dimensional seismic work area. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of this article or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of this article. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 It is a flow chart of a method for constructing a model for predicting the depth of erosion valley incision according to an embodiment of this invention;

[0020] Figure 2 The figure is a flow chart of a method for determining a first mapping group of samples according to an embodiment of this invention;

[0021] Figure 3 The flowchart of a method for determining the correlation between seismic attributes and the depth of incision of karst valleys in an embodiment of this invention is shown;

[0022] Figure 4 It is a flow chart of a method for determining a forward modeling sample seismic profile of a sample dissolution valley according to an embodiment of this invention;

[0023] Figure 5 The figure is a flow chart of a method for determining a sample forward model according to an embodiment of this invention;

[0024] Figure 6 It is a flow chart of a method for predicting the depth of the erosion valley incision according to an embodiment of this invention;

[0025] Figure 7 Shown is a flow chart of a method for predicting undercut depth according to an embodiment of this invention;

[0026] Figure 8 It is a schematic diagram of the structure of a device for constructing a karst valley incision depth prediction model according to an embodiment of this invention;

[0027] Fig. 9 It is a schematic diagram of a seismic interpretation section of the development characteristics of sample dissolution gullies under a complex fault background in an embodiment of this specification;

[0028] Fig. 10A , Fig. 10B , Fig. 10C , Fig. 10D , Fig.10E , Fig.10F , Figure 10G , Fig. 10H and Fig.10I It is a schematic diagram of a group of forward modeling of sample dissolution gullies under a complex fault background in an embodiment of this article;

[0029] Fig.11A , Fig. 11B , Fig. 11C , Fig.11D , Fig.11E , Fig.11F , Fig.11G and Fig.11H It is a schematic diagram of a forward simulation section of a dissolution valley model of eight samples with different depths and widths according to an embodiment of this invention;

[0030] Fig.12 The figure shows the plane distribution characteristics prediction map of the karst gullies in the target work area under a complex fault background in the embodiment of this article;

[0031] Fig.13 The figure shows a prediction diagram of the incision depth of a karst gully in the embodiment of this article;

[0032] Fig.14 Shown is a schematic diagram of the structure of a computer device according to an embodiment of this invention.

[0033] Description of the accompanying symbols:

[0034] 801. A sample first mapping group determining unit;

[0035] 802. A second sample mapping group determining unit;

[0036] 803, training sample data set determination unit;

[0037] 804. Determination unit of the prediction model for the depth of incision of karst gullies;

[0038] 1402. Computer equipment;

[0039] 1404, processor;

[0040] 1406. Memory;

[0041] 1408, driving mechanism;

[0042] 1410, input / output module;

[0043] 1412. Input device;

[0044] 1414. Output device;

[0045] 1416. Presentation equipment;

[0046] 1418. Graphical user interface;

[0047] 1420, network interface;

[0048] 1422, communication link;

[0049] 1424. Communication bus. DETAILED DESCRIPTION

[0050] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this specification.

[0051] It should be noted that the terms "first", "second", etc. in the specification and claims of this article and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of this article described here can be implemented in an order other than those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, device, product or equipment that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0052] This specification provides method operation steps as described in the embodiments or flow charts, but more or fewer operation steps may be included based on routine or non-creative work. The order of steps listed in the embodiments is only one way of executing the order of many steps and does not represent the only execution order. When the system or device product is executed in practice, it can be executed in the order of the method shown in the embodiments or the drawings or in parallel.

[0053] It should be noted that the method for constructing a karst gully incision depth prediction model and the prediction method in this article can be used in the field of oil and gas exploration. This article does not limit the application field of the method for constructing a karst gully incision depth prediction model, the prediction method and the device.

[0054] Figure 1The flowchart of a method for constructing a prediction model of the incision depth of a karst gully according to an embodiment of the present invention is shown. In this method, a prediction model of a karst gully with a known incision depth is established. Specifically, the following steps are included:

[0055] Step 101 : determining a first mapping group of samples according to the incision depth of the sample dissolution valley and the seismic attribute values ​​of different types of seismic attributes at different incision depths.

[0056] In this step, a sample dissolution valley with a known downcutting depth in the sample work area is selected, and the downcutting depth of the sample dissolution valley is used to read the seismic attribute values ​​of different types of seismic attributes at different downcutting depths of the sample dissolution valley, and the seismic attribute values ​​and the corresponding downcutting depths are used as the first mapping group of the sample.

[0057] Specifically, the first mapping group of the sample is determined by constructing a sample dissolution valley forward model and a forward sample seismic profile. The sample dissolution valley forward model can be a dissolution valley model with different depths and different widths. For example, the sample dissolution valley forward model has 8 different depths and 4 different widths, thus forming a total of 8*4 types of 32 dissolution valleys, and multiple valley incision depth seismic attribute values ​​are read in each valley.

[0058] The incision depth of the sample dissolution valley is obtained by analyzing the valley model of the sample dissolution valley. For details on constructing the valley model of the sample dissolution valley, see Figure 2 Description. Different types of seismic attributes are extracted from the forward modeling section of the sample dissolution valley by using seismic attribute extraction technology. That is, different types of seismic attributes are extracted from each section of the forward modeling section of the sample dissolution valley.

