Shale oil dessert intelligent evaluation method and device

Abstract

The application discloses a shale oil dessert intelligent evaluation method and device, wherein the method comprises the following steps: according to known well logging interpretation results and oil testing conditions, the dessert type of a shale oil layer section in a well is divided, and a dessert type label is formed; a wellside seismic trace containing the dessert type label is extracted, seismic attributes of the shale oil layer section are calculated, and a training data set is formed; a shale oil dessert evaluation model is constructed by using a machine learning algorithm, the input of the shale oil dessert evaluation model is the seismic attributes of the shale oil layer section, and the output is the dessert type of the shale oil layer section; the shale oil dessert evaluation model is trained according to the training data set, and a trained shale oil dessert evaluation model is obtained; the seismic attributes of a shale oil layer section of a label-free seismic trace are input into the trained shale oil dessert evaluation model, and the dessert type of the shale oil layer section of the label-free seismic trace is predicted. The application can improve the accuracy of the prediction result, reduce the difference of the prediction result, and improve the operation efficiency.

Description

Technical Field

[0001] This invention relates to the field of geophysical oil and gas exploration technology, and in particular to a method and apparatus for intelligent evaluation of shale oil sweet spots. Background Technology

[0002] This section is intended to provide background or context for the embodiments of the invention set forth in the claims. The description herein is not an admission that it is prior art simply because it is included in this section.

[0003] As oil and gas exploration continues to deepen, conventional oil and gas resource exploration has gradually shifted towards unconventional oil and gas exploration. Currently, shale oil sweet spot evaluation mainly focuses on: source rock quality, reservoir sweet spot quality, engineering sweet spot quality, as well as source rock characteristics (thickness, total organic carbon), reservoir lithology, physical properties (porosity, fractures), oil-bearing capacity, brittleness, and geostress characteristics. Shale oil is classified into several types: clastic, carbonate, and mixed sedimentary, and the key geophysical evaluation parameters differ for each type.

[0004] Existing methods for evaluating shale oil sweet spots include two main categories: parametric comprehensive evaluation and model comprehensive evaluation. The parametric comprehensive evaluation method primarily involves overlaying various evaluation parameters onto a graph, taking the intersection of the distribution of areas above the standard for all evaluation parameters, and combining this with the continuous distribution area and economic viability to determine the distribution of shale oil sweet spots. The model comprehensive evaluation method, based on the comprehensive evaluation parameter standards for shale oil layers and drilling and testing data, calculates the weighting coefficient for each individual parameter in the evaluation area, and then weighted sums these coefficients to calculate the comprehensive evaluation coefficient for the sweet spot. A higher comprehensive evaluation coefficient indicates a higher evaluation level for the sweet spot.

[0005] However, the above two methods have the following drawbacks: the accuracy of the prediction results is low, the difference in the prediction results is large, and the computational efficiency is low. Summary of the Invention

[0006] This invention provides a method for intelligent evaluation of shale oil sweetness, which improves the accuracy of prediction results, reduces the variability of prediction results, and improves computational efficiency. The method includes:

[0007] Based on known well logging interpretation results and oil testing conditions, the sweet spot types of shale oil intervals in the well are classified, and sweet spot type labels are formed.

[0008] Seismic traces near wells with dessert type tags are extracted, and the seismic properties of shale oil sections are calculated to form a training dataset.

[0009] A shale oil sweet spot evaluation model is constructed using machine learning algorithms. The input of the shale oil sweet spot evaluation model is the seismic attributes of the shale oil interval, and the output is the sweet spot type of the shale oil interval.

[0010] The shale oil sweet spot evaluation model is trained based on the training dataset to obtain a well-trained shale oil sweet spot evaluation model;

[0011] By inputting the seismic attributes of shale oil strata from unlabeled seismic traces into a trained shale oil sweet spot evaluation model, the sweet spot type of shale oil strata from unlabeled seismic traces can be predicted.

[0012] This invention also provides a smart evaluation device for shale oil sweetness, used to improve the accuracy of prediction results, reduce the variability of prediction results, and improve computational efficiency. The device includes:

[0013] The dessert type tagging module is used to classify the dessert type of shale oil intervals in the well based on known well logging interpretation results and oil testing results, and to generate dessert type tags.

[0014] The training dataset formation module is used to extract well-side seismic traces with dessert type labels, calculate the seismic properties of shale oil intervals, and form the training dataset.

