Permeability Calculation Method and Device Based on Rock Physics Constrained Neural Network

Abstract

An embodiment of the present invention discloses a method and device for calculating permeability based on a rock physics constrained neural network. The method includes: extracting corresponding depth point logging parameters based on core samples, and preprocessing the logging parameters and corresponding core data to construct a first sample set and a second sample set; constructing and training a first neural network according to the first sample set; constructing and training a second constrained neural network according to the second sample set; wherein, the second constrained neural network consists of an input layer, a constraint layer, a hidden layer and an output layer, and the constraint layer is a network layer where constraint elements are specified to be added; the constraint elements are calculated according to the output result of the first neural network, input to the constraint layer of the second constrained neural network and used to constrain the calculation of the second constrained neural network; obtaining the logging parameters of the well to be predicted, and inputting them into the first neural network and the second constrained neural network respectively to obtain the permeability of the well to be predicted. The present invention improves the prediction accuracy of the permeability of the entire well section.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of data processing in oil exploration, and in particular, to a method and device for calculating permeability based on a petrophysical-constrained neural network. Background Art

[0002] Permeability is a measure of the ability of a rock to allow fluid to pass through under a pressure difference, and is a very important parameter in oil and gas field exploration and development. Establishing a mathematical model of permeability with other petrophysical parameters and calculating permeability using the mathematical model is currently the only available full-well-section permeability evaluation method. The mathematical models of permeability can be roughly divided into two types. One is the classical porosity-permeability formula, such as the Timur formula, which has low calculation accuracy due to reasons such as regional applicability. The other is the permeability model established based on cores, such as the FZI. This method is greatly affected by the distribution range of cores. The premise for establishing a permeability model based on cores is that the lithology of the cores taken is representative in the full well section, that is, the lithology of the cores and the lithology of the full well section follow the same distribution. However, due to the high difficulty and cost of downhole coring operations, coring operations are generally only carried out in the target layer. Therefore, it is difficult for the core lithology distribution to be consistent with the full well section distribution. At the same time, affected by rock heterogeneity, the permeability model established based on cores will show phenomena such as low calculation accuracy and poor generalization ability.

[0003] At the current technical level, the main ways to obtain rock permeability include direct measurement through experiments or engineering means and indirect calculation based on geophysical methods. The direct measurement methods include core experiment measurement method, wireline formation testing method, drillstem formation testing method, etc., and the indirect calculation methods include well logging interpretation method and seismic interpretation method, etc. Due to factors such as the high cost of formation testing method and core experiment measurement method and the limited depth range of measurement, this method is only limited to the calibration of well logging permeability calculation. Currently, the most commonly used method is still to establish a formula or model using conventional well logging data and calculate using the method of calibrating permeability through experimental analysis. Due to the empirical nature of the formula or model, the influence of human factors, regional differences, and the complex and cumbersome model establishment steps, the final calculation model often cannot be widely applied or the calculation accuracy is insufficient. For some methods using machine learning algorithms for reservoir permeability prediction, a large number of core permeability samples are required, and because the machine learning algorithm is simply and rigidly applied for model training and prediction without considering the influence of core distribution, the application effect of this method in permeability calculation is not good and the calculation accuracy is generally low. Summary of the Invention

[0004] In view of the above problems, the embodiments of the present invention are proposed to provide a method and device for calculating permeability that overcome the above problems or at least partially solve the above problems.

[0005] According to one aspect of the embodiments of the present invention, a method for calculating permeability based on a petrophysical constrained neural network is provided. The method includes:

[0006] Extracting corresponding depth-point logging parameters based on core samples, and preprocessing the logging parameters and corresponding core data to construct a first sample set and a second sample set;

[0007] Constructing and training a first neural network according to the first sample set;

[0008] Constructing and training a second constrained neural network according to the second sample set; wherein, the second constrained neural network is composed of an input layer, a constraint layer, a hidden layer and an output layer, and the constraint layer is a network layer where constraint elements are specified to be added; the constraint elements are calculated based on the output result of the first neural network, input to the constraint layer of the second constrained neural network and used to constrain the calculation of the second constrained neural network;

[0009] Obtaining the logging parameters of the well to be predicted, and inputting them into the first neural network and the second constrained neural network respectively to obtain the permeability of the well to be predicted.