[0059] In this step, different types of seismic attributes include, but are not limited to: maximum amplitude, peak amplitude, absolute amplitude, instantaneous amplitude, maximum energy, average energy, energy gradient, arc length, coherence, maximum likelihood coherence, eigenvalue coherence, waveform difference, maximum curvature, minimum curvature, most positive curvature, average curvature, attenuation gradient, first-order derivative of instantaneous amplitude, absorption coefficient, coherent energy gradient attribute, etc. The embodiment of this specification selects these 20 seismic attributes to construct the first mapping group of samples, that is, the forward modeling profile of each karst gully model has these 20 attribute profiles.

[0060] Specifically, different cutting depths include but are not limited to: the maximum cutting depth, half the cutting depth, one quarter of the cutting depth, three eighths of the cutting depth, etc.

[0061] Step 102, based on the correlation between all seismic attributes in the first sample mapping group and the incision depth of the karst valley, select the seismic attribute with a correlation greater than a preset correlation threshold as the first seismic attribute, and the seismic attribute value of the first seismic attribute and different incision depths form a second sample mapping group.

[0062] In this step, it is necessary to select seismic attributes with a greater correlation with the downcutting depth from a large number of seismic attributes to construct sample data for model training. Specifically, along the seismic interpretation horizon of the sample dissolution valley, read the seismic attribute values ​​at different downcutting depths in the forward modeling profile of the sample dissolution valley in the time domain and the downcutting depth values ​​at different downcutting depths of the forward model of the sample dissolution valley in the depth domain, respectively, and create an intersection diagram of seismic attributes and downcutting depths. The correlation between seismic attributes and the downcutting depth of the dissolution valley is analyzed through the intersection diagram. For a detailed description of analyzing the correlation between seismic attributes and downcutting depth, see Figure 2 describe.

[0063] The seismic attribute with a correlation greater than a preset correlation threshold is selected as the first seismic attribute. The first seismic attribute can be understood as a preferred seismic attribute. In one embodiment of the specification of this article, five seismic attributes, namely, maximum curvature, coherent energy gradient, attenuation gradient, absorption coefficient and average curvature attributes, are selected as the first seismic attribute for constructing training samples for the prediction model of the incision depth of the karst valley. Furthermore, the attribute values ​​of the first seismic attribute at different incision depths and the depth values ​​at different incision depths are used to form data pairs as the second sample mapping group.

[0064] Step 103, using the second sample mapping group as a training sample data set. In the above steps, if Q first seismic attributes with large correlation are selected, and the corresponding dissolution valley incision depths are read, the depth values ​​of the Q first seismic attributes and the dissolution valley incision depths are used as a set of data pairs, that is, a sample data. Then each sample data has Q+1 features, where the incision depth is the label data in the sample data. For example, 5 first seismic attributes and corresponding incision depths of the dissolution valley are selected in step 102, then each sample data in the second sample mapping group includes 6 features. The second mapping sample group is used as a training sample data set for training the model. In the model training process of the embodiment of this specification, 8*4 types of 32 sample dissolution valley forward models are involved, and 5 incision depths of each model are read, that is, 32*5 total 160 sample data are obtained.

[0065] Step 104: Use a random forest regression algorithm to train the training sample data set to establish a karst gully incision depth prediction model that reflects the mapping relationship between the first seismic attribute and the sample karst gully incision depth.

[0066] The random forest multivariate regression method is used to determine the three main parameters of the random forest regression method, including: the number of training sample subsets (i.e. the number of decision trees) k, the extracted feature attributes m, and the maximum depth of tree growth r. These three parameters need to be optimized through repeated experiments based on the average error minimization criterion.

[0067] On this basis, regression analysis was performed on 160 sample data to establish a mapping relationship between the incision depth of the karst gully and five preferred seismic attributes, namely, the prediction mapping of the incision depth of the karst gully.

[0068] Specifically, the Bagging method is used to randomly extract several samples from the initial training sample data set to construct k training sample subsets, where the training sample subsets can be composed of θ1, θ2, …, θ k In some embodiments of the present specification, 100 samples are randomly selected from 160 sample data to construct 100 training sample subsets, which are the training sample data sets. Therefore, the number k of training sample subsets is determined to be 100.

[0069] Furthermore, using the random subspace method, we randomly select m feature attributes from all feature attributes, split the nodes, and select the optimal splitting method among these m features to build a single regression decision tree. In this step, we set m=3, repeat the above steps, and build multiple regression decision trees, namely {H(x,θ1),H(x,θ2),…,H(x,θ k )} and maximize the growth of each tree to form a forest.

[0070] Through repeated experiments, the maximum depth r of the tree is selected as 6. Furthermore, the prediction results of all regression decision trees are averaged to obtain the final prediction result. That is:

[0071]

[0072] Where: is the final prediction result of random forest; h(x,θ i ) is the decision number H(x,θ i ), x is the independent variable, k is the number of decision trees, θ i is the i-th training sample subset, and k is the total number of training sample subsets.

[0073] Figure 2 The flowchart of a method for determining a first mapping group of samples in an embodiment of this invention is shown, which specifically includes the following steps:

[0074] Step 201, obtain the three-dimensional sample seismic data volume, well logging sample data and sample seismic interpretation horizon of the sample dissolution valley, and determine the forward modeling sample seismic profile of the sample dissolution valley. Specifically, firstly, a forward modeling model of the sample dissolution valley is established based on the three-dimensional sample seismic data volume, well logging sample data and sample seismic interpretation horizon. Further, through forward modeling simulation, a forward modeling sample seismic profile is obtained.