[0015] A construction module is used to build a shale oil sweet spot evaluation model using machine learning algorithms. The input of the shale oil sweet spot evaluation model is the seismic attributes of the shale oil segment, and the output is the sweet spot type of the shale oil segment.

[0016] The training module is used to train the shale oil sweet spot evaluation model based on the training dataset, and obtain the trained shale oil sweet spot evaluation model.

[0017] The prediction module is used to input the seismic attributes of shale oil segments from unlabeled seismic traces into a trained shale oil sweet spot evaluation model to predict the sweet spot type of the shale oil segments from unlabeled seismic traces.

[0018] This invention also 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, it implements the above-described intelligent evaluation method for shale oil sweet spots.

[0019] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described intelligent evaluation method for shale oil sweet spots.

[0020] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described intelligent evaluation method for shale oil sweet spots.

[0021] Compared with existing technologies, this invention classifies shale oil zones in wells into sweet spot types based on known well logging interpretation results and oil testing data, forming sweet spot type labels. It extracts well-side seismic traces containing sweet spot type labels, calculates the seismic attributes of shale oil zones, and forms a training dataset. This enables the organic fusion of multiple information sources. A shale oil sweet spot evaluation model is constructed using machine learning algorithms. The input of this model is the seismic attributes of the shale oil zones, and the output is the sweet spot type of the shale oil zones. The model is trained using the training dataset to obtain a trained shale oil sweet spot evaluation model. The seismic attributes of shale oil zones from unlabeled seismic traces are input into the trained model to predict the sweet spot type of the unlabeled seismic traces. This comprehensively considers multiple seismic attributes, avoids calculating each single parameter, improves the accuracy of prediction results, reduces the variability of prediction results, and increases computational efficiency. Attached Figure Description

[0022] 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. In the drawings:

[0023] Figure 1 This is a flowchart illustrating the intelligent evaluation method for shale oil sweetness provided in an embodiment of the present invention;

[0024] Figure 2 This is a specific example diagram of a sweet spot type of shale oil formation provided in an embodiment of the present invention;

[0025] Figure 3 This is a specific example diagram of the training dataset provided in an embodiment of the present invention;

[0026] Figure 4 This is a specific example diagram of the intelligent evaluation method for shale oil sweetness provided in the embodiments of the present invention;

[0027] Figure 5 This is a specific example diagram of the input to the shale oil sweetness evaluation model provided in this embodiment of the invention;

[0028] Figure 6 This is a specific example of the shale oil sweet spot prediction result provided in an embodiment of the present invention;

[0029] Figure 7 This is a specific example diagram of the intelligent evaluation method for shale oil sweetness provided in the embodiments of the present invention;

[0030] Figure 8 This is a schematic diagram of the intelligent evaluation device for shale oil sweetness provided in an embodiment of the present invention;

[0031] Figure 9 This is a specific example diagram of the intelligent evaluation device for shale oil sweetness provided in the embodiments of the present invention. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0033] This invention provides a method for intelligent evaluation of shale oil sweetness. Figure 1 This is a flowchart illustrating the intelligent evaluation method for shale oil sweetness provided in an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes the following steps:

[0034] Step 101: Based on the known logging interpretation results and oil testing results, classify the sweet spot type of the shale oil interval in the well and form a sweet spot type label;

[0035] Step 102: Extract well-side seismic traces with dessert type labels, calculate the seismic properties of shale oil sections, and form a training dataset;

[0036] Step 103: Construct a shale oil sweet spot evaluation model using machine learning algorithms. The input of the shale oil sweet spot evaluation model is the seismic attributes of the shale oil strata, and the output is the sweet spot type of the shale oil strata.

[0037] Step 104: Train the shale oil sweet spot evaluation model based on the training dataset to obtain the trained shale oil sweet spot evaluation model;

[0038] Step 105: Input the seismic attributes of the shale oil strata in the unlabeled seismic traces into the trained shale oil sweet spot evaluation model to predict the sweet spot type of the shale oil strata in the unlabeled seismic traces.