[0010] According to another aspect of the embodiments of the present invention, a device for calculating permeability based on a petrophysical constrained neural network is provided. The device includes:

[0011] A sample construction module, adapted to extract corresponding depth-point logging parameters based on core samples, and preprocess the logging parameters and corresponding core data to construct a first sample set and a second sample set;

[0012] A first training module, adapted to construct and train a first neural network according to the first sample set;

[0013] A second training module, adapted to construct and train a second constrained neural network according to the second sample set; wherein, the second constrained neural network is composed of an input layer, a constraint layer, a hidden layer and an output layer, and the constraint layer is a network layer where constraint elements are specified to be added; the constraint elements are calculated based on the output result of the first neural network, input to the constraint layer of the second constrained neural network and used to constrain the calculation of the second constrained neural network;

[0014] A prediction module, adapted to obtain the logging parameters of the well to be predicted, and input them into the first neural network and the second constrained neural network respectively to obtain the permeability of the well to be predicted.

[0015] According to still another aspect of the embodiments of the present invention, a computing device is provided, including: a processor, a memory, a communication interface and a communication bus. The processor, the memory and the communication interface complete communication with each other through the communication bus;

[0016] The memory is used to store at least one executable instruction, and the executable instruction causes the processor to perform operations corresponding to the above-mentioned permeability calculation method based on the rock physics constrained neural network.

[0017] According to another aspect of the embodiments of the present invention, a computer storage medium is provided. The storage medium stores at least one executable instruction, and the executable instruction causes the processor to perform operations corresponding to the above-mentioned permeability calculation method based on the rock physics constrained neural network.

[0018] According to the permeability calculation method and device based on the rock physics constrained neural network provided by the embodiments of the present invention, by adding constraint elements to the constraint layer of the second constrained neural network, the calculation of the second constrained neural network is constrained, the influence of the core distribution is reduced, and a neural network that more conforms to the actual situation and has better generalization ability is constructed, so that the permeability of the full well section logging can be predicted quickly and accurately.

[0019] The above description is only an overview of the technical solutions of the embodiments of the present invention. In order to be able to understand the technical means of the embodiments of the present invention more clearly, it can be implemented according to the content of the description. And in order to make the above and other purposes, features and advantages of the embodiments of the present invention more obvious and understandable, the following specifically illustrates the specific embodiments of the present invention. Description of the Drawings

[0020] By reading the following detailed description of the preferred embodiments, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the embodiments of the present invention. And throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0021] Figure 1 Shows a flowchart of a permeability calculation method based on a rock physics constrained neural network according to an embodiment of the present invention;

[0022] Figure 2 Shows a schematic structural diagram of a permeability calculation device based on a rock physics constrained neural network according to an embodiment of the present invention;

[0023] Figure 3 Shows a schematic structural diagram of a computing device according to an embodiment of the present invention. Detailed Embodiments

[0024] Exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present invention can be more thoroughly understood and the scope of the present invention can be fully conveyed to those skilled in the art.

[0025] Figure 1 The flowchart of a permeability calculation method based on a petrophysical constraint neural network according to an embodiment of the present invention is shown. As Figure 1 shown, the method includes the following steps:

[0026] Step S101, extracting corresponding depth point logging parameters based on core samples, and preprocessing the logging parameters and corresponding core data to construct a first sample set and a second sample set.

[0027] In this embodiment, cores are used as samples, and logging parameters such as gamma, neutron, and density, which have a very high correlation with permeability and irreducible water saturation, are extracted therefrom. Using the input data of the logging parameter sample set, a first sample set and a second sample set are constructed respectively. The first sample set is composed of logging parameters and the matched core irreducible water saturation, and the second sample set is composed of logging parameters and the matched core permeability.

[0028] Furthermore, to ensure the accuracy during subsequent training, the logging parameters and corresponding core data are preprocessed. Specifically, a crossplot is drawn based on the porosity in the logging parameters and the core permeability, and the value range of the logging parameters is determined according to the crossplot to eliminate the logging parameters corresponding to outliers. These outliers are outside the value range and affect the accuracy during training. Or, if there are no corresponding logging parameters in the core data, the corresponding logging parameters are interpolated using the interpolation method according to the adjacent upper and lower depths of the core data to complete the missing values required for the logging parameters. Considering that the logging parameters of different logs are obtained by different acquisition methods, different well conditions, different instruments used for acquisition, etc., resulting in different logging parameter standards, the logging parameters are normalized. The normalization is performed according to the maximum and minimum values of the logging parameters, so that their response differences are in the same order of magnitude, thereby unifying the measurement units of the logging parameters and enabling better discovery of their regularity during training. During normalization, if the value of a certain logging parameter is m, the maximum value is x, and the minimum value is y, the normalization can be performed in the way of (m - y) / (x - y). The normalization can be set according to the actual situation, such as normalizing the logging parameters to the 0 - 1 interval, etc. The specific calculation method of the normalization is not limited here. When preprocessing the logging parameters, one or more of the above can be selected according to the specific implementation situation for data preprocessing, making the logging data more regular and facilitating accurate training of the neural network.