[0075] In this step, the offset stacked three-dimensional seismic data volume after amplitude preservation processing is obtained; all wells that penetrate the target layer (wherein the target layer is the stratum where the sample dissolution valley is developed) are selected in the sample work area, and among these wells, several wells that have density, acoustic wave, well diameter curve and geological stratification data are preferably obtained, and the density, acoustic wave, well diameter curve and well logging geological stratification data of several wells are obtained, so as to obtain well logging sample data. Further, according to the seismic geological calibration results of the synthetic seismic records, a one-to-one correspondence is established between the seismic reflection interface in the time domain and the geological stratification in the depth domain, and multiple seismic interpretation horizons corresponding to the main geological stratifications in the sample work area are obtained, so as to obtain sample seismic interpretation horizons. Fig. 9 The figure shows a schematic diagram of a seismic interpretation profile of the development characteristics of sample dissolution valleys under a complex fault background in an embodiment of this specification. The figure shows the main reflection layers above and below the target layer (development layer) of the sample work area. In the figure, the number of seismic interpretation horizons is 10, that is, 10 seismic reflection horizons can be obtained in the sample work area.

[0076] In this step, a forward seismic model of the sample dissolution valley is preliminarily constructed based on the acquired three-dimensional sample seismic data volume, logging sample data and sample seismic interpretation layer of the sample dissolution valley. Then, based on the forward seismic model of the sample dissolution valley, a corresponding forward seismic profile of the sample dissolution valley is constructed. In the process of constructing the forward seismic model of the sample dissolution valley, first, based on the analysis results of the geological characteristics of the sample dissolution valley, the lateral width and vertical depth of the forward model of the sample dissolution valley are determined. Further, based on the development layer of the sample dissolution valley and combined with the interpretation results of the seismic data, it is determined that the forward model of the sample dissolution valley is designed with several layers in the upper and lower parts of the dissolution valley development layer. Furthermore, based on the analysis of the geological characteristics of the sample dissolution valley, the width and approximate depth range of the dissolution valley are obtained, and several forward models with different depths and widths are designed to participate in the model forward analysis. For a detailed description of the construction of the forward seismic model and the forward seismic profile, please refer to Figure 4 .

[0077] Step 202, according to the sample seismic interpretation horizon of the sample dissolution valley, read the seismic attribute values ​​of each seismic attribute at different downcutting depth steps of the sample dissolution valley on the forward modeling sample seismic profile in the time domain. In this step, along the sample seismic interpretation horizon of the sample dissolution valley, read the seismic attribute values ​​at different downcutting depths in the forward modeling sample seismic profile of the sample dissolution valley in the time domain. Among them, the values ​​of all seismic attributes on each seismic profile in the sample forward modeling sample seismic profile can be read.

[0078] Step 203, according to the sample seismic interpretation horizon of the sample dissolution valley, read the depth values ​​at different downcutting depths of the sample dissolution valley on the forward model in the depth domain, and the depth values ​​and the seismic attribute values ​​of each seismic attribute form a first sample mapping group. In this step, the seismic interpretation along the sample dissolution valley is to read the depth values ​​at different downcutting depths corresponding to the forward sample seismic profile in step 202 in the sample forward model of the sample dissolution valley in the depth domain.

[0079] Figure 3 The flowchart of a method for determining the correlation between seismic attributes and the depth of the karst valley incision according to an embodiment of the present invention is shown, which specifically includes the following steps:

[0080] Step 301, along the interpretation horizon of the karst gully, read the seismic attribute values ​​at different incision depths in the time domain forward modeling sample seismic profile. And read the incision depth value in the corresponding depth domain sample forward model, create an intersection diagram between the seismic attribute and the incision depth of the karst gully, and analyze the correlation between the seismic attribute and the incision depth of the karst gully through the intersection diagram.

[0081] from Fig. 9 The forward model of the sample dissolution valley sample shown in the figure selects one of the forward modeling sample seismic profiles R1 (such as FIG. 10A to FIG. 10H Taking a model in the example, a dissolution valley of a certain width is selected, and along the interpretation horizon of the sample dissolution valley, the seismic attribute values ​​of a corresponding seismic attribute Attribute are read at the maximum, 1 / 2 and 1 / 4 of the dissolution valley downcut depth on the forward modeling sample seismic section in the time domain, respectively: Attribute- max , Attribute- 1 / 2 , Attribute- 1 / 4 At the same time, read the maximum incision depth value Depth- max , Depth- 1 / 2 And the cutting depth value Depth- 1 / 4 , thus obtaining a set of data on the incision depth of the karst valley and the seismic attribute values ​​of the R1 profile.

[0082] In the same way, in step 302, the incision depth and attribute value data of other width karst gullies in the forward modeling sample seismic profile are read, that is, the G group of data of the incision depth and attribute value of the karst gullies in the R1 profile is obtained. Further, in the same way as the R1 profile, the different incision depths of the karst gullies on other forward modeling sample seismic profiles and the corresponding seismic attribute values ​​are obtained.

[0083] Step 303, read the dissolution valley incision depth and the corresponding seismic attribute value on other forward modeling sample seismic profiles. Specifically, read the different valley incision depths and corresponding attribute value data on other R-1 forward modeling profiles in the R profile, thereby obtaining the corresponding data of the R*G group incision depth and Attribute attribute. With depth as the horizontal coordinate and attribute value as the vertical coordinate, establish an intersection diagram for the R*G group valley incision depth and the corresponding Attribute attribute value data, analyze the correlation between the two, and obtain the fitting coefficient C-R1 of the two, thereby obtaining the correlation between the valley incision depth and a certain seismic attribute Attribute.

