Pulse neutron gamma logging fracture-cavity identification method and device and medium

Abstract

The invention provides a pulsed neutron gamma logging fracture-cavity identification method and device and a medium, and relates to the technical field of petroleum and natural gas development, and the method comprises the steps: building a numerical calculation model according to an entity structure of a pulsed neutron gamma instrument and scale well parameters, so as to obtain first simulation response data of different detectors through simulation; performing precision optimization on the numerical calculation model through actually measured response data of the pulsed neutron gamma instrument in combination with the first simulation response data to obtain an optimized numerical calculation model; simulating by using the optimized numerical calculation model to obtain second simulation response data under various well-side cave conditions; and based on the second simulation response data, training a preset neural network model to obtain a pulsed neutron gamma logging fracture-cavity identification model. Therefore, the difficulty of performing fracture-cavity size calculation and fluid type identification by using the gamma information of the plurality of detectors is reduced, and the calculation accuracy is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of oil and gas development, and more specifically, to a method, device and medium for identifying fractures and vugs by pulsed neutron gamma logging. Background Art

[0002] Fractures and caves in carbonate reservoirs are the main storage spaces and seepage channels. The degree of fracture development, the filling degree of fractures and vugs, and the type of filling materials determine the storage capacity and permeability of the reservoir. Therefore, accurately identifying fractures and the filling materials of fractures and vugs is of guiding significance and plays a crucial role in the evaluation of carbonate reservoirs. Currently, the main methods for identifying fractures and vugs include dipole shear wave remote detection, seismic data AVO inversion, and electrical imaging logging. The identification of holes near the well and far from the well can already be achieved. However, there are still deficiencies in accurately calculating the size of fractures and vugs near the well and determining the fluid type and saturation in the caves.

[0003] Pulsed neutron gamma technology has been used in aspects such as calculating oil and gas saturation. However, it is difficult to determine the size of fractures and vugs near the well and the fluid type by combining the information of multiple gamma detectors. The effective areas of different detectors overlap, and the neutron gamma response is affected by multiple factors with complex rules. The error is relatively large when using the traditional ratio method to determine the size of fractures and vugs near the well and identify the fluid type.

[0004] In view of the problems of the existing technology, the present invention provides a method, device and medium for identifying fractures and vugs by pulsed neutron gamma logging. Summary of the Invention

[0005] In view of the problems of the existing technology, the present invention provides a method, device and medium for identifying fractures and vugs by pulsed neutron gamma logging, and the method includes:

[0006] Establishing a numerical calculation model based on the physical structure of the pulsed neutron gamma instrument and the calibration well parameters to simulate and obtain the first simulated response data of different detectors;

[0007] Optimizing the accuracy of the numerical calculation model by combining the measured response data of the pulsed neutron gamma instrument with the first simulated response data to obtain an optimized numerical calculation model;

[0008] Using the optimized numerical calculation model to simulate and obtain the second simulated response data under various conditions of caves near the well;

[0009] Training a preset neural network model based on the second simulated response data to obtain a pulsed neutron gamma logging fracture and vug identification model.

[0010] According to an embodiment of the present invention, the numerical calculation model is established through the following steps:

[0011] Determine the number of detectors, the physical structure parameters of the pulsed neutron gamma instrument, and the calibration well parameters, and establish a multi-dimensional Monte Carlo numerical calculation model through the number of detectors, the physical structure parameters, and the calibration well parameters as the numerical calculation model, wherein the pulsed neutron gamma instrument includes a D-T source and at least two detectors;

[0012] The physical structure parameters of the pulsed neutron gamma instrument include: the size of the pulsed neutron gamma instrument, the material of the pulsed neutron gamma instrument, the density of the pulsed neutron gamma instrument, the radiation source, the type of the radiation source, the type of the detector crystal, the size of the detector crystal, the type of the shielding body, and the size of the shielding body.

[0013] According to an embodiment of the present invention, the measured response data is obtained through the following steps:

[0014] Place the pulsed neutron gamma instrument in calibration wells with different formation porosities, densities, and borehole sizes to obtain the responses of different detectors, and obtain the measured response data, wherein the measured response data includes inelastic gamma counts, capture gamma counts, total gamma counts, and gamma time spectra of different detectors.