[0029] Step S102: Construct and train a first neural network based on the first sample set.

[0030] The input layer of the first neural network is the logging parameters of the first sample set, and the output layer is the irreducible water saturation of the first sample set. The first neural network has a first specified number of layers. For example, the first neural network has a three-layer neural network structure. The number of neurons in the first layer is twice the number of neurons in the input layer, the number of neurons in the second layer is equal to the number of neurons in the input layer, and finally the output layer outputs the irreducible water saturation. When training the first neural network using the first sample set, the first sample set can be divided by the K-Fold cross-validation method into a test sample set and a validation sample set for training, and the training parameters of the first neural network can be adjusted to complete the training of the first neural network.

[0031] Step S103: Construct and train a second constrained neural network based on the second sample set.

[0032] The input layer of the second constrained neural network is the logging parameters of the second sample set, and the output layer is the core permeability of the second sample set. Considering that it is difficult for the core lithology distribution to be consistent with the distribution of the entire well section, the permeability model established only based on the core will have problems such as low calculation accuracy and poor generalization ability. In this embodiment, when constructing the second constrained neural network, the second constrained neural network is based on a neural network structure and consists of an input layer, a constraint layer, a hidden layer, and an output layer. Among them, constraint elements are added in the constraint layer. The constraint elements are connected to the output layer of the first neural network and are not directly connected to the neurons of the input layer of the second constrained neural network. The neurons in the constraint layer other than the constraint elements are fully connected to the neurons of the input layer of the second constrained neural network, that is, the other neurons except the constraint elements in the constraint layer are determined by the input layer of the second constrained neural network. The input of the constraint elements of the second constrained neural network is provided by the output of the first neural network and is calculated from the output result of the first neural network and the porosity of the logging parameters. The output of the constraint elements is fully connected to the hidden layer, and the weights and biases of the constraint elements are adjusted together with other neurons. The input of the constraint elements to this constraint layer is a rock physics model constraint layer. Specifically, the constraint elements of the second constrained neural network are constrained by the rock physics model and are calculated according to the rock physics model. The rock physics model can be obtained according to the following formula:

[0033]

[0034] where K is a constraint element parameter, which is used to characterize the permeability calculated using the formula, C is a specified coefficient, which can be set as a constant term, and its value is not limited here. φ is the porosity of the logging parameters input to the second constrained neural network, S wirris the irreducible water saturation output by the first neural network. Here, the well logging parameters used by the first neural network when calculating the irreducible water saturation are the same well logging parameters as those input to the second constrained neural network. By using this formula to constrain the second constrained neural network, the second constrained neural network can better adapt to the well logging parameters of the entire well section and improve the prediction accuracy. The above formula is for illustrative purposes. In specific implementation, the formula corresponding to the appropriate rock physics model can be selected according to the implementation situation, and no limitation is made here.

[0035] The second constrained neural network has a second specified number of layers. For example, the second constrained neural network has a five-layer neural network structure. The second specified layer, such as the first two layers, is a fully connected layer. Among them, the number of neurons in the first layer is twice that of the second layer, which is used for feature extraction; the number of neurons in the third layer is halved on the basis of the second layer. The third layer is a constraint layer, and an additional input neuron is added as a constraint element. This layer is a rock physics model constraint layer; the number of neurons in the fourth layer is twice that of the third layer and is fully connected to the neurons in the third layer, which is used for data refinement; the last layer is the output layer.

[0036] After training the first neural network, based on the trained first neural network and the second sample set, train the second constrained neural network. During training, the second sample set can be divided by using the K-Fold cross-validation method and divided into a test sample set and a validation sample set for training.

[0037] The types of the first neural network and the second constrained neural network include, for example, single-well models, regional models, and general models. The type is determined according to the quantity size, distribution range, etc. of the sample set. Construct appropriate types of the first neural network and the second constrained neural network according to the implementation situation.

[0038] Step S104: Obtain the well logging parameters of the well to be predicted and input them into the first neural network and the second constrained neural network respectively to obtain the permeability of the well to be predicted.

[0039] The trained first neural network and the second constrained neural network can be applied to the entire well section. Based on the well logging parameters of the well to be predicted, input them into the trained first neural network and the second constrained neural network, and the permeability of the well to be predicted can be obtained, which helps to complete the well logging permeability evaluation of the well to be predicted.