[0084] According to the same method as the correlation calculation of the Attribute attribute, the fitting coefficients between the incision depth of the karst gully and the other Q-1 attributes are obtained in turn, thereby realizing the correlation analysis between the Q seismic attributes and the incision depth of the karst gully. In this embodiment, the correlation between the incision depth of the karst gully and the corresponding different seismic attribute values ​​is analyzed in turn with 20 kinds of seismic attributes.

[0085] Taking the maximum curvature attribute as an example, the correlation analysis process between seismic attributes and the depth of dissolution valley incision is specifically explained. First, for the forward sections of 8 sample dissolution valleys, the maximum curvature attribute and the value of the dissolution valley incision depth are read. Taking the forward section of the sample dissolution valley with a depth of 15 meters (Depth = 15) as an example, for the dissolution valley with a width of 158 meters (Width = 158), along the interpreted horizon of the dissolution valley, the corresponding maximum curvature attribute (Attribute = maximum curvature) values ​​are read at the maximum, 1 / 2 and 1 / 4 of the dissolution valley incision depth on the forward section in the time domain, which are 0.27, 0.13 and 0.06 respectively (that is, Attribute- max =0.27, Attribute- 1 / 2 =0.13, Attribute- 1 / 4 =0.06), and the corresponding valley incision depth values ​​were read on the forward model of the depth domain sample at a depth of 15 meters, which were 15.6 meters, 7.8 meters and 3.9 meters respectively (i.e. Depth- ma x=15.6, Depth- 1 / 2 =7.8, Depth-1 / 4 =3.9), and a set of data on the incision depth and maximum curvature attribute values ​​of the karst gully was obtained. According to the same method, on the forward section of the karst gully with a depth of 15 meters, the data of the karst gullies with other widths (widths of 120 meters (Width=120), 210 meters (Width=210) and 309 meters (Width=309)) were read, that is, the data on the incision depth and maximum curvature attribute of the karst gully were read respectively, thereby obtaining 4 sets (G=4) of data on the incision depth and maximum curvature attribute of the karst gully forward section with a depth of 15 meters. According to the same method as the forward modeling profile with a depth of 15 meters, the different valley cutting depths and corresponding attribute value data on the other 7 forward modeling profiles (i.e., the forward modeling profiles with depths of 5 meters, 10 meters, 20 meters, 25 meters, 30 meters, 35 meters and 50 meters respectively) are read, thereby obtaining 32 groups (R*G=32) of corresponding data of cutting depths and maximum curvature attributes.

[0086] The depth of the karst valley is used as the horizontal coordinate, and the maximum curvature attribute value is used as the vertical coordinate to form multiple sets of data. These data are cross-analyzed to obtain the fitting coefficient of the karst valley depth and seismic attributes. The fitting coefficient is the correlation. The larger the fitting coefficient, the higher the correlation.

[0087] Figure 4 The flowchart of the method for determining the forward modeling sample seismic profile of the sample dissolution valley in the embodiment of this article is shown, which specifically includes the following steps:

[0088] Step 401, analyze the well logging sample data and the three-dimensional sample seismic data to determine the geological characteristics of the sample dissolution valley, wherein the geological characteristics include the width range and depth range of the sample dissolution valley. In this step, the geological characteristics of the dissolution valley can be determined by comprehensive analysis of the three-dimensional sample seismic data body and the well logging sample data of the sample dissolution valley. Through analysis, the sedimentary background, development layer and lithological characteristics of the sample dissolution valley can be determined; secondly, the seismic data is fully utilized, and the seismic reflection characteristics of the sample dissolution valley are identified through comprehensive analysis of seismic profiles and seismic attributes.

[0089] Specifically, the geological characteristics of the sample dissolution valley can be determined through a comprehensive analysis of the well section, seismic section and seismic attributes of the sample work area. First, the well section is compared and analyzed to determine the sedimentary background, development layer and lithological characteristics of the sample dissolution valley. Through well analysis, it is determined that the sample dissolution valley is developed on the carbonate rock unconformity surface. Due to long-term weathering and leaching, a series of dissolution valleys are developed on the top of the unconformity surface. The development layer is the Natih Formation of the Cretaceous System. The main lithology is limestone, and the lithology filling the dissolution valley is the mudstone of the Fiqa Formation. Secondly, make full use of seismic data, through a comprehensive analysis of the seismic section and seismic attributes of the sample dissolution valley, according to the morphology of the sample dissolution valley on the seismic section, identify the seismic reflection characteristics of the sample dissolution valley, and determine the width range and downcutting depth range of the sample dissolution valley.

[0090] Step 402: According to the width range and depth range of the sample dissolution valley, sample forward modeling of sample dissolution valleys with different undercut depths and different widths is constructed. According to the geological characteristics of the sample dissolution valley determined in step 401, a sample forward modeling of the sample dissolution valley is established through comprehensive analysis of seismic data and seismic attribute data. Fig. 9 As shown, the sample dissolution valley has obvious downcutting characteristics on the seismic profile, and the seismic reflection phase axis at its lower part has a certain downward pull phenomenon. The width of the sample dissolution valley ranges from 100m to 350m, and the depth ranges from 5m to 60m.

[0091] Step 403, using the wave equation forward modeling method of Tesseral software, determine the sample dissolution valley forward seismic profile corresponding to the sample dissolution valley sample forward model with different incision depths and different widths.