[0015] According to an embodiment of the present invention, the optimized numerical calculation model is obtained through the following steps:

[0016] Calculate the relative error between the measured response data and the first simulated response data;

[0017] When the relative error is less than the error threshold, determine the numerical calculation model as the optimized numerical calculation model;

[0018] When the relative error is not less than the error threshold, adjust the structure of the numerical calculation model until the relative error is less than the error threshold, and determine the numerical calculation model as the optimized numerical calculation model.

[0019] According to an embodiment of the present invention, the cave conditions beside the well include: fracture-cavity size, fracture-cavity fluid type, matrix porosity, lithology, and density; the second simulated response data includes inelastic gamma counts, capture gamma counts, total gamma counts, and gamma time spectra of different detectors.

[0020] According to an embodiment of the present invention, the pulsed neutron gamma logging fracture-cavity identification model is obtained through the following steps:

[0021] Construct a training set of the preset neural network model, and the training set is the value obtained through the second simulated response data;

[0022] Initialize the weights and biases of the preset neural network model;

[0023] Input the training set into the preset neural network model for training, calculate the loss function and accuracy of the preset neural network model. When the loss function and the accuracy meet the preset conditions, output the pulsed neutron gamma logging fracture and cave identification model;

[0024] When the loss function and the accuracy do not meet the preset conditions, calculate the updated weights and biases based on the backpropagation mechanism for the loss function, and calculate the updated loss function and accuracy until the loss function and the accuracy meet the preset conditions, and output the pulsed neutron gamma logging fracture and cave identification model.

[0025] According to an embodiment of the present invention, the structure of the pulsed neutron gamma logging fracture and cave identification model is 5 layers, with 3 hidden layers, and the neurons in the hidden layers are 4, 8, and 8.

[0026] According to an embodiment of the present invention, the loss function adopts L2 regularization, and the expression of the loss function is as follows:

[0027]

[0028] where g is the activation function, w and b are the weights and biases between the connection layers; p is the number of iterations; m is the number of samples; x is the net input vector of neurons in each layer, is the output vector of the training set, and χ is the regularization coefficient;

[0029] Among them, the training algorithm of the pulsed neutron gamma logging fracture and cave identification model is the Adam algorithm; the activation function adopts the ReLU function.

[0030] According to another aspect of the present invention, there is also provided a storage medium, which contains a series of instructions for executing the method steps described in any one of the above.

[0031] According to another aspect of the present invention, there is also provided a pulsed neutron gamma logging fracture and cave identification device, which executes the method described in any one of the above. The device includes:

[0032] A building module, configured to establish a numerical calculation model based on the physical structure of the pulsed neutron gamma instrument and the calibration well parameters to simulate and obtain the first simulated response data of different detectors;

[0033] An optimization module, configured to optimize the accuracy of the numerical calculation model by combining the measured response data of the pulsed neutron gamma instrument with the first simulated response data to obtain an optimized numerical calculation model;

[0034] A simulation module, configured to simulate and obtain second simulation response data under various cave conditions beside wells by using the optimized numerical calculation model;

[0035] A training module, configured to train a preset neural network model based on the second simulation response data to obtain a pulsed neutron gamma logging fracture and cave identification model.

[0036] The present invention provides a method, device and medium for pulsed neutron gamma logging fracture and cave identification. Compared with the prior art, the following advantages are achieved:

[0037] By using the measured response data to optimize the numerical calculation model, the accuracy of the optimized numerical calculation model is improved. The second simulation response data is obtained by using the optimized numerical calculation model, and the neural network model is trained by the second simulation response data, so as to obtain an accurate pulsed neutron gamma logging fracture and cave identification model, realizing the calculation of fracture and cave sizes and the identification of fluid types under different formation porosities, densities and wellbore sizes. The difficulty of calculating fracture and cave sizes and identifying fluid types by using gamma information of multiple detectors is reduced, and the accuracy of calculation is ensured. At the same time, in actual operation, no petroleum professionals are required for processing, and only the measured data is needed to obtain the fracture and cave sizes and fluid types.