[0040] According to the permeability calculation method based on the rock physics constrained neural network provided by the embodiments of the present invention, the neural network structure is improved. By adding neurons as constraint elements in the constraint layer of the second constrained neural network, the calculation of the second constrained neural network is constrained, the influence of the core distribution is reduced, and a neural network that is more in line with the actual situation and has better generalization ability is constructed, so that the permeability of the well logging of the entire well section can be predicted quickly and accurately.

[0041] Figure 2 FIG. 2 shows a schematic diagram of the structure of a permeability calculation device based on a rock physics constrained neural network provided by an embodiment of the present invention. Figure 2 As shown, the device comprises:

[0042] A sample building module 210 is adapted to extract corresponding depth point logging parameters based on the core samples, and pre-process the logging parameters and the corresponding core data to build a first sample set and a second sample set;

[0043] A first training module 220, adapted to construct and train a first neural network according to a first sample set;

[0044] The second training module 230 is adapted to construct and train a second constrained neural network according to the second sample set; wherein the second constrained neural network is composed of an input layer, a constraint layer, a hidden layer, and an output layer, and the constraint layer is a network layer for specifying the addition of constraint elements; the constraint elements are calculated according to the output results of the first neural network, input to the constraint layer of the second constrained neural network and constrain the calculation of the second constrained neural network;

[0045] The prediction module 240 is adapted to obtain the logging parameters of the well to be predicted and input them into the first neural network and the second constraint neural network respectively to obtain the permeability of the well to be predicted.

[0046] Optionally, the build sample module 210 is further adapted to:

[0047] Normalize the logging parameters to unify the measurement units of the logging parameters;

[0048] Draw the intersection chart, remove outliers, and use interpolation method to fill in the default values ​​required by logging parameters.

[0049] Optionally, the constraint elements of the constraint layer are connected to the output layer of the first neural network and are disconnected from the neurons of the input layer of the second constraint neural network, and the neurons of the non-constraint elements in the constraint layer are fully connected to the neurons of the input layer of the second constraint neural network.

[0050] Optionally, the constraint layer of the second constraint neural network is constrained by a rock physics model, and constraint element calculations of the constraint layer are obtained based on the rock physics model.

[0051] Optionally, the constraint element input of the second constraint neural network is provided by the output of the first neural network, the constraint element output is fully connected to the hidden layer, and the weight and bias of the constraint element are adjusted together with other neurons.

[0052] Optionally, the first sample set consists of logging parameters and the matched irreducible water saturation; the second sample set consists of logging parameters and the matched core permeability; the input layer of the first neural network is the logging parameters of the first sample set; the output layer is the irreducible water saturation of the first sample set; the input layer of the second constrained neural network is the logging parameters of the second sample set; the output layer is the permeability of the second sample set.

[0053] Optionally, the types of the first neural network and the second constrained neural network include: single-well model, regional model, and general model; the type is determined according to the quantity and / or distribution range of the sample set.

[0054] The descriptions of the above modules refer to the corresponding descriptions in the method embodiments and will not be elaborated herein.

[0055] An embodiment of the present invention also provides a non-volatile computer storage medium, and the computer storage medium stores at least one executable instruction, and the executable instruction can execute the permeability calculation method based on the petrophysical constrained neural network in any of the above method embodiments.

[0056] Figure 3 The structural schematic diagram of a computing device according to an embodiment of the present invention is shown, and the specific embodiments of the embodiment of the present invention do not limit the specific implementation of the computing device.

[0057] As Figure 3 shown, the computing device may include: a processor 302, a communication interface 304, a memory 306, and a communication bus 308.

[0058] It is characterized in that:

[0059] The processor 302, the communication interface 304, and the memory 306 communicate with each other through the communication bus 308.

[0060] The communication interface 304 is used to communicate with network elements of other devices such as clients or other servers.

[0061] The processor 302 is used to execute the program 310, and specifically can execute the relevant steps in the above method embodiment of the permeability calculation method based on the petrophysical constrained neural network.

[0062] Specifically, the program 310 may include program code, and the program code includes computer operation instructions.

[0063] The processor 302 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention. One or more processors included in the computing device may be of the same type of processor, such as one or more CPUs; or may be of different types of processors, such as one or more CPUs and one or more ASICs.

[0064] A memory 306 for storing a program 310. The memory 306 may include high-speed RAM memory and may also include non-volatile memory, such as at least one disk memory.

[0065] The program 310 may specifically be configured to cause the processor 302 to execute the permeability calculation method based on the petrophysical constraint neural network in any of the above method embodiments. For the specific implementation of each step in the program 310, reference may be made to the corresponding steps and units in the above embodiments of the permeability calculation based on the petrophysical constraint neural network, which will not be elaborated herein. Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described devices and modules may refer to the corresponding process descriptions in the foregoing method embodiments, which will not be repeated herein.