[0092] In this step, the forward seismic profile of the sample dissolution valley is obtained by wave equation forward simulation. First, the observation system of the sample forward model is determined, and the three parameters of the trace spacing, coverage number and shot point mode are set. The shot point mode with the shot point at the center of the detection point is adopted, and the trace spacing and coverage number parameters are set to be consistent with the parameters of the three-dimensional seismic data of the sample area, so as to be closer to the simulation of the actual seismic data. For example, the trace spacing is set to 25 meters and the coverage number is set to 120 times.

[0093] Secondly, according to the actual geological conditions, the type of seismic wavelet and wave equation is selected for forward simulation. According to the principle that the spectrum of the seismic wavelet is similar to the spectrum characteristics of the 3D seismic data, the seismic wavelet is determined by the spectrum characteristics of the sample 3D seismic data, for example, the seismic wavelet is a 30HZ Ricker wavelet.

[0094] According to the geological characteristics of the sample dissolution valley development layer, the acoustic medium wave equation was used for forward simulation, and the pre-stack single-shot gathers of the sample dissolution valley sample forward model were obtained. Static correction, dynamic correction, pre-stack time migration and stacking processing were performed on the gathers obtained from the sample dissolution valley to obtain the forward simulation seismic profiles corresponding to multiple dissolution valley forward models.

[0095] Figure 5 The flowchart of a method for determining a sample forward model in an embodiment of this invention is shown. In the embodiment of this specification, by determining the two physical parameters of velocity and density of each layer in the sample forward model, the method specifically includes the following steps:

[0096] Step 501, determine the drilling geological layers corresponding to the top and bottom surfaces of each layer segment in the sample dissolution valley sample forward model. In this step, the drilling geological layers corresponding to the top and bottom surfaces of a layer segment (eg, Layer D) in the sample forward model are determined.

[0097] Step 502, according to the drilling geological stratification corresponding to the top surface and the bottom surface, determine the velocity and density parameters of each layer. The top surface and the bottom surface are respectively recorded as DU and DN. The two layers of the top surface DU and the bottom surface DN are respectively used as longitudinal constraints to statistically analyze the velocity and density distribution characteristics of seismic waves in the wells within the sample study area. Specifically, by analyzing the seismic wave velocity and density distribution histograms of the DN layer and the DU layer, the velocity value and density value corresponding to the peak of the DU-DN layer velocity and density distribution histogram are read, and the velocity value and density value are the velocity and density values ​​of the layer.

[0098] Step 503, according to the density and velocity, determine the velocity and density of the sample dissolution valley sample forward model. The velocity and density values ​​of each other layer segment in the sample dissolution valley sample forward model are determined by the method in step 501 and step 502. In some embodiments of this specification, the density and velocity parameters of the 11 layers in the sample forward model are respectively counted. As shown in Table 1, Table 1 shows the velocity and density parameters of each layer segment in the sample dissolution valley sample forward model. Taking the Layer3 layer segment as an example, the specific acquisition steps of its parameters are described. First, the geological stratification corresponding to the top and bottom surfaces of the Layer3 layer segment is determined. These two geological stratifications are used as top and bottom constraints, and the velocity distribution characteristics of the five wells in the sample area between the two layers are counted to obtain the velocity distribution histogram, and the velocity value of 3239 meters / second at the peak of the histogram is read, which is the velocity value of the Layer3 layer segment. According to the same method for obtaining the velocity parameters, the density value of the Layer3 layer segment is obtained as 2269 kilograms / cubic meters, thereby completing the acquisition of the velocity and density parameters of the Layer3 segment. According to the same method as that of obtaining the parameters of Layer 3, the velocity and density parameters of each layer in the forward model of the sample dissolution valley are obtained. The parameter values ​​of each layer are statistically shown in Table 1.

[0099]

[0100] Table 1 Velocity and density parameters of each layer in the forward model of the sample dissolution valley

[0101] Figure 6 The flowchart of a method for predicting the incision depth of a karst gully in an embodiment of this invention is shown. The method applies the incision depth prediction model of the karst gully described above to predict the incision depth of an unknown karst gully in a target work area. The method specifically includes the following steps:

[0102] Step 601, obtain the three-dimensional seismic data body, well logging data and seismic interpretation horizon of the karst gully in the target work area. The process of obtaining the three-dimensional seismic data body, well logging data and seismic interpretation horizon in this step is the same as the method of obtaining the three-dimensional sample seismic data body, well logging sample data and sample seismic interpretation horizon of the sample karst gully. In this step, the three-dimensional seismic data body is the offset stacking seismic data after amplitude preservation processing. The well logging data are the density, acoustic wave and well diameter curves of the wells that drilled through the target layer in the target work area, as well as the well logging geological stratification data; the seismic interpretation horizon is the T seismic reflection phase axes corresponding to the main geological stratification in the target work area after calibration through synthetic records. The T seismic reflection phase axes are tracked and interpreted in the three-dimensional area to obtain T key seismic interpretation horizons.

[0103] Furthermore, based on the above seismic data, comparative analysis is performed to determine the sedimentary background, development layer sections and lithological characteristics of the karst valleys in the target work area, the reflection characteristics of the karst valleys on the seismic profile are analyzed through seismic data, and the width and depth ranges of the karst valleys in the target work area are determined in combination with seismic attributes, thereby determining the geological characteristics of the karst valleys in the target work area.

[0104] Step 602, extracting the first seismic attribute from the three-dimensional seismic data volume of the karst gully in the target work area. In some embodiments of the present specification, when predicting the incision depth of the karst gully in the target work area, the stratigraphic structure and underground morphology corresponding to the same work area have certain similarities. Therefore, the karst gully incision depth prediction model constructed by the sample karst gully in the same target work area can be used to predict the karst gully incision depth of other unknown karst gullies in the work area.