[0039] Depend on Figure 1As shown in the flowchart, compared with the technical solutions in the prior art, the embodiments of the present invention classify the sweet spot types of shale oil intervals in the well according to the known well logging interpretation results and oil testing conditions, forming sweet spot type labels; extract the wellside seismic traces containing sweet spot type labels, calculate the seismic attributes of the shale oil intervals, and form a training dataset; this enables the organic integration of multiple information; a shale oil sweet spot evaluation model is constructed using machine learning algorithms, the input of which is the seismic attribute of the shale oil interval, and the output is the sweet spot type of the shale oil interval; the shale oil sweet spot evaluation model is trained according to the training dataset to obtain a trained shale oil sweet spot evaluation model; the seismic attributes of the shale oil intervals in the unlabeled seismic traces are input into the trained shale oil sweet spot evaluation model to predict the sweet spot type of the shale oil intervals in the unlabeled seismic traces; thus, multiple seismic attributes are comprehensively considered, avoiding the calculation of each single parameter, which can improve the accuracy of the prediction results, reduce the variability of the prediction results, and improve the computational efficiency.

[0040] In practice, the sweet spot type of the shale oil layer in the well is first classified according to the known logging interpretation results and oil testing results, and a sweet spot type label is formed. Figure 2 This is a specific example of a sweet spot type for a shale oil formation provided in an embodiment of the present invention. In this example, taking a tight oil formation as an example, the sweet spot type of a shale oil formation may include, for example, the following: multi-stage sand-stacked thick layer type (Type I), thick sand and thin mud interlayer type (Type II), and shale type with thin sand interlayers (Type III).

[0041] After generating dessert type labels, well-side seismic traces containing dessert type labels are extracted, and the seismic properties of shale oil sections are calculated to form a training dataset.

[0042] In one embodiment, the seismic properties of the shale oil zone include one or any combination of the following seismic properties: amplitude, frequency, phase, curvature, coherence, fracture density, fracture orientation, sand body thickness, total organic carbon, porosity, lithology, brittleness, oil content, and formation pressure.

[0043] In one embodiment, each row in the training dataset represents a sample, where a sample includes a known seismic attribute and a corresponding known dessert type; each column represents a feature, specifically, a feature may be, for example, a seismic attribute of a shale oil segment or a dessert type of a shale oil segment.

[0044] Figure 3This is a specific example diagram of the training dataset provided in this embodiment of the invention. In this example, the training dataset may have 90 samples, with each row representing one sample and each column representing one feature. A sample may include, for example, known seismic attributes and corresponding known sweet spot types. A feature may include, for example, survey line number, connecting line number, well name, reservoir thickness, reservoir porosity, root mean square amplitude, Poisson's ratio, brittleness, source rock thickness, total organic carbon content of source rock, and sweet spot type.

[0045] Figure 4 This is a specific example diagram of the intelligent evaluation method for shale oil sweetness provided in this embodiment of the invention, as shown in the figure. Figure 4 As shown, in this example, Figure 1 The process shown may also include the following steps:

[0046] Step 401: Preprocess the training dataset.

[0047] In one embodiment, the training dataset is preprocessed according to one or any combination of the following preprocessing methods: outlier removal, feature value range adjustment, type feature encoding, feature combination and optimization.

[0048] After forming the training dataset, a shale oil sweet spot evaluation model is constructed using machine learning algorithms. The input of the shale oil sweet spot evaluation model is the seismic attributes of the shale oil strata, and the output is the sweet spot type of the shale oil strata.

[0049] Figure 5 This is a specific example diagram of the input to the shale oil sweet spot evaluation model provided in this embodiment of the invention. In this example, the input to the shale oil sweet spot evaluation model may include one or any combination of the following seismic attributes of shale oil strata: source rock TOC, sand body thickness, porosity, oil content, brittleness, fractures, and pore pressure.

[0050] In one embodiment, the machine learning algorithm includes any one of the following: random forest algorithm, decision tree algorithm, logistic regression algorithm, support vector machine algorithm, and neural network algorithm.

[0051] After constructing a shale oil sweet spot evaluation model using machine learning algorithms, the model is trained using a training dataset to obtain a well-trained shale oil sweet spot evaluation model.

[0052] In one embodiment, training a shale oil sweet spot evaluation model based on a training dataset to obtain a trained shale oil sweet spot evaluation model includes: selecting a known cross-entropy function as a cost function to measure the difference between the predicted sweet spot type and the known sweet spot type; optimizing the parameters of the shale oil sweet spot evaluation model using a stochastic gradient descent algorithm based on the difference between the predicted sweet spot type and the known sweet spot type; determining the optimized parameters of the shale oil sweet spot evaluation model when the value of the cost function reaches the target value, thus obtaining the trained shale oil sweet spot evaluation model.