[0038] Other features and advantages of the present invention will be described in the following specification, and will be partially obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention can be realized and obtained by the structures specifically pointed out in the specification, claims and drawings. Description of the Drawings

[0039] The drawings are used to provide further understanding of the present invention, and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention, but do not constitute a limitation to the present invention. In the drawings:

[0040] Figure 1 Shows a flowchart of a method for pulsed neutron gamma logging fracture and cave identification according to an embodiment of the present invention;

[0041] Figure 2 Shows a schematic diagram of a multi-dimensional Monte Carlo numerical calculation model according to an embodiment of the present invention;

[0042] Figure 3 Shows a comparison diagram of the fracture and cave sizes predicted by the pulsed neutron gamma logging fracture and cave identification model and the fracture and cave sizes of the optimized numerical model according to an embodiment of the present invention;

[0043] Figure 4 Shows a schematic structural diagram of a pulsed neutron gamma logging fracture and cave identification device according to an embodiment of the present invention.

[0044] In the drawings, like parts are designated by like reference numerals. Additionally, the drawings are not drawn to scale.

[0045] The meanings of the various reference numerals in the drawings are as follows: 1 is a D-T source that emits neutrons with an energy of 14 MeV, 2 is a tungsten-nickel-iron shield, 3 is a near gamma detector, 4 is a medium gamma detector, 5 is a far gamma detector, 6 is a wellbore, 7 is a fracture-vug, and 8 is a basement formation. Detailed implementation manners

[0046] To make the objectives, technical solutions, and advantages of the present invention clearer, the following further describes the embodiments of the present invention in detail with reference to the accompanying drawings.

[0047] In the prior art (CN114117898A), a forward modeling method for logging-while-drilling gamma-ray logging based on a machine learning algorithm is mentioned. It describes establishing formation models with different radioactive intensities, layer thicknesses, densities, and different dips; simulating the transport process of gamma rays in the formation to obtain the response relationship of the logging-while-drilling gamma detector count under different geological conditions; establishing a database for the response relationship, training the simulated data with a neural network algorithm, and constructing a neural network model suitable for logging-while-drilling gamma forward modeling to form a fast forward modeling method for logging-while-drilling gamma-ray logging. However, the prior art (CN114117898A) does not mention the calculation of fracture-vug size and the identification of fluid types.

[0048] In the prior art (CN115457348A), a method of using machine learning to preprocess and train rock pictures and label pictures is mentioned. Using the DeepLabv3+ model, automatic segmentation of image pores is achieved by using the pore-fracture contours and pixel point classification in the pictures. However, the prior art (CN115457348A) belongs to the category of rock physics image recognition and processing and has no relevant connection with this application. And the prior art does not mention the related technology of using pulsed neutron gamma technology for fracture-vug identification.

[0049] In view of the above-mentioned defects of the prior art, the present invention provides a method, device, and storage medium for identifying fracture-vugs by pulsed neutron gamma logging.

[0050] Figure 1 The flowchart of a method for identifying fracture-vugs by pulsed neutron gamma logging according to an embodiment of the present invention is shown. The method includes:

[0051] S101, establishing a numerical calculation model based on the physical structure of the pulsed neutron gamma instrument and the calibration well parameters to simulate and obtain the first simulated response data of different detectors.

[0052] Among them, the calibrated well parameters can be fracture-cavity size, fracture-cavity fluid type, matrix porosity, lithology, and density. The first simulated response data can include: inelastic gamma counts, capture gamma counts, total gamma counts, and gamma time spectra of different detectors, and the gamma time spectra are used to determine the formation neutron capture cross-section.

[0053] Exemplarily, a numerical calculation model can be established based on the physical structure of the pulsed neutron gamma instrument, the calibrated well parameters, and in combination with the Monte Carlo method to simulate and obtain the first simulated response data of different detectors.

[0054] S102. Optimize the accuracy of the numerical calculation model by combining the measured response data of the pulsed neutron gamma instrument with the first simulated response data to obtain an optimized numerical calculation model.