[0105] Therefore, based on the first seismic attribute determined during the training process of the karst gully incision depth prediction model, the karst gully in the target work area selects the same first seismic attribute, and the first seismic attribute and the value corresponding to the first seismic attribute are extracted along the target layer from the seismic profile of the karst gully in the target work area.

[0106] Step 603: Input the value of the first seismic attribute at the target layer of the three-dimensional seismic data volume into the karst valley incision depth prediction model to obtain the predicted incision depth of the karst valley. Input the value of the first seismic attribute into the karst valley incision depth prediction model described above to predict the incision depth of the karst valley in the target work area.

[0107] The embodiments of this specification can effectively reduce the interference of complex faults on the prediction of the incision depth of the karst valley, realize the precise prediction of the incision depth of the karst valley, provide reliable basic parameters for the evaluation and optimization of closures related to the karst valley, and provide strong technical support for the exploration of fields related to the karst valley.

[0108] Figure 7 The flowchart of a method for predicting the cutting depth in the embodiment of this invention is shown, which specifically includes the following steps:

[0109] Step 701, along the seismic interpretation horizon of the karst gully in the target work area, extract the slices along the first seismic attribute to obtain the plane prediction range of the karst gully in the target work area. In this step, the first seismic attribute is used to predict the plane distribution range of the karst gully in the target work area. Further, the seismic attribute with the highest correlation with the downcut depth is selected from the first seismic attribute, and along the seismic interpretation horizon of the karst gully in the target work area, extract the slices along the layer of the seismic attribute to obtain the plane distribution prediction range of the karst gully in the target work area.

[0110] Step 702: Create a distribution polygon of the karst gullies in the plane prediction range of the karst gullies in the target work area, and determine the plane distribution constraint boundary of the karst gullies in the target work area, such as Fig.12 In the field of oil and gas exploration, the embodiments of this specification can not only identify the planar distribution of the karst gullies, but also accurately predict the depth of the karst gullies, which complies with the current demand for accurate characterization of key geological parameters.

[0111] Step 703, read the seismic attribute value of the first seismic attribute of the target layer at each CDP within the plane distribution constraint boundary of the karst gully in the target work area, and input the seismic attribute value of the first seismic attribute into the karst gully incision depth prediction model. In some embodiments of the present specification, each CDP is regarded as a sample to be predicted, and the plane distribution boundary of the karst gully is used as a constraint, and the value of the first seismic attribute at each CDP is read in sequence along the seismic interpretation horizon of the karst gully. For example, the maximum curvature, coherent energy gradient, attenuation gradient, absorption coefficient and average curvature attribute body are used as five first seismic attributes, and the values ​​of the first seismic attributes constitute five known features at each CDP, and the incision depth of the karst gully at each CDP is the feature to be predicted. The karst gully incision depth prediction model that reflects the mapping relationship between the first seismic attribute and the sample karst gully incision depth is established by using the random forest regression method. The unknown karst gully incision depth characteristics are predicted using the five known seismic attribute characteristics at each CDP, and the predicted value of the karst gully incision depth at each CDP is obtained. This enables the prediction of the karst gully incision depth within the karst gully distribution range of the three-dimensional work area, such as Fig.13 shown.

[0112] like Figure 8 FIG. 1 is a schematic diagram of a device for constructing a karst gully incision depth prediction model according to an embodiment of the present invention. The basic structure of the device for constructing a karst gully incision depth prediction model is described in the figure. The functional units and modules therein can realize the construction of the karst gully incision depth prediction model by software, or by using a general chip or a specific chip. The device specifically includes:

[0113] The sample first mapping group determination unit 801 is used to determine the sample first mapping group according to the incision depth of the sample karst valley and the seismic attributes of different types of seismic attributes at different incision depths;

[0114] A second sample mapping group determining unit 802 is used to select a seismic attribute with a correlation greater than a preset correlation threshold as a first seismic attribute according to the correlation between all seismic attributes in the first sample mapping group and the incision depth of the karst gully, and the seismic attribute value of the first seismic attribute and different incision depths form a second sample mapping group;

[0115] A training sample data set determining unit 803 is configured to use the second sample mapping group as a training sample data set;

[0116] The karst gully incision depth prediction model determination unit 804 is used to train the sample data set using a random forest regression algorithm, establish a mapping relationship between the first seismic attribute and the sample karst gully incision depth, and obtain a karst gully incision depth prediction model.

[0117] Fig. 9 The figure shows a schematic diagram of a seismic interpretation profile of the development characteristics of sample dissolution valleys under a complex fault background in an embodiment of this specification. The figure shows the main reflection layers above and below the target layer (development layer) of the sample work area. In the figure, the number of seismic interpretation horizons is 10, that is, 10 seismic reflection horizons can be obtained in the sample work area.

[0118] like Fig. 10A , Fig. 10B , Fig. 10C , Fig. 10D , Fig.10E , Fig.10F , Figure 10G , Fig. 10H and Fig.10I The figure shows a schematic diagram of a group of forward modeling of sample dissolution valleys under a complex fault background in the embodiment of this article. In each model in the figure, there are four different widths of dissolution valleys, 120 meters wide, 158 meters wide, 210 meters wide and 309 meters wide, and 8 different downcutting depths are designed for these four different widths of dissolution valley models, which are 5, 10 meters, 15 meters, 20 meters, 25 meters, 30 meters, 35 meters and 50 meters respectively. This group of models can fully simulate the seismic reflection characteristics of dissolution valleys of different widths and different downcutting depths, thereby completing the establishment of the forward modeling of the sample dissolution valley.