[0053] After obtaining the trained shale oil sweet spot evaluation model, the seismic attributes of the shale oil segments in the unlabeled seismic traces are input into the trained shale oil sweet spot evaluation model to predict the sweet spot type of the shale oil segments in the unlabeled seismic traces.

[0054] Figure 6 This is a specific example diagram of the shale oil sweetness prediction results provided in an embodiment of the present invention, as shown in the figure. Figure 6 As shown in the example, this study uses a tight oil formation as an example to predict the distribution map of sweet spot types in the tight oil formation. Figure 6 The Mu53, Li92 and shown Figure 6 The other numbers shown are all well numbers. The accuracy rate of blind well verification is 90%. The prediction results are basically consistent with the geological understanding of the area. The shale oil sweet spot evaluation method provided in this embodiment of the invention can be applied to guide the deployment of exploration and development well locations.

[0055] The following is a specific embodiment illustrating the application of the intelligent evaluation method for shale oil sweetness according to the present invention. Figure 7 This is a specific example diagram of the intelligent evaluation method for shale oil sweetness provided in this embodiment of the invention, as shown in the figure. Figure 7 As shown, in this example:

[0056] Based on known well logging interpretation results and oil testing, the sweet spot types of shale oil strata in the well are classified, and sweet spot type labels are formed.

[0057] Seismic traces near wells with "sweet spot" type tags are extracted, and seismic attributes of shale oil intervals are calculated to form a training dataset. The seismic attributes of shale oil intervals include one or any combination of the following seismic attributes: amplitude, frequency, phase, curvature, coherence, fracture density, fracture orientation, sand body thickness, total organic carbon, porosity, lithology, brittleness, oil content, and formation pressure.

[0058] The training dataset is preprocessed to obtain the processed training dataset. The preprocessing method is one or any combination of the following methods: outlier removal, feature value range adjustment, type feature encoding, feature combination and optimization.

[0059] A shale oil sweet spot evaluation model is constructed using machine learning algorithms. The input of the shale oil sweet spot evaluation model is the seismic attributes of the shale oil strata, and the output is the sweet spot type of the shale oil strata.

[0060] The shale oil sweet spot evaluation model is trained using a training dataset to obtain a trained shale oil sweet spot evaluation model. This process includes: selecting a known cross-entropy function as the cost function to measure the difference between the predicted sweet spot type and the known sweet spot type; optimizing the parameters of the shale oil sweet spot evaluation model using a stochastic gradient descent algorithm based on the difference between the predicted and known sweet spot types; and determining the optimized parameters of the shale oil sweet spot evaluation model when the cost function reaches a target value, thus obtaining the trained shale oil sweet spot evaluation model.

[0061] By inputting the seismic attributes of shale oil strata from unlabeled seismic traces into a trained shale oil sweet spot evaluation model, the sweet spot type of shale oil strata from unlabeled seismic traces can be predicted.

[0062] This invention also provides a smart evaluation device for shale oil sweetness, as described in the following embodiments. Since the principle by which this device solves the problem is similar to the smart evaluation method for shale oil sweetness, the implementation of this device can refer to the implementation of the smart evaluation method for shale oil sweetness, and repeated details will not be elaborated further.

[0063] This invention provides an intelligent evaluation device for shale oil sweetness. Figure 8 This is a schematic diagram of the intelligent evaluation device for shale oil sweetness provided in an embodiment of the present invention, as shown below. Figure 8 As shown, the device includes the following modules:

[0064] The dessert type label forming module 81 is used to classify the dessert type of shale oil intervals in the well based on known well logging interpretation results and oil testing results, and form dessert type labels;

[0065] Training dataset formation module 82 is used to extract well-side seismic traces with dessert type labels, calculate the seismic properties of shale oil intervals, and form a training dataset.

[0066] Module 83 is used to construct a shale oil sweet spot evaluation model using machine learning algorithms. The input of the shale oil sweet spot evaluation model is the seismic attributes of the shale oil segment, and the output is the sweet spot type of the shale oil segment.

[0067] Training module 84 is used to train the shale oil sweet spot evaluation model based on the training dataset to obtain the trained shale oil sweet spot evaluation model.