[0055] Exemplarily, the measured response data can be obtained by using the pulsed neutron gamma instrument, and the relative error between the first simulated response data and the measured response data can be calculated to optimize the accuracy of the data calculation model to obtain an optimized numerical calculation model.

[0056] S103. Use the optimized numerical calculation model to simulate and obtain the second simulated response data under various near-wellbore cave conditions.

[0057] Exemplarily, the near-wellbore cave conditions can include: fracture-cavity size, fracture-cavity fluid type, matrix porosity, lithology, and density. The second simulated response data can include: inelastic gamma counts, capture gamma counts, total gamma counts, and gamma time spectra of different detectors. The second simulated response data under various near-wellbore cave conditions can be simulated by using the optimized numerical calculation model to construct a database, and the database parameters include six parameters such as cave radius, fluid type, matrix lithology, oil-gas saturation, matrix porosity, and wellbore diameter.

[0058] As shown in Table 1, it is the parameter range of the database. In the table, Ri is the cave radius, Fluid Type is the fluid type, and there are 3 types of fluid types, including fresh water, oil, and natural gas, with densities of 1.0 g / cm 3 、0.87 g / cm 3 、0.2 g / cm 3 , in this application, only 2 kinds of fluid mixtures are considered, namely water / oil, water / gas, and oil / gas. Satur is the saturation. Lith is the matrix lithology, including sandstone, limestone, and dolomite, and the matrix porosity The range is 0 - 40%, where CAL is the borehole diameter. According to the data in Table 1, this database has 30×3×10×3×8×5, that is, a total of 108,000 response vectors. Taking the three-detector case as an example, each detector records inelastic scattering gamma counts, capture gamma counts, total gamma counts, and gamma time spectra. Then the number of data points in the database is 108,000×12, that is, a total of 1,296,000.

[0059] Table 1 Parameter ranges of the database

[0060]

[0061] S104, based on the second simulated response data, train the preset neural network model to obtain a pulsed neutron gamma logging fracture-vug identification model.

[0062] Exemplarily, the second simulated response data can be input into the preset neural network model for training, and then the trained neural network model can be output as a pulsed neutron gamma logging fracture-vug identification model. Inputting the second simulated response data into the pulsed neutron gamma logging fracture-vug identification model, the output results can be obtained: fracture-vug size, fracture-vug fluid type, matrix porosity, lithology, and density.

[0063] The numerical calculation model is optimized through the measured response data, improving the accuracy of the optimized numerical calculation model. The second simulated response data is obtained using the optimized numerical calculation model, and the neural network model is trained through the second simulated response data, thereby obtaining an accurate pulsed neutron gamma logging fracture-vug identification model, realizing the calculation of fracture-vug size and the identification of fluid type under different formation porosities, densities, and borehole sizes. It reduces the difficulty of calculating the fracture-vug size and identifying the fluid type using the gamma information of multiple detectors, and ensures the accuracy of the calculation. At the same time, in actual operation, no petroleum professionals are required for processing, and only the measured data is needed to obtain the fracture-vug size and fluid type.

[0064] In a possible embodiment, a numerical calculation model is established through the following steps:

[0065] Determine the number of detectors, physical structure parameters, and calibration well parameters included in the pulsed neutron gamma instrument, and establish a multi-dimensional Monte Carlo numerical calculation model through the number of detectors, physical structure parameters, and calibration well parameters as the numerical calculation model. Among them, the pulsed neutron gamma instrument includes a D-T source and at least two detectors.

[0066] The physical structure parameters of the pulsed neutron gamma instrument include: the size of the pulsed neutron gamma instrument, the material of the pulsed neutron gamma instrument, the density of the pulsed neutron gamma instrument, the radiation source, the type of the radiation source, the type of the detector crystal, the size of the detector crystal, the type of the shielding body, and the size of the shielding body.

