A method and apparatus for predicting reservoir thickness

Abstract

This application provides a method and apparatus for predicting reservoir thickness. The method includes: acquiring well logging information and seismic information, along with corresponding sample category labels, from a seismic database of the area to be measured, as well as seismic attribute data of the reservoir to be measured; expanding the dominant frequency and bandwidth of the seismic information based on the well logging information and seismic information to obtain expanded seismic bandwidth information; calculating the correlation between the seismic attribute data and the well logging information based on the well logging information, seismic information, and expanded seismic bandwidth information; constructing a reservoir prediction model based on the well logging information, seismic information, corresponding sample category labels, and correlation data; and inputting the seismic attribute data of the reservoir to be measured into the reservoir prediction model to obtain the reservoir thickness prediction result. Based on bidirectional frequency-extended seismic data, this method solves the technical problem that existing reservoir thickness prediction methods cannot accurately predict reservoir thickness due to the narrow bandwidth and low dominant frequency width of deep seismic data.

Description

Technical Field

[0001] This invention relates to the field of reservoir thickness prediction technology, and in particular to a method and apparatus for predicting reservoir thickness. Background Technology

[0002] Seismic exploration practice has shown that there are various direct or indirect complex statistical relationships between seismic attributes and stratigraphic characteristics, and thin reservoir information is generally included in such statistical relationships.

[0003] Reservoir thickness refers to the thickness of the oil-producing portion of an oil-bearing layer, i.e., the thickness of the oil-bearing layer with movable oil. The study of reservoir thickness boundaries essentially involves determining the boundary between oil-bearing and non-oil-bearing layers. This should be based on core analysis, single-layer well testing, and full utilization of well logging interpretation data. Specifically, through the analysis of rock physical parameters, pore structure, and relative permeability relationships, the lower limits of porosity, permeability, and oil saturation of movable oil-bearing reservoirs are determined. Based on this, qualitative and quantitative interpretation of well logging data is conducted to establish well logging standards for defining the effective thickness of oil-bearing layers. Alternatively, single-layer test data can be directly used to establish interpretation standards for the effective logging thickness boundaries.

[0004] In reality, deep seismic frequencies are often narrow and the dominant frequency width is not high, leading to technical defects such as excessive reliance on well data in reservoir prediction and excessive human influence during operation. Summary of the Invention

[0005] This invention provides a method and apparatus for predicting reservoir thickness, which is based on the prediction of deep thin reservoirs using two-way extended frequency seismic data. This solves the technical problem that existing reservoir thickness prediction methods cannot accurately predict reservoir thickness due to the narrow frequency band and low dominant frequency width of deep seismic data.

[0006] In a first aspect, the present invention provides a method for predicting reservoir thickness, comprising:

[0007] Acquire reservoir prediction training samples from the seismic database of the area to be measured, as well as seismic attribute data of the reservoir to be measured; the reservoir prediction training samples include well logging information and seismic information and corresponding sample category labels;

[0008] Based on the well logging information and the seismic information, the dominant frequency and bandwidth of the seismic frequency band information in the seismic information are expanded to obtain the expanded seismic frequency band information;

[0009] Based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data in the seismic information, combined with the extended seismic frequency band information, the correlation data between the seismic attribute data and the well logging curve data, the lithology category labels, and the reservoir thickness data is calculated.

[0010] Based on the well logging information and seismic information, the corresponding sample category labels, and the correlation data, a reservoir prediction model is constructed.

[0011] The seismic attribute data of the reservoir to be tested is input into the reservoir prediction model to obtain the reservoir thickness prediction result of the reservoir to be tested.

[0012] Optionally, based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data in the seismic information, combined with the extended seismic frequency band information, correlation data between the seismic attribute data and the well logging curve data, the lithology category labels, and the reservoir thickness data is calculated, including:

[0013] Statistical analysis is performed on the logging curve data in the logging information to obtain the lithology category label and the reservoir thickness data;

[0014] Seismic attribute data of the area to be measured are extracted from the seismic interpretation horizon data in the seismic information.

[0015] Based on the well logging data, the lithology category label, the reservoir thickness data, and the seismic attribute data, combined with the extended seismic frequency band information, the correlation data between the seismic attribute data and the well logging data, the lithology category label, and the reservoir thickness data is calculated.

[0016] Optionally, based on the well logging information and seismic information, the corresponding sample category labels, and the correlation data, a reservoir prediction model is constructed, including:

[0017] Based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data and correlation data in the seismic information, a preliminary reservoir prediction model is constructed.

[0018] Based on the well logging information and seismic information and the corresponding sample category labels, the preliminary reservoir prediction model is trained to obtain the trained preliminary reservoir prediction model.

[0019] Based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data and corresponding sample category labels in the seismic information, the trained preliminary reservoir prediction model is verified to obtain the reservoir prediction model.