[0119] Fig.11A , Fig. 11B , Fig. 11C , Fig.11D , Fig.11E , Fig.11F , Fig.11G and Fig.11H The figure shows a schematic diagram of a forward simulation profile of 8 sample dissolution valley models with different depths and widths in an embodiment of this article. In the figure, each seismic profile has 435 CDP channels, and the reflection characteristics of the upper and lower layers of the sample dissolution valley layer segment are clearly visible on the seismic profile. As the depth of the dissolution valley increases from 5 meters to 50 meters, its downcutting characteristics on the seismic profile are gradually obvious, thereby completing the acquisition of the forward simulation profile of the sample dissolution valley model.

[0120] Fig.12The diagram shows a predicted plane distribution feature of the karst gully in the target work area under a complex fault background in an embodiment of this invention. The seismic attribute with the highest correlation with the downcut depth is selected from the first seismic attribute of the karst gully in the target work area, and the slices along the layer of the seismic attribute are extracted along the seismic interpretation layer of the karst gully in the target work area to obtain the predicted plane distribution range of the karst gully in the target work area. The distribution polygon of the karst gully is created in the predicted plane range of the karst gully in the target work area, and the plane distribution constraint boundary of the karst gully in the target work area is determined.

[0121] Fig.13 The figure shows a prediction diagram of the incision depth of a karst gully in an embodiment of this invention. In some embodiments of this specification, the drilling value and the predicted value of the incision depth of the karst gully within the distribution range of the karst gully in the study area are compared. Taking the BR1 well in the figure as an example, the incision depth of the karst gully revealed by drilling is 24 meters, while the incision depth of the karst gully predicted in this patent is 22.5 meters. The comparative analysis shows that the depth value obtained by the incision depth prediction method of the karst gully in the embodiment of this specification is very close to the incision depth value of the karst gully revealed by drilling, and the agreement is high, thereby verifying the reliability of the incision depth prediction method of the karst gully in the embodiment of this specification.

[0122] The embodiments of this specification predict the depth of dissolution valley incision under the background of complex faults, effectively reduce the interference of complex faults on the prediction of dissolution valley incision depth, realize the fine prediction of dissolution valley incision depth, improve the prediction accuracy of dissolution valley incision depth, provide reliable basic parameters for the evaluation and optimization of closures related to dissolution valleys, and provide strong technical support for the exploration of dissolution valleys. It has strong feasibility, high credibility, and a wide range of applications, and has positive significance for the expansion and promotion of exploration in the field of dissolution valleys.

[0123] like Fig.14As shown, a computer device provided in an embodiment of the present invention, the computer device 1402 may include one or more processors 1404, such as one or more central processing units (CPUs), each of which may implement one or more hardware threads. The computer device 1402 may also include any memory 1406, which is used to store any kind of information such as code, settings, data, etc. Non-restrictive, for example, the memory 1406 may include any one or more combinations of the following: any type of RAM, any type of ROM, flash memory device, hard disk, optical disk, etc. More generally, any memory may use any technology to store information. Further, any memory may provide volatile or non-volatile retention of information. Further, any memory may represent a fixed or removable component of the computer device 1402. In one case, when the processor 1404 executes an associated instruction stored in any memory or a combination of memories, the computer device 1402 may perform any operation of the associated instruction. The computer device 1402 also includes one or more drive mechanisms 1408 for interacting with any memory, such as a hard disk drive mechanism, an optical disk drive mechanism, etc.

[0124] The computer device 1402 may also include an input / output module 1410 (I / O) for receiving various inputs (via input devices 1412) and for providing various outputs (via output devices 1414). A specific output mechanism may include a presentation device 1416 and an associated graphical user interface (GUI) 1418. In other embodiments, the input / output module 1410 (I / O), input device 1412, and output device 1414 may not be included, and the computer device 1402 may be used as a computer device in a network. The computer device 1402 may also include one or more network interfaces 1420 for exchanging data with other devices via one or more communication links 1423. One or more communication buses 1424 couple the components described above together.

[0125] The communication link 1423 may be implemented in any manner, for example, through a local area network, a wide area network (e.g., the Internet), a point-to-point connection, etc., or any combination thereof. The communication link 1423 may include any combination of hardwired links, wireless links, routers, gateway functions, name servers, etc. governed by any protocol or combination of protocols.

[0126] Corresponds to Figures 1 to 7 The method in this article also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method are executed.

[0127] The embodiment of the present invention also provides a computer readable instruction, wherein when the processor executes the instruction, the program therein causes the processor to execute the following Figures 1 to 7 The method shown.

[0128] It should be understood that in the various embodiments of this document, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this document.

[0129] It should also be understood that in the embodiments of this article, the term "and / or" is only a description of the association relationship of the associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0130] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this article.

[0131] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0132] In the several embodiments provided herein, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, or can be electrical, mechanical or other forms of connection.

[0133] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of this article.

[0134] In addition, each functional unit in each embodiment of this invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit may be implemented in the form of hardware or in the form of software functional unit.

[0135] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this article is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of this article. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0136] Specific embodiments are used in this article to illustrate the principles and implementation methods of this article. The description of the above embodiments is only used to help understand the methods and core ideas of this article. At the same time, for general technicians in this field, according to the ideas of this article, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as a limitation on this article.