[0067] As shown Figure 2 in the figure, it is a schematic diagram of a multi-dimensional Monte Carlo numerical calculation model. In the figure, 1 is a D-T source, emitting neutrons with an energy of 14 MeV, 2 is a tungsten-nickel-iron shield, 3 is a near gamma detector, 4 is a medium gamma detector, 5 is a far gamma detector, 6 is a wellbore, 7 is a fracture-vug, and 8 is a bedrock formation. For example, according to the number of detectors, it can be determined that the multi-dimensional Monte Carlo numerical calculation model can be a three-dimensional Monte Carlo numerical calculation model. Then, the physical structure parameters and calibration well parameters of the pulsed neutron gamma instrument can be determined by measurement. Next, the physical structure parameters and calibration well parameters of the pulsed neutron gamma instrument are input into an existing model established based on the Monte Carlo method for calculating in a three-dimensional complex geometric structure to establish a multi-dimensional Monte Carlo numerical calculation model as the numerical calculation model.

[0068] In a possible embodiment, the measured response data is obtained through the following steps:

[0069] Place the pulsed neutron gamma instrument in calibration wells with different formation porosities, densities, and wellbore sizes to obtain the responses of different detectors, and obtain the measured response data. Among them, the measured response data includes inelastic gamma counts, capture gamma counts, total gamma counts, and gamma time spectra of different detectors. In this way, the measured response data measured is more accurate.

[0070] In a possible embodiment, the optimized numerical calculation model is obtained through the following steps:

[0071] Calculate the relative error between the measured response data and the first simulated response data;

[0072] When the relative error is less than the error threshold, determine the numerical calculation model as the optimized numerical calculation model;

[0073] When the relative error is not less than the error threshold, adjust the structure of the numerical calculation model so that when the relative error is less than the error threshold, determine the numerical calculation model as the optimized numerical calculation model.

[0074] For example, the relative error between the measured response data and the first simulated response data can be calculated by the following formula.

[0075]

[0076] Among them, R M represents the measured response data, R S represents the first simulated response data, and Δ R represents the relative error.

[0077] When the relative error Δ RWhen it is less than the error threshold ε, the numerical calculation model is determined as the optimized numerical calculation model. At this time, the optimized numerical calculation model can completely replace the pulsed neutron gamma instrument for generating the second simulated response data and the subsequent establishment of the database. Exemplarily, the value of ε can be determined according to the actual application scenario. In this application, the value is 2%. When the relative error Δ R is not less than ε, the parameters of the numerical calculation model are adjusted, and the relative error is recalculated until the numerical calculation model when the relative error is less than the error threshold is determined as the optimized numerical calculation model. In this way, the determined optimized numerical model is more accurate.

[0078] In a possible embodiment, the pulsed neutron gamma logging fracture and cave identification model is obtained through the following steps:

[0079] Construct a training set for the preset neural network model, and the training set is the value obtained from the second simulated response data;

[0080] Initialize the weights and biases of the preset neural network model, input the training set into the preset neural network model for training, calculate the loss function and accuracy of the preset neural network model. When the loss function and accuracy meet the preset conditions, output the pulsed neutron gamma logging fracture and cave identification model;

[0081] When the loss function and accuracy do not meet the preset conditions, calculate the updated weights and biases of the loss function based on the backpropagation mechanism, and calculate the updated loss function and accuracy until the loss function and accuracy meet the preset conditions, and output the pulsed neutron gamma logging fracture and cave identification model.

[0082] Exemplarily, the second simulated response data can be normalized first, and then the normalized second simulated response data is divided into a test set and a training set through a random function. Among them, the training set accounts for 95% of the total database samples, and the test set accounts for 5% of the total database samples.

[0083] Secondly, the weights w, biases b of the preset neural network model can be initialized, and the initial number p of the iteration times is set to 1. Then, input the training set into the preset neural network model, calculate the loss function and accuracy of the preset neural network model. When the loss function and accuracy meet the preset conditions, that is, the loss function is less than the target threshold and the accuracy is less than the accuracy threshold, stop training, and output the preset neural network model as the pulsed neutron gamma logging fracture and cave identification model. When the loss function and accuracy do not meet the preset conditions, calculate the updated weights and biases of the loss function based on the backpropagation mechanism, input the training set into the preset neural network model, calculate the updated loss function and accuracy, and further determine whether the loss function and accuracy meet the preset conditions, and so on in a loop until the loss function and accuracy meet the preset conditions, and output the pulsed neutron gamma logging fracture and cave identification model.