[0020] Optionally, based on the well logging information and seismic information and the corresponding sample category labels, the preliminary reservoir prediction model is trained to obtain the trained preliminary reservoir prediction model, including:

[0021] The seismic attribute data from the seismic information is input into the preliminary reservoir prediction model to generate corresponding sample categories;

[0022] The training error is determined based on the seismic attribute data and corresponding sample category labels and sample categories in the well logging information and seismic information;

[0023] Based on the training error, the preliminary reservoir prediction model is adjusted to obtain the optimal parameters, and the preliminary reservoir prediction model is optimized using the optimal parameters to obtain the trained preliminary reservoir prediction model.

[0024] In a second aspect, the present invention provides a reservoir thickness prediction device, comprising:

[0025] The acquisition module is used to acquire reservoir prediction training samples from the seismic database of the area to be measured, as well as seismic attribute data of the reservoir to be measured; the reservoir prediction training samples include well logging information and seismic information and corresponding sample category labels;

[0026] The expansion module is used to expand the dominant frequency and bandwidth of the seismic frequency band information in the seismic information based on the well logging information and the seismic information, so as to obtain the expanded seismic frequency band information;

[0027] The calculation module is used to calculate the correlation data between the seismic attribute data and the well logging curve data, the lithology category label and the reservoir thickness data based on the well logging curve data, the lithology category label and the reservoir thickness data in the well logging information, and the seismic attribute data in the seismic information, combined with the extended seismic frequency band information.

[0028] The construction module is used to construct a reservoir prediction model based on the well logging information, seismic information, corresponding sample category labels, and correlation data;

[0029] The prediction module is used to input the seismic attribute data of the reservoir to be tested into the reservoir prediction model to obtain the reservoir thickness prediction result of the reservoir to be tested.

[0030] Optionally, the computing module includes:

[0031] The statistics submodule is used to perform statistics on the logging curve data in the logging information to obtain the lithology category label and the reservoir thickness data;

[0032] An extraction submodule is used to extract seismic attribute data of the area to be measured from the seismic interpretation horizon data in the seismic information.

[0033] The calculation submodule is used to calculate the correlation data between the seismic attribute data and the well logging data, the lithology category label, the reservoir thickness data, and the seismic attribute data, combined with the extended seismic frequency band information.

[0034] Optionally, the building module includes:

[0035] A submodule is constructed to build a preliminary reservoir prediction model based on the well logging curve data, lithology category labels and reservoir thickness data in the well logging information, and the seismic attribute data and correlation data in the seismic information.

[0036] The training submodule is used to train the preliminary reservoir prediction model based on the well logging information, seismic information and corresponding sample category labels, so as to obtain the trained preliminary reservoir prediction model.

[0037] The verification submodule is used to verify the trained preliminary reservoir prediction model based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data and corresponding sample category labels in the seismic information, so as to obtain the reservoir prediction model.

[0038] Thirdly, this application provides an electronic device including a processor and a memory, the memory storing computer-readable instructions that, when executed by the processor, perform the steps of the method provided in the first aspect above.

[0039] Fourthly, this application provides a storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps of the method provided in the first aspect above.

[0040] Fifthly, this application provides a computer program product that, when executed by a processor, performs the steps of the method provided in the first aspect above.

[0041] As can be seen from the above technical solutions, the present invention has the following advantages:

[0042] This invention provides a method and apparatus for predicting reservoir thickness, comprising: acquiring reservoir prediction training samples from a seismic database of a region to be measured, and seismic attribute data of the reservoir to be measured, wherein the reservoir prediction training samples include well logging information and seismic information and corresponding sample category labels; expanding the dominant frequency and bandwidth of the seismic frequency band information in the seismic information based on the well logging information and the seismic information to obtain expanded seismic frequency band information; calculating the correlation data between the seismic attribute data and the well logging data, the lithological category labels, and the reservoir thickness data based on the well logging information, the seismic attribute data in the seismic information, and the expanded seismic frequency band information; constructing a reservoir prediction model based on the well logging information, the seismic information, the corresponding sample category labels, and the correlation data; and inputting the seismic attribute data of the reservoir to be measured into the reservoir prediction model to obtain the reservoir thickness prediction result of the reservoir to be measured. This method is based on a two-way extended frequency seismic data method to predict deep thin reservoirs, which solves the technical problem that existing reservoir thickness prediction methods cannot accurately predict reservoir thickness due to the narrow frequency band and low dominant frequency width of deep seismic data. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 This is a flowchart illustrating the steps of a reservoir thickness prediction method according to an embodiment of the present invention.

[0045] Figure 2 This is a flowchart illustrating the steps of a second embodiment of the reservoir thickness prediction method of the present invention.

[0046] Figure 3 This is a schematic diagram showing the effect of 3D seismic frequency topology processing in region A before the processing.

[0047] Figure 4 This is a schematic diagram showing the effect of 3D seismic frequency conversion processing in region A.