[0068] Prediction module 85 is used to input the seismic attributes of the shale oil strata in the unlabeled seismic traces into the trained shale oil sweet spot evaluation model, and predict the sweet spot type of the shale oil strata in the unlabeled seismic traces.

[0069] In one embodiment, the seismic properties of the shale oil zone include one or any combination of the following seismic properties: amplitude, frequency, phase, curvature, coherence, fracture density, fracture orientation, sand body thickness, total organic carbon, porosity, lithology, brittleness, oil content, and formation pressure.

[0070] Figure 9 This is a specific example diagram of the intelligent evaluation device for shale oil sweetness provided in the embodiments of the present invention, such as... Figure 9 As shown, in this example, Figure 8 The shale oil sweet spot intelligent evaluation device shown also includes:

[0071] The preprocessing module 91 is used to preprocess the training dataset before the training module 84 trains the shale oil sweet spot evaluation model based on the training dataset.

[0072] In one embodiment, the preprocessing module 91 is specifically used to preprocess the training dataset according to one or any combination of the following preprocessing methods: outlier removal, feature value range adjustment, type feature encoding, feature combination and optimization.

[0073] In one embodiment, the machine learning algorithm includes any one of the following: random forest algorithm, decision tree algorithm, logistic regression algorithm, support vector machine algorithm, and neural network algorithm.

[0074] In one embodiment, the training module 84 is specifically used to: select a known cross-entropy function as the cost function to measure the difference between the predicted dessert type and the known dessert type; based on the difference between the predicted dessert type and the known dessert type, use the stochastic gradient descent algorithm to optimize the parameters of the shale oil dessert evaluation model; when the value of the cost function reaches the target value, determine the optimized parameters of the shale oil dessert evaluation model, and obtain the trained shale oil dessert evaluation model.

[0075] This invention also 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, it implements the above-described intelligent evaluation method for shale oil sweet spots.

[0076] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described intelligent evaluation method for shale oil sweet spots.

[0077] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described intelligent evaluation method for shale oil sweet spots.

[0078] Compared with existing technologies, this invention classifies shale oil zones in wells into sweet spot types based on known well logging interpretation results and oil testing data, forming sweet spot type labels. It extracts well-side seismic traces containing sweet spot type labels, calculates the seismic attributes of shale oil zones, and forms a training dataset. This enables the organic fusion of multiple information sources. A shale oil sweet spot evaluation model is constructed using machine learning algorithms. The input of this model is the seismic attributes of the shale oil zones, and the output is the sweet spot type of the shale oil zones. The model is trained using the training dataset to obtain a trained shale oil sweet spot evaluation model. The seismic attributes of shale oil zones from unlabeled seismic traces are input into the trained model to predict the sweet spot type of the unlabeled seismic traces. This comprehensively considers multiple seismic attributes, avoids calculating each single parameter, improves the accuracy of prediction results, reduces the variability of prediction results, and increases computational efficiency.

[0079] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0080] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0081] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0082] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0083] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for intelligent evaluation of shale oil sweetness, characterized in that, include: Based on known well logging interpretation results and oil testing conditions, the sweet spot types of shale oil intervals in the well are classified, and sweet spot type labels are formed. The dessert types include: multi-stage sand-stacked thick-layered type, thick sand and thin mud interlayered type, and shale type with thin sand interlayers; Seismic traces near wells with dessert type labels are extracted, and the seismic attributes of shale oil intervals are calculated to form a training dataset. Each row in the training dataset represents a sample, where a sample includes known seismic attributes and the corresponding known dessert type. Each column represents a feature, which is either the seismic attribute of the shale oil interval or the dessert type of the shale oil interval. A shale oil sweet spot evaluation model is constructed using machine learning algorithms. The input of the shale oil sweet spot evaluation model is the seismic attributes of the shale oil interval, and the output is the sweet spot type of the shale oil interval. The shale oil sweet spot evaluation model is trained based on the training dataset to obtain a well-trained shale oil sweet spot evaluation model; The seismic attributes of shale oil strata in unlabeled seismic traces are input into a trained shale oil sweet spot evaluation model to predict the sweet spot type of shale oil strata in unlabeled seismic traces. The shale oil sweetness evaluation model is trained based on the training dataset to obtain a well-trained shale oil sweetness evaluation model, including: The known cross-entropy function is selected as the cost function to measure the difference between the predicted dessert type and the known dessert type; Based on the difference between the predicted dessert type and the known dessert type, the stochastic gradient descent algorithm is used to optimize the parameters of the shale oil dessert evaluation model. When the value of the cost function reaches the target value, the optimized parameters of the shale oil dessert evaluation model are determined, and the trained shale oil dessert evaluation model is obtained.