[0084] Such asFigure 3 As shown in the figure, it is a comparison chart of the fracture-vug size predicted by the fracture-vug identification model of pulsed neutron gamma logging and the fracture-vug size (actual fracture-vug size) of the optimized numerical model. It can be seen from the figure that the prediction result of the fracture-vug identification model of pulsed neutron gamma logging is in good agreement with the actual fracture-vug size. In this way, the determined fracture-vug identification model of pulsed neutron gamma logging is more accurate.

[0085] In a possible embodiment, the structure of the fracture-vug identification model of pulsed neutron gamma logging is 5 layers, with 3 hidden layers, and the neurons in the hidden layers are 4, 8, and 8. Exemplarily, the number of neurons can be determined by algorithms such as K-means and ROLS. In this way, the training accuracy can be improved and the number of training times can be reduced, thereby improving the accuracy of the fracture-vug identification model of pulsed neutron gamma logging.

[0086] In a possible embodiment, the loss function adopts L2 regularization, and the expression of the loss function is as follows:

[0087]

[0088] where g is the activation function, w and b are the weights and biases between the connection layers; p is the number of iterations; m is the number of samples; x is the net input vector of neurons in each layer, is the output vector of the training set, and χ is the regularization coefficient;

[0089] Among them, the training algorithm of the fracture-vug identification model of pulsed neutron gamma logging is the Adam algorithm; the activation function adopts the ReLU function.

[0090] Among them, the training algorithm of the fracture-vug identification model of pulsed neutron gamma logging is the Adam algorithm, the batch size is 0.75% of the training set, and the learning rate variation range is 10 -4 ~10 -3 . The input layer is the normalized vector x new , and y is the output vector y of the training set new . During the training process, there are multiple solutions in the training results, that is, there may be multiple fracture-vug sizes. Therefore, the detection depths of different detectors are added to minimize the multiple solutions as much as possible. In this way, the accuracy and speed of the training results are improved.

[0091] Based on the same inventive concept, in this embodiment, a fracture-vug identification device for pulsed neutron gamma logging is provided. Figure 4 It is a block diagram of a fracture-vug identification device for pulsed neutron gamma logging provided by an embodiment of the present application. As Figure 4 shown, the device may include:

[0092] A building module 510, configured to establish a numerical calculation model according to the physical structure of the pulsed neutron gamma instrument and the calibration well parameters, so as to simulate and obtain the first simulated response data of different detectors;

[0093] An optimization module 520 is configured to optimize the accuracy of a numerical calculation model by combining the measured response data of a pulsed neutron gamma instrument with the first simulated response data, so as to obtain an optimized numerical calculation model;

[0094] A simulation module 530 is configured to use the optimized numerical calculation model to simulate and obtain second simulated response data under various wellbore cave conditions;

[0095] A training module 540 is configured to train a preset neural network model based on the second simulated response data to obtain a pulsed neutron gamma logging fracture and cave identification model.

[0096] A pulsed neutron gamma logging fracture and cave identification method provided by the present invention can also cooperate with a computer-readable storage medium. A computer program is stored on the storage medium, and the computer program is executed to run a pulsed neutron gamma logging fracture and cave identification method. The computer program can run computer instructions, and the computer instructions include computer program codes, and the computer program codes can be in the form of source codes, object codes, executable files or some intermediate forms, etc.

[0097] The computer-readable storage medium may include: any entity or device capable of carrying computer program codes, recording media, USB flash drives, mobile hard disks, magnetic disks, optical disks, computer memories, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0098] It should be noted that the content included in the computer-readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice within the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable storage medium does not include electrical carrier signals and telecommunication signals.