[0048] Figure 5 A schematic diagram illustrating the correlation between seismic attribute data and well logging curve data, lithological category labels, and reservoir thickness data;

[0049] Figure 6 This is a flowchart illustrating the third embodiment of the reservoir thickness prediction method of the present invention.

[0050] Figure 7 This is a flowchart illustrating the steps of a fourth embodiment of the reservoir thickness prediction method of the present invention.

[0051] Figure 8 This is a schematic diagram illustrating the correlation analysis results between reservoir model predictions and actual results.

[0052] Figure 9 A schematic diagram of the predicted lithological probability of the sandstone and mudstone section in the middle of Group B;

[0053] Figure 10 This is a plane contour map of the predicted thickness of the sandstone in the middle of Group B;

[0054] Figure 11 This is a structural block diagram of an embodiment of a reservoir thickness prediction device according to the present invention. Detailed Implementation

[0055] This invention provides a method and apparatus for predicting reservoir thickness, which is based on two-way frequency-spreading seismic data for predicting deep thin reservoirs. This solves the technical problem that existing reservoir thickness prediction methods cannot accurately predict reservoir thickness due to the narrow frequency band and low dominant frequency width of deep seismic data.

[0056] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0057] Example 1, please refer to Figure 1 , Figure 1 This is a flowchart illustrating the steps of a reservoir thickness prediction method according to an embodiment of the present invention, including:

[0058] Step S101: Obtain reservoir prediction training samples from the seismic database of the area to be tested, as well as seismic attribute data of the reservoir to be tested; the reservoir prediction training samples include well logging information and seismic information and corresponding sample category labels;

[0059] Step S102: Based on the well logging information and the seismic information, expand the dominant frequency and bandwidth of the seismic frequency band information in the seismic information to obtain the expanded seismic frequency band information;

[0060] Step S103: Based on the well logging curve data, lithology category label and reservoir thickness data in the well logging information, and the seismic attribute data in the seismic information, combined with the extended seismic frequency band information, calculate the correlation data between the seismic attribute data and the well logging curve data, the lithology category label and the reservoir thickness data.

[0061] Step S104: Construct a reservoir prediction model based on the well logging information, seismic information, corresponding sample category labels, and correlation data;

[0062] Step S105: Input the seismic attribute data of the reservoir to be tested into the reservoir prediction model to obtain the reservoir thickness prediction result of the reservoir to be tested.

[0063] The reservoir thickness prediction method provided in this embodiment of the invention involves acquiring reservoir prediction training samples from a seismic database of the area to be measured, as well as seismic attribute data of the reservoir to be measured. The reservoir prediction training samples include well logging information, seismic information, and corresponding sample category labels. Based on the well logging information and the seismic information, the dominant frequency and bandwidth of the seismic frequency band information in the seismic information are expanded to obtain expanded seismic frequency band information. Based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data in the seismic information, combined with the expanded seismic frequency band information, the correlation data between the seismic attribute data and the well logging curve data, the lithology category labels, and the reservoir thickness data are calculated. Based on the well logging information, the seismic information, the corresponding sample category labels, and the correlation data, a reservoir prediction model is constructed. The seismic attribute data of the reservoir to be measured is input into the reservoir prediction model to obtain the reservoir thickness prediction result of the reservoir to be measured. This method is based on a two-way extended frequency seismic data method to predict deep thin reservoirs, which solves the technical problem that existing reservoir thickness prediction methods cannot accurately predict reservoir thickness due to the narrow frequency band and low dominant frequency width of deep seismic data.

[0064] Example 2, please refer to Figure 2 , Figure 2 This is a flowchart illustrating the steps of a second embodiment of the reservoir thickness prediction method of the present invention, including:

[0065] Step S201: Obtain reservoir prediction training samples from the seismic database of the area to be tested, as well as seismic attribute data of the reservoir to be tested; the reservoir prediction training samples include well logging information and seismic information and corresponding sample category labels;

[0066] Step S202: Based on the well logging information and the seismic information, expand the dominant frequency and bandwidth of the seismic frequency band information in the seismic information to obtain the expanded seismic frequency band information;

[0067] In this embodiment of the invention, based on well logging information and seismic information, parameters such as the burial depth of the target layer (reservoir to be measured) in the study area, seismic frequency band information, and well data frequency band information are determined, such as Butterworth wavelet and other parameters. On the basis of well control, the low-frequency information of the seismic data is first determined and the high-frequency part is expanded. Then, the high-frequency part of the seismic data is determined and the low-frequency information is expanded to obtain seismic data with a significant improvement in both the dominant frequency and the bandwidth, i.e., the expanded seismic frequency band information.