2. The intelligent evaluation method for shale oil sweetness as described in claim 1, characterized in that, The seismic properties of shale oil strata include one or any combination of the following seismic properties: Amplitude, frequency, phase, curvature, coherence, fracture density, fracture orientation, sand body thickness, total organic carbon, porosity, lithology, brittleness, oil content, and formation pressure.

3. The intelligent evaluation method for shale oil sweetness as described in claim 1, characterized in that, Before training the shale oil sweet spot evaluation model based on the training dataset, the following steps are also included: Preprocess the training dataset.

4. The intelligent evaluation method for shale oil sweetness as described in claim 3, characterized in that, Preprocess the training dataset using one or any combination of the following preprocessing methods: Outlier removal, feature value range adjustment, type feature encoding, and feature combination.

5. The intelligent evaluation method for shale oil sweetness as described in claim 1, characterized in that, Machine learning algorithms include any one of the following: random forest algorithm, decision tree algorithm, logistic regression algorithm, support vector machine algorithm, and neural network algorithm.

6. A smart evaluation device for shale oil sweetness, characterized in that, include: The dessert type tagging module is used to classify the dessert type of shale oil intervals in the well based on known well logging interpretation results and oil testing results, and to generate dessert type tags. The dessert types include: multi-stage sand-stacked thick-layered type, thick sand and thin mud interlayered type, and shale type with thin sand interlayers; The training dataset formation module is used to extract well-side seismic traces with dessert type labels, calculate the seismic attributes of shale oil intervals, and form a training dataset. Each row in the training dataset represents a sample, where a sample includes known seismic attributes and the corresponding known dessert type. Each column represents a feature, where a feature is either the seismic attribute of the shale oil interval or the dessert type of the shale oil interval. A construction module is used to build a shale oil sweet spot evaluation model using machine learning algorithms. The input of the shale oil sweet spot evaluation model is the seismic attributes of the shale oil segment, and the output is the sweet spot type of the shale oil segment. The training module is used to train the shale oil sweet spot evaluation model based on the training dataset, and obtain the trained shale oil sweet spot evaluation model. The prediction module is used to input the seismic attributes of the shale oil strata in the unlabeled seismic traces into the trained shale oil sweet spot evaluation model to predict the sweet spot type of the shale oil strata in the unlabeled seismic traces. The training module is specifically used for: The known cross-entropy function is selected as the cost function to measure the difference between the predicted dessert type and the known dessert type; Based on the difference between the predicted dessert type and the known dessert type, the stochastic gradient descent algorithm is used to optimize the parameters of the shale oil dessert evaluation model. When the value of the cost function reaches the target value, the optimized parameters of the shale oil dessert evaluation model are determined, and the trained shale oil dessert evaluation model is obtained.

7. The intelligent evaluation device for shale oil sweetness as described in claim 6, characterized in that, The seismic properties of shale oil strata include one or any combination of the following seismic properties: Amplitude, frequency, phase, curvature, coherence, fracture density, fracture orientation, sand body thickness, total organic carbon, porosity, lithology, brittleness, oil content, and formation pressure.

8. The intelligent evaluation device for shale oil sweetness as described in claim 6, characterized in that, It also includes a preprocessing module for use before the training module trains the shale oil sweetness evaluation model based on the training dataset: Preprocess the training dataset.

9. The intelligent evaluation device for shale oil sweetness as described in claim 8, characterized in that, The preprocessing module is specifically used to preprocess the training dataset according to one or any combination of the following preprocessing methods: Outlier removal, feature value range adjustment, type feature encoding, and feature combination.

10. The intelligent evaluation device for shale oil sweetness as described in claim 6, characterized in that, Machine learning algorithms include any one of the following: random forest algorithm, decision tree algorithm, logistic regression algorithm, support vector machine algorithm, and neural network algorithm.

11. 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, it implements the intelligent evaluation method for shale oil sweet spots according to any one of claims 1 to 5.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the intelligent evaluation method for shale oil sweet spots according to any one of claims 1 to 5.

13. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the intelligent evaluation method for shale oil sweet spots according to any one of claims 1 to 5.

Images