[0099] In summary, the present invention provides a pulsed neutron gamma logging fracture and cave identification method, device and medium. Compared with the prior art, the following advantages are achieved:

[0100] This method optimizes the numerical calculation model through measured response data, improving the accuracy of the optimized numerical calculation model. The optimized numerical calculation model is used to obtain the second simulated response data, and the neural network model is trained through the second simulated response data, thereby obtaining an accurate fracture and cave identification model for pulsed neutron gamma logging, realizing the calculation of fracture and cave sizes and the identification of fluid types under different formation porosities, densities, and borehole sizes. It reduces the difficulty of calculating fracture and cave sizes and identifying fluid types using gamma information from multiple detectors, and ensures the accuracy of the calculation. At the same time, in actual operation, no petroleum professionals are required for processing, and only the measured data is needed to obtain the fracture and cave sizes and fluid types.

[0101] It should be understood that the embodiments disclosed in the present invention are not limited to the specific structures, processing steps, or materials disclosed herein, but should extend to equivalent alternatives of these features understood by those of ordinary skill in the relevant art. It should also be understood that the terms used herein are only for the purpose of describing specific embodiments and do not mean to limit.

[0102] In the description of the present invention, unless otherwise specified, the meaning of "a plurality" is two or more; the orientation or positional relationships indicated by the terms "upper", "lower", "left", "right", "inner", "outer", "front end", "rear end", "head", "tail", etc. are based on the orientation or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation of the present invention. In addition, the terms "first", "second", "third", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0103] In the description of the present invention, it should be noted that unless otherwise clearly specified and defined, the terms "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0104] Certain terms are used throughout this application document to refer to specific system components. As those skilled in the art will recognize, the same components can typically be referred to by different names, and thus this application document is not intended to distinguish between components that differ only in name and not in function. In this application document, the terms "comprise", "include", and "have" are used in an open-ended fashion and should therefore be interpreted to mean "including but not limited to...". Additionally, the terms "substantially", "essentially", or "approximately" as may be used herein refer to the tolerances accepted by the industry for the corresponding terms. The term "coupled" as may be employed herein includes direct coupling and indirect coupling via additional components, elements, circuits, or modules, where for indirect coupling, the intervening components, elements, circuits, or modules do not change the information of the signal but may adjust its current level, voltage level, and / or power level. Inferred coupling (e.g., where one element is coupled to another element by inference) includes direct and indirect coupling between the two elements in the same manner as "coupled".

[0105] As used in the specification, the phrase "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the invention. Thus, the appearances of the phrase "one embodiment" or "an embodiment" throughout the specification are not necessarily all referring to the same embodiment.

[0106] Embodiments of the present invention are given for purposes of illustration and description and are not exhaustive or limit the invention to the disclosed form. Many modifications and variations are obvious to those of ordinary skill in the art. The embodiments are chosen and described in order to better explain the principles of the invention and its practical application, and to enable those of ordinary skill in the art to understand the invention and design various embodiments with various modifications suitable for a particular purpose.

[0107] Although the embodiments disclosed in the present invention are as above, the content described is only an embodiment adopted for the convenience of understanding the present invention and is not intended to limit the present invention. Any person skilled in the art within the technical field to which the present invention pertains, without departing from the spirit and scope disclosed by the present invention, may make any modifications and changes in the form of implementation and details, but the scope of patent protection of the present invention shall still be subject to the scope defined by the appended claims.

Claims

1. A pulsed neutron gamma logging fracture and cavity identification method, characterized in that: The method comprises: A numerical calculation model is established according to the physical structure of the pulsed neutron gamma instrument and the calibration well parameters to simulate and obtain the first simulation response data of different detectors; By combining the measured response data of the pulsed neutron gamma instrument with the first simulated response data, the accuracy of the numerical calculation model is optimized to obtain an optimized numerical calculation model; Using the optimized numerical calculation model to simulate and obtain second simulation response data under various well-near cave conditions; Based on the second simulation response data, the preset neural network model is trained to obtain a pulsed neutron gamma logging fracture and cavity identification model.