[0068] For a detailed implementation, please refer to Figures 3-4 , Figure 3 This is a schematic diagram showing the effect of 3D seismic frequency topology processing in region A before the processing. Figure 4 This is a schematic diagram of the effect of three-dimensional seismic frequency topology processing in region A. Two-way frequency topology processing was carried out using the seismic processing results data of region A to enhance the dominant frequency and bandwidth of seismic data in group B, and improve the abundance of seismic data response to stratigraphic information.

[0069] Step S203: Statistically analyze the logging curve data in the logging information to obtain lithology category labels and reservoir thickness data;

[0070] In this embodiment of the invention, electrical logging curves such as natural gamma, spontaneous potential, sonic velocity, and density are used to statistically analyze the lithological category label and reservoir thickness information of the target layer (the reservoir to be tested).

[0071] Step S204: Extract seismic attribute data of the area to be measured from the seismic interpretation horizon data in the seismic information;

[0072] In this embodiment of the invention, various seismic attribute data of the target segment are extracted from the interpretation layer data in the seismic information, including statistical seismic attributes such as amplitude, frequency, and phase.

[0073] Step S205: Based on the well logging curve data, the lithology category label, the reservoir thickness data, and the seismic attribute data, and in conjunction with the extended seismic frequency band information, calculate the correlation data between the seismic attribute data and the well logging curve data, the lithology category label, and the reservoir thickness data;

[0074] In this embodiment of the invention, the correlation data between the seismic attribute data and the well logging data, the lithology category label, the reservoir thickness data, the seismic attribute data, and the extended seismic frequency band information are calculated.

[0075] For a detailed implementation, please refer to Figure 5 , Figure 5This diagram illustrates the correlation between seismic attribute data and well logging data, lithological category labels, and reservoir thickness data. The study investigates the development of sandstone reservoirs, calcareous bands, and other non-reservoir elements in Group B of Area A. Lithological classification and reservoir thickness determination were achieved using sonic velocity, gamma curves, and density curves. Different lithological category labels were constructed, and well-side seismic attributes were extracted. Cross-plot analysis was then performed to determine the correlation between seismic attributes and well logging data, lithology, and reservoir thickness, yielding the correlation data between seismic attribute data and well logging data, lithological category labels, and reservoir thickness data.

[0076] Step S206: Construct a reservoir prediction model based on the well logging information, seismic information, corresponding sample category labels, and correlation data;

[0077] Step S207: Input the seismic attribute data of the reservoir to be tested into the reservoir prediction model to obtain the reservoir thickness prediction result of the reservoir to be tested.

[0078] The reservoir thickness prediction method provided in this embodiment of the invention involves acquiring reservoir prediction training samples from a seismic database of the area to be measured, as well as seismic attribute data of the reservoir to be measured. The reservoir prediction training samples include well logging information, seismic information, and corresponding sample category labels. Based on the well logging information and the seismic information, the dominant frequency and bandwidth of the seismic frequency band information in the seismic information are expanded to obtain expanded seismic frequency band information. Based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data in the seismic information, combined with the expanded seismic frequency band information, the correlation data between the seismic attribute data and the well logging curve data, the lithology category labels, and the reservoir thickness data are calculated. Based on the well logging information, the seismic information, the corresponding sample category labels, and the correlation data, a reservoir prediction model is constructed. The seismic attribute data of the reservoir to be measured is input into the reservoir prediction model to obtain the reservoir thickness prediction result of the reservoir to be measured. This method is based on a two-way extended frequency seismic data method to predict deep thin reservoirs, which solves the technical problem that existing reservoir thickness prediction methods cannot accurately predict reservoir thickness due to the narrow frequency band and low dominant frequency width of deep seismic data.

[0079] Example 3, please refer to Figure 6 , Figure 6 This is a flowchart illustrating the steps of a third embodiment of the reservoir thickness prediction method of the present invention, including:

[0080] Step S301: Obtain reservoir prediction training samples from the seismic database of the area to be tested, as well as seismic attribute data of the reservoir to be tested; the reservoir prediction training samples include well logging information and seismic information and corresponding sample category labels;

[0081] Step S302: Based on the well logging information and the seismic information, expand the dominant frequency and bandwidth of the seismic frequency band information in the seismic information to obtain the expanded seismic frequency band information;

[0082] Step S303: Based on the well logging curve data, lithology category label and reservoir thickness data in the well logging information, and the seismic attribute data in the seismic information, combined with the extended seismic frequency band information, calculate the correlation data between the seismic attribute data and the well logging curve data, the lithology category label and the reservoir thickness data.

[0083] In this embodiment of the invention, the well logging curve data in the well logging information are statistically analyzed to obtain the lithology category label and the reservoir thickness data. Seismic attribute data of the area to be measured is extracted from the seismic interpretation stratigraphic data in the seismic information. Based on the well logging curve data, the lithology category label, the reservoir thickness data, and the seismic attribute data, combined with the extended seismic frequency band information, the correlation data between the seismic attribute data and the well logging curve data, the lithology category label, and the reservoir thickness data is calculated.