Claims

1. A method for constructing a model for predicting the depth of karst gullies, characterized in that: The method comprises: Determine a first mapping group of the sample according to the incision depth of the sample dissolution valley and the seismic attribute values ​​of different types of seismic attributes at different incision depths; According to the correlation between all seismic attributes in the first sample mapping group and the incision depth of the karst gully, a seismic attribute with a correlation greater than a preset correlation threshold is selected as the first seismic attribute, and the seismic attribute value of the first seismic attribute and different incision depths form a second sample mapping group; Using the second sample mapping group as a training sample data set; The training sample data set is trained using a random forest regression algorithm to establish a karst gully incision depth prediction model that reflects the mapping relationship between the first seismic attribute and the sample karst gully incision depth.

2. The method for constructing a karst gully incision depth prediction model according to claim 1, characterized in that: According to the incision depth of the sample dissolution valley and the seismic attributes of different types of seismic attributes at different incision depths, determining the first mapping group of the sample includes: Acquire a three-dimensional sample seismic data volume, well logging sample data and sample seismic interpretation horizon of the sample dissolution valley, and establish a forward modeling sample seismic profile of the sample dissolution valley; According to the sample seismic interpretation horizon of the sample dissolution valley, the seismic attribute values ​​of each seismic attribute at different incision depths of the sample dissolution valley on the forward modeling sample seismic section in the time domain are read; According to the sample seismic interpretation layer of the sample dissolution valley, the depth values ​​of the sample dissolution valley at different incision depths on the sample dissolution valley forward model in the depth domain are read, and the depth values ​​and the seismic attribute values ​​of each seismic attribute form a first sample mapping group.

3. The method for constructing a karst gully incision depth prediction model according to claim 2, characterized in that: Determining the correlation between seismic attributes and the depth of incision of karst valleys includes: The first sample mapping group is cross-analyzed to determine the fitting coefficients of different incision depths of the sample dissolution valleys in each sample mapping group and the seismic attributes, and to obtain the correlation between different types of seismic attributes and the incision depths of the dissolution valleys.

4. The method for constructing a karst gully incision depth prediction model according to claim 2, characterized in that: According to the three-dimensional sample seismic data volume, well logging sample data and sample seismic interpretation horizon of the sample dissolution valley, a forward modeling sample seismic profile of the sample dissolution valley is determined, and the method comprises: Analyze the well logging sample data and the three-dimensional sample seismic data to determine the geological characteristics of the sample dissolution valley, wherein the geological characteristics include the width range and depth range of the sample dissolution valley; According to the width range and depth range of the sample dissolution valley, forward modeling models of sample dissolution valleys with different incision depths and different widths are respectively constructed; The forward modeling method of the wave equation is used to determine the forward modeling seismic profile of the sample dissolution valley corresponding to the forward modeling model of the sample dissolution valley with different incision depths and different widths.

5. The method for constructing a karst gully incision depth prediction model according to claim 1, characterized in that: The different cutting depths include: a maximum cutting depth, a half cutting depth, and a quarter cutting depth.

6. The method for constructing a model for predicting the depth of dissolution gullies according to claim 4, characterized in that: After constructing the forward model with different undercut depths and different widths, the method further includes: Determine the drilling geological layers corresponding to the top and bottom surfaces of each layer segment in the forward model of the sample dissolution gully; Determine the velocity and density parameters of each layer according to the drilling geological layers corresponding to the top surface and the bottom surface; The velocity and density of the sample dissolution valley forward model are determined according to the density and velocity.

7. A method for predicting the depth of incision of a karst gully, characterized in that: The method predicts the incision depth using the incision depth prediction model of the erosion gully according to claims 1 to 5, comprising: Obtain 3D seismic data volume, well logging data and seismic interpretation horizons of the karst gullies in the target work area; Extract the first seismic attribute from the 3D seismic data volume of the karst gully in the target work area; The value of the first seismic attribute at the target layer of the three-dimensional seismic data volume is input into the karst gully incision depth prediction model to obtain the predicted incision depth of the karst gully in the target work area.

8. The method for predicting the depth of the erosion valley according to claim 7, characterized in that: Before inputting the value of the first seismic attribute into the karst valley incision depth prediction model, the method includes: Extracting the first seismic attribute along-layer slices from the seismic interpretation horizon of the karst gully in the target work area to obtain the plane prediction range of the karst gully in the target work area; Creating a distribution polygon of the karst gullies in the plane prediction range of the karst gullies in the target work area, and determining the plane distribution constraint boundary of the karst gullies in the target work area; The seismic attribute value of the first seismic attribute of the target layer at each CDP within the plane distribution constraint boundary of the karst gully in the target work area is read, and the seismic attribute value of the first seismic attribute is input into the karst gully incision depth prediction model.

9. A device for constructing a model for predicting the depth of incision of a karst gully, characterized in that: The device comprises: A sample first mapping group determination unit, used to determine the sample first mapping group according to the incision depth of the sample karst valley and the seismic attributes of different types of seismic attributes at different incision depths; A second sample mapping group determination unit is used to select a seismic attribute with a correlation greater than a preset correlation threshold as a first seismic attribute according to the correlation between all seismic attributes in the first sample mapping group and the incision depth of the karst gully, and the seismic attribute value of the first seismic attribute and different incision depths form a second sample mapping group; A training sample data set determining unit, configured to use the second sample mapping group as a training sample data set; The karst gully incision depth prediction model determination unit is used to train the sample data set using a random forest regression algorithm, establish a mapping relationship between the first seismic attribute and the sample karst gully incision depth, and obtain a karst gully incision depth prediction model.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 8 is implemented.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images