2. The method according to claim 1, characterized in that The numerical calculation model is established by the following steps: Determine the number of detectors, entity structure parameters and calibration well parameters included in the pulsed neutron gamma instrument, and establish a multidimensional Monte Carlo numerical calculation model as the numerical calculation model through the number of detectors, the entity structure parameters and the calibration well parameters, wherein the pulsed neutron gamma instrument includes a DT source and at least two detectors; The physical structural parameters of the pulsed neutron gamma instrument include: the size of the pulsed neutron gamma instrument, the material of the pulsed neutron gamma instrument, the density of the pulsed neutron gamma instrument, the radiation source, the type of the radiation source, the type of the detector crystal, the size of the detector crystal, the type of the shielding body, and the size of the shielding body.

3. The method according to claim 1 or 2, characterized in that The measured response data is obtained by the following steps: The pulsed neutron gamma instrument is placed in a plurality of calibrated wells with different formation porosity, density and borehole size to obtain responses of different detectors, and obtain the measured response data, wherein the measured response data includes inelastic gamma counts, captured gamma counts, total gamma counts and gamma time spectra of different detectors.

4. The method according to any one of claims 1 to 3, characterized in that The optimized numerical calculation model is obtained by the following steps: Calculating a relative error between the measured response data and the first simulated response data; When the relative error is less than an error threshold, determining the numerical calculation model as the optimized numerical calculation model; When the relative error is not less than the error threshold, the structure of the numerical calculation model is adjusted until the relative error is less than the error threshold, and the numerical calculation model is determined as the optimized numerical calculation model.

5. The method according to any one of claims 1 to 4, characterized in that The wellbore cave conditions include: fracture cave size, fracture cave fluid type, bedrock porosity, lithology and density; the second simulation response data includes inelastic gamma counts, captured gamma counts, total gamma counts and gamma time spectra of different detectors.

6. The method according to any one of claims 1 to 5, characterized in that The pulsed neutron gamma logging fracture-cavity identification model is obtained by the following steps: Constructing a training set of the preset neural network model, wherein the training set is a value obtained through the second simulated response data; Initializing the weights and biases of the preset neural network model; Inputting the training set into the preset neural network model for training, calculating the loss function and precision of the preset neural network model, and outputting the pulsed neutron gamma logging fracture-cavity identification model when the loss function and the precision meet preset conditions; When the loss function and the accuracy do not meet the preset conditions, the loss function is used to calculate the updated weights and biases based on the back propagation mechanism, and the updated loss function and accuracy are calculated until the loss function and the accuracy meet the preset conditions, and the pulse neutron gamma logging fracture identification model is output.

7. The method according to any one of claims 1 to 6, characterized in that The structure of the pulsed neutron gamma logging fracture and cave identification model is 5 layers, the hidden layer is 3 layers, and the neurons of the hidden layer are 4, 8, and 8.

8. The method according to claim 6 or 7, characterized in that The loss function adopts L2 regularization, and the expression of the loss function is as follows: Among them, g is the activation function, w, b are the weights and biases between the connection layers; p is the number of iterations; m is the number of samples; x is the net input vector of each neuron layer, y i j is the output vector of the training set, χ is the regularization coefficient; Among them, the training algorithm of the pulse neutron gamma logging fracture and cave identification model is the Adam algorithm; the activation function adopts the ReLU function.

9. A storage medium, characterized in that: It contains a series of instructions for executing the method steps as claimed in any one of claims 1 to 8.

10. A pulsed neutron gamma logging fracture and cavity identification device, characterized in that: Execute the method according to any one of claims 1 to 8, wherein the device comprises: Establishing a module, used to establish a numerical calculation model according to the physical structure of the pulsed neutron gamma instrument and the calibration well parameters, so as to simulate and obtain the first simulation response data of different detectors; An optimization module, configured to optimize the accuracy of the numerical calculation model by combining the measured response data of the pulsed neutron gamma instrument with the first simulated response data to obtain an optimized numerical calculation model; A simulation module, used to use the optimized numerical calculation model to simulate and obtain second simulation response data under various well-near cave conditions; The training module is used to train the preset neural network model based on the second simulation response data to obtain a pulsed neutron gamma logging fracture and cavity identification model.

Images