[0084] Step S304: Construct a preliminary reservoir prediction model based on the well logging curve data, lithology category label, reservoir thickness data, seismic attribute data, and correlation data in the well logging information;

[0085] In this embodiment of the invention, a machine learning model is constructed. Based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, the seismic attribute data and correlation data in the seismic information, a preliminary reservoir prediction model is constructed with lithofacies classification and sandstone thickness as the objective function.

[0086] In the specific implementation, the modeling steps are as follows: Initial selection of seismic attributes: Taking the logging curves that are sensitive to lithology and reservoir thickness as the target, the statistical relationship between seismic attributes and logging curves is clarified by using methods such as cross plots, histograms, and scatter plots, the lower limit of correlation is determined, and the best seismic attributes are selected for further analysis.

[0087] Step S305: Based on the well logging information and seismic information and the corresponding sample category labels, train the preliminary reservoir prediction model to obtain the trained preliminary reservoir prediction model;

[0088] In this embodiment of the invention, the preliminary reservoir prediction model is trained based on the well logging information, seismic information, and corresponding sample category labels to obtain the trained preliminary reservoir prediction model.

[0089] In the specific implementation, the training steps for the prediction model are as follows: Using statistical data from well points, including sandstone thickness and lithofacies classification, as well as the seismic attributes selected in the previous step, the data is randomly divided proportionally. One part is used for the machine learning model, and the other part is used to test the model's training stability and accuracy. This deep learning model is based on a Wide&Deep hybrid model, which is a hybrid model composed of a single-layer Wide part and multiple-layer Deep parts. The Wide part gives the model strong memorization capabilities, while the Deep part gives it generalization capabilities. This model combines the advantages of logistic regression and deep neural networks. After the model is trained and reaches the set accuracy, the results are used for supervised learning classification or regression. For a complete Wide&Deep model, its training process consists of the following three steps: First, training the Wide part, which is essentially a generalized linear model used for calculating lithology-related inversion attribute volumes. Second, training the Deep part, which is mainly an Embedding+MLP neural network model. Large-scale sparse features are transformed into low-dimensional dense features through embedding. The features are then concatenated and input into the MLP to uncover hidden data patterns. AdaGrad is used in the DNN model of the Deep part to obtain more accurate solutions. The third step is joint training. The Wide & Deep models combine their outputs for training. During training, the outputs of the Deep and Wide parts are reused in a logistic regression model to make the final prediction and output probability values. Constructing different levels of learning methods is crucial. In multi-attribute thin-layer prediction, some attributes can directly indicate formation lithology and reservoir thickness and can be trained in the Wide part. Other attributes require deep learning to effectively indicate formation lithology and reservoir thickness and can be learned in the Deep part, increasing the breadth of data learning to obtain better learning results. In the training process of deep neural networks, if unsupervised training is used, the learning accuracy will be greatly reduced, and the obtained parameters are often local optima. By using drilling curves, lithology, and reservoir thickness as labels, local fine-tuning is performed during training. Specifically, the difference between the output of the learning model and the sample labels is used as the error, which is backpropagated to generate driving functions and optimize the neuron parameters of each layer so that the parameters of the entire network reach the optimal state.

[0090] Step S306: Based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data and corresponding sample category labels in the seismic information, verify the trained preliminary reservoir prediction model to obtain the reservoir prediction model;

[0091] In this embodiment of the invention, the trained preliminary reservoir prediction model is tested based on the well logging information, the seismic information, and the corresponding sample category labels. A "blind well" is randomly selected to test the reservoir prediction accuracy, verify the accuracy of the trained preliminary reservoir prediction model, and obtain the reservoir prediction model.

[0092] Step S307: Input the seismic attribute data of the reservoir to be tested into the reservoir prediction model to obtain the reservoir thickness prediction result of the reservoir to be tested.

[0093] The reservoir thickness prediction method provided in this embodiment of the invention involves acquiring reservoir prediction training samples from a seismic database of the area to be measured, as well as seismic attribute data of the reservoir to be measured. The reservoir prediction training samples include well logging information, seismic information, and corresponding sample category labels. Based on the well logging information and the seismic information, the dominant frequency and bandwidth of the seismic frequency band information in the seismic information are expanded to obtain expanded seismic frequency band information. Based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data in the seismic information, combined with the expanded seismic frequency band information, the correlation data between the seismic attribute data and the well logging curve data, the lithology category labels, and the reservoir thickness data are calculated. Based on the well logging information, the seismic information, the corresponding sample category labels, and the correlation data, a reservoir prediction model is constructed. The seismic attribute data of the reservoir to be measured is input into the reservoir prediction model to obtain the reservoir thickness prediction result of the reservoir to be measured. This method is based on a two-way extended frequency seismic data method to predict deep thin reservoirs, which solves the technical problem that existing reservoir thickness prediction methods cannot accurately predict reservoir thickness due to the narrow frequency band and low dominant frequency width of deep seismic data.

[0094] Example 4, please refer to Figure 7 , Figure 7 This is a flowchart illustrating the steps of a fourth embodiment of the reservoir thickness prediction method of the present invention, including:

[0095] Step S401: Obtain reservoir prediction training samples from the seismic database of the area to be tested, as well as seismic attribute data of the reservoir to be tested; the reservoir prediction training samples include well logging information and seismic information and corresponding sample category labels;

[0096] Step S402: Based on the well logging information and the seismic information, expand the dominant frequency and bandwidth of the seismic frequency band information in the seismic information to obtain the expanded seismic frequency band information;

[0097] Step S403: Based on the well logging curve data, lithology category label and reservoir thickness data in the well logging information, and the seismic attribute data in the seismic information, combined with the extended seismic frequency band information, calculate the correlation data between the seismic attribute data and the well logging curve data, the lithology category label and the reservoir thickness data.

[0098] In this embodiment of the invention, the well logging curve data in the well logging information are statistically analyzed to obtain the lithology category label and the reservoir thickness data. Seismic attribute data of the area to be measured is extracted from the seismic interpretation stratigraphic data in the seismic information. Based on the well logging curve data, the lithology category label, the reservoir thickness data, and the seismic attribute data, combined with the extended seismic frequency band information, the correlation data between the seismic attribute data and the well logging curve data, the lithology category label, and the reservoir thickness data is calculated.

[0099] Step S404: Construct a preliminary reservoir prediction model based on the well logging curve data, lithology category label, reservoir thickness data, seismic attribute data, and correlation data in the well logging information;

[0100] In this embodiment of the invention, a machine learning model is constructed. Based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, the seismic attribute data and correlation data in the seismic information, a preliminary reservoir prediction model is constructed with lithofacies classification and sandstone thickness as the objective function.

[0101] Step S405: Input the seismic attribute data from the seismic information into the preliminary reservoir prediction model to generate the corresponding sample category;

[0102] Step S406: Determine the training error based on the seismic attribute data and corresponding sample category labels and sample categories in the well logging information and seismic information;

[0103] Step S407: Based on the training error, adjust the preliminary reservoir prediction model to obtain the optimal parameters, and use the optimal parameters to optimize the preliminary reservoir prediction model to obtain the trained preliminary reservoir prediction model.

[0104] In this embodiment of the invention, the difference between the output of the preliminary reservoir prediction model and the sample label is used as the error. This error is then backpropagated to generate a driving function, which optimizes and adjusts the neuron parameters of each layer to achieve the optimal state of the entire network parameters, thus obtaining the trained preliminary reservoir prediction model.

[0105] For a detailed implementation, please refer to Figure 8 , Figure 8This diagram illustrates the correlation between reservoir model predictions and actual results. A Wide & Deep network was constructed using seismic attributes in area A. Impedance and gamma attributes, which are highly correlated with curves and lithology, were set as initial neurons in the Wide part, while other attribute volumes were set as Deep neurons. Verification wells were randomly selected for deep learning, and some learning parameters were adjusted based on test errors to optimize the learning results.

[0106] Step S408: Based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data and corresponding sample category labels in the seismic information, verify the trained preliminary reservoir prediction model to obtain the reservoir prediction model;

[0107] In this embodiment of the invention, the trained preliminary reservoir prediction model is tested based on the well logging information, the seismic information, and the corresponding sample category labels. A "blind well" is randomly selected to test the reservoir prediction accuracy, verify the accuracy of the trained preliminary reservoir prediction model, and obtain the reservoir prediction model.

[0108] Step S409: Input the seismic attribute data of the reservoir to be tested into the reservoir prediction model to obtain the reservoir thickness prediction result of the reservoir to be tested.

[0109] In this embodiment of the invention, seismic attribute data of the reservoir to be measured is input into the reservoir prediction model to obtain the reservoir thickness prediction result of the reservoir to be measured.

[0110] For a detailed implementation, please refer to Figures 9-10 , Figure 9 This is a schematic diagram of the predicted lithology probability of the central sandstone and mudstone section of Group B. Figure 10 To predict the thickness of the sandstone in the middle of Group B, the optimal Wide&Deep model parameters were selected to complete the prediction of lithology and sandstone thickness in the whole area, and the thickness planar maps of each sand body were compiled to obtain the reservoir thickness prediction results of the reservoir to be tested.

[0111] The reservoir thickness prediction method provided in this embodiment of the invention involves acquiring reservoir prediction training samples from a seismic database of the area to be measured, as well as seismic attribute data of the reservoir to be measured. The reservoir prediction training samples include well logging information, seismic information, and corresponding sample category labels. Based on the well logging information and the seismic information, the dominant frequency and bandwidth of the seismic frequency band information in the seismic information are expanded to obtain expanded seismic frequency band information. Based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data in the seismic information, combined with the expanded seismic frequency band information, the correlation data between the seismic attribute data and the well logging curve data, the lithology category labels, and the reservoir thickness data are calculated. Based on the well logging information, the seismic information, the corresponding sample category labels, and the correlation data, a reservoir prediction model is constructed. The seismic attribute data of the reservoir to be measured is input into the reservoir prediction model to obtain the reservoir thickness prediction result of the reservoir to be measured. This method is based on a two-way extended frequency seismic data method to predict deep thin reservoirs, which solves the technical problem that existing reservoir thickness prediction methods cannot accurately predict reservoir thickness due to the narrow frequency band and low dominant frequency width of deep seismic data.

[0112] Please see Figure 11 , Figure 11 A structural block diagram of an embodiment of a reservoir thickness prediction device according to the present invention includes:

[0113] The acquisition module 501 is used to acquire reservoir prediction training samples from the seismic database of the area to be measured, as well as seismic attribute data of the reservoir to be measured; the reservoir prediction training samples include well logging information and seismic information and corresponding sample category labels;

[0114] The expansion module 502 is used to expand the dominant frequency and bandwidth of the seismic frequency band information in the seismic information based on the well logging information and the seismic information, so as to obtain the expanded seismic frequency band information;

[0115] The calculation module 503 is used to calculate the correlation data between the seismic attribute data and the well logging curve data, the lithology category label and the reservoir thickness data based on the well logging curve data, the lithology category label and the reservoir thickness data in the well logging information, and the seismic attribute data in the seismic information, combined with the extended seismic frequency band information.

[0116] Construction module 504 is used to construct a reservoir prediction model based on the well logging information and seismic information, the corresponding sample category labels and the correlation data;

[0117] The prediction module 505 is used to input the seismic attribute data of the reservoir to be tested into the reservoir prediction model to obtain the reservoir thickness prediction result of the reservoir to be tested.

[0118] In an optional embodiment, the computing module 503 includes:

[0119] The statistics submodule is used to perform statistics on the logging curve data in the logging information to obtain the lithology category label and the reservoir thickness data;

[0120] An extraction submodule is used to extract seismic attribute data of the area to be measured from the seismic interpretation horizon data in the seismic information.

[0121] The calculation submodule is used to calculate the correlation data between the seismic attribute data and the well logging data, the lithology category label, the reservoir thickness data, and the seismic attribute data, combined with the extended seismic frequency band information.

[0122] In an optional embodiment, the building module 504 includes:

[0123] A submodule is constructed to build a preliminary reservoir prediction model based on the well logging curve data, lithology category labels and reservoir thickness data in the well logging information, and the seismic attribute data and correlation data in the seismic information.

[0124] The training submodule is used to train the preliminary reservoir prediction model based on the well logging information, seismic information and corresponding sample category labels, so as to obtain the trained preliminary reservoir prediction model.

[0125] The verification submodule is used to verify the trained preliminary reservoir prediction model based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data and corresponding sample category labels in the seismic information, so as to obtain the reservoir prediction model.

[0126] In an optional embodiment, the training submodule includes:

[0127] The generation unit is used to input the seismic attribute data in the seismic information into the preliminary reservoir prediction model and generate the corresponding sample category;

[0128] An error unit is used to determine the training error based on the seismic attribute data and corresponding sample category labels and sample categories in the well logging information and seismic information;

[0129] An optimization unit is used to adjust the preliminary reservoir prediction model based on the training error to obtain optimal parameters, and to optimize the preliminary reservoir prediction model using the optimal parameters to obtain the trained preliminary reservoir prediction model.

[0130] This invention also provides an electronic device, including a memory and a processor. The memory stores a computer program, which, when executed by the processor, causes the processor to perform the steps of the reservoir thickness prediction method as described in any of the above embodiments.

[0131] This invention also provides a computer storage medium storing a computer program thereon, which, when executed by the processor, implements the steps of the reservoir thickness prediction method as described in any of the above embodiments.

[0132] This invention also provides a computer program product storing a computer program, which, when executed by a processor, implements the steps of the method for predicting reservoir thickness under artificial intervention as described in any of the above embodiments.

[0133] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0134] In the several embodiments provided in this application, it should be understood that the methods, apparatuses, electronic devices, and storage media disclosed in this invention can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0135] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0136] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0137] If the integrated unit is implemented as 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 the present invention, in essence, 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. This computer software product is stored in a readable storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned readable storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0138] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting reservoir thickness, characterized in that, include Acquire reservoir prediction training samples from the seismic database of the area to be measured, as well as seismic attribute data of the reservoir to be measured; the reservoir prediction training samples include well logging information and seismic information and corresponding sample category labels; Based on the well logging information and the seismic information, the dominant frequency and bandwidth of the seismic frequency band information in the seismic information are expanded to obtain the expanded seismic frequency band information; Based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data in the seismic information, combined with the extended seismic frequency band information, the correlation data between the seismic attribute data and the well logging curve data, the lithology category labels, and the reservoir thickness data is calculated. Based on the well logging information and seismic information, the corresponding sample category labels, and the correlation data, a reservoir prediction model is constructed. The seismic attribute data of the reservoir to be tested is input into the reservoir prediction model to obtain the reservoir thickness prediction result of the reservoir to be tested. Based on the well logging information and seismic information, the corresponding sample category labels, and the correlation data, a reservoir prediction model is constructed, including: Based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data and correlation data in the seismic information, a preliminary reservoir prediction model is constructed. Based on the well logging information and seismic information and the corresponding sample category labels, the preliminary reservoir prediction model is trained to obtain the trained preliminary reservoir prediction model. Based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data and corresponding sample category labels in the seismic information, the trained preliminary reservoir prediction model is verified to obtain the reservoir prediction model.

2. The method for predicting reservoir thickness according to claim 1, characterized in that, Based on the well logging curve data, lithological category labels, and reservoir thickness data in the well logging information, and the seismic attribute data in the seismic information, combined with the extended seismic frequency band information, the correlation data between the seismic attribute data and the well logging curve data, the lithological category labels, and the reservoir thickness data is calculated, including: Statistical analysis is performed on the logging curve data in the logging information to obtain the lithology category label and the reservoir thickness data; Seismic attribute data of the area to be measured are extracted from the seismic interpretation horizon data in the seismic information. Based on the well logging data, the lithology category label, the reservoir thickness data, and the seismic attribute data, combined with the extended seismic frequency band information, the correlation data between the seismic attribute data and the well logging data, the lithology category label, and the reservoir thickness data is calculated.

3. The method for predicting reservoir thickness according to claim 1, characterized in that, Based on the well logging information and seismic information and the corresponding sample category labels, the preliminary reservoir prediction model is trained to obtain the trained preliminary reservoir prediction model, including: The seismic attribute data from the seismic information is input into the preliminary reservoir prediction model to generate corresponding sample categories; The training error is determined based on the seismic attribute data and corresponding sample category labels and sample categories in the well logging information and seismic information; Based on the training error, the preliminary reservoir prediction model is adjusted to obtain the optimal parameters, and the preliminary reservoir prediction model is optimized using the optimal parameters to obtain the trained preliminary reservoir prediction model.

4. A reservoir thickness prediction device, characterized in that, include: The acquisition module is used to acquire reservoir prediction training samples from the seismic database of the area to be measured, as well as seismic attribute data of the reservoir to be measured; the reservoir prediction training samples include well logging information and seismic information and corresponding sample category labels; The expansion module is used to expand the dominant frequency and bandwidth of the seismic frequency band information in the seismic information based on the well logging information and the seismic information, so as to obtain the expanded seismic frequency band information; The calculation module is used to calculate the correlation data between the seismic attribute data and the well logging curve data, the lithology category label and the reservoir thickness data based on the well logging curve data, the lithology category label and the reservoir thickness data in the well logging information, and the seismic attribute data in the seismic information, combined with the extended seismic frequency band information. The construction module is used to construct a reservoir prediction model based on the well logging information, seismic information, corresponding sample category labels, and correlation data; The prediction module is used to input the seismic attribute data of the reservoir to be tested into the reservoir prediction model to obtain the reservoir thickness prediction result of the reservoir to be tested. The building module includes: A submodule is constructed to build a preliminary reservoir prediction model based on the well logging curve data, lithology category labels and reservoir thickness data in the well logging information, and the seismic attribute data and correlation data in the seismic information. The training submodule is used to train the preliminary reservoir prediction model based on the well logging information, seismic information and corresponding sample category labels, so as to obtain the trained preliminary reservoir prediction model. The verification submodule is used to verify the trained preliminary reservoir prediction model based on the well logging curve data, lithology category labels, and reservoir thickness data in the well logging information, and the seismic attribute data and corresponding sample category labels in the seismic information, so as to obtain the reservoir prediction model.

5. The reservoir thickness prediction device according to claim 4, characterized in that, The computing module includes: The statistics submodule is used to perform statistics on the logging curve data in the logging information to obtain the lithology category label and the reservoir thickness data; An extraction submodule is used to extract seismic attribute data of the area to be measured from the seismic interpretation horizon data in the seismic information. The calculation submodule is used to calculate the correlation data between the seismic attribute data and the well logging data, the lithology category label, the reservoir thickness data, and the seismic attribute data, combined with the extended seismic frequency band information.

6. An electronic device, characterized in that, It includes a processor and a memory, the memory storing computer-readable instructions, which, when executed by the processor, perform the method as described in any one of claims 1-3.

7. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it performs the method as described in any one of claims 1-3.

8. A computer program product, characterized in that, When the computer program product is executed by a processor, it performs the method as described in any one of claims 1-3.

Images