[0066] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems may also be used in conjunction with the teachings provided herein. The structure required to construct such systems will be apparent from the above description. In addition, the embodiments of the present invention are not directed to any specific programming language. It should be understood that the content of the embodiments of the present invention described herein can be implemented using various programming languages, and the description of the specific language above is for the purpose of disclosing the preferred embodiments of the present invention.

[0067] In the specification provided herein, a large number of specific details are set forth. However, it is understood that the embodiments of the present invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0068] Similarly, it should be understood that, in order to streamline the embodiments of the present invention and assist in understanding one or more of the various inventive aspects, in the above description of the exemplary embodiments of the present invention, the various features of the embodiments of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, the disclosed method should not be construed as reflecting an intention that the claimed embodiments of the present invention require more features than are expressly recited in each claim. Rather, as reflected in the following claims, the inventive aspects lie in less than all the features of the single embodiments disclosed previously. Thus, the claims following the detailed description are hereby expressly incorporated into the detailed description, with each claim standing on its own as a separate embodiment of the present invention.

[0069] Those skilled in the art can understand that the modules in the devices in the embodiments can be adaptively changed and disposed in one or more devices different from the embodiments. The modules or units or components in the embodiments can be combined into one module or unit or component, and in addition, they can be divided into multiple sub-modules or sub-units or sub-components. Except that at least some of such features and / or processes or units are mutually exclusive, any combination can be adopted to combine all the features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all the processes or units of any method or device so disclosed. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) can be replaced by an alternative feature that provides the same, equivalent, or similar purpose.

[0070] In addition, those skilled in the art can understand that although some of the embodiments herein include certain features included in other embodiments rather than other features, the combination of the features of different embodiments means that it is within the scope of the present invention and forms different embodiments. For example, in the following claims, any one of the claimed embodiments can be used in any combination.

[0071] Each component embodiment of the present invention can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art should understand that a microprocessor or a digital signal processor (DSP) can be used in practice to implement some or all of the functions of some or all of the components according to the embodiments of the present invention. Embodiments of the present invention can also be implemented as a device or apparatus program (e.g., a computer program and a computer program product) for performing part or all of the methods described herein. Such a program implementing the embodiments of the present invention can be stored on a computer-readable medium, or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, or provided on a carrier signal, or in any other form.

[0072] It should be noted that the above embodiments illustrate the embodiments of the present invention rather than limit the present invention, and those skilled in the art can design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in the claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. Embodiments of the present invention can be implemented by means of hardware including several different elements and by means of a suitably programmed computer. In the unit claims listing several devices, several of these devices can be embodied by the same item of hardware. The use of the words first, second, and third, etc. does not denote any order. These words can be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be construed as limiting the order of execution.

Claims

1. A method for calculating permeability based on a rock physics-constrained neural network, characterized in that, The method includes: Extracting corresponding depth-point logging parameters based on core samples, and preprocessing the logging parameters and corresponding core data to construct a first sample set and a second sample set; Constructing and training a first neural network according to the first sample set; Constructing and training a second constrained neural network according to the second sample set; wherein, the second constrained neural network consists of an input layer, a constraint layer, a hidden layer, and an output layer, and the constraint layer is a network layer with specified constraint elements added; the constraint elements are calculated based on the output result of the first neural network, input to the constraint layer of the second constrained neural network and used to constrain the calculation of the second constrained neural network; the constraint elements of the constraint layer are in a connected state with the output layer of the first neural network and in a non-connected state with the neurons of the input layer of the second constrained neural network, and the neurons of the non-constraint elements in the constraint layer are in a fully connected state with the neurons of the input layer of the second constrained neural network; the output of the constraint elements is fully connected to the hidden layer, and the weights and biases of the constraint elements are jointly adjusted with other neurons; Obtaining the logging parameters of the well to be predicted, and inputting them into the first neural network and the second constrained neural network respectively to obtain the permeability of the well to be predicted.

2. The method according to claim 1, wherein The preprocessing of the logging parameters and corresponding core data further includes: Normalizing the logging parameters to unify the measurement units of the logging parameters; Drawing a crossplot, removing outliers, and using interpolation to fill in the missing values required for the logging parameters.

3. The method according to claim 1, wherein The constraint layer of the second constrained neural network is constrained by a rock physics model, and the constraint elements of the constraint layer are calculated according to the rock physics model.

4. The method according to claim 1, characterized in that, The input of the constraint elements of the second constrained neural network is provided by the output of the first neural network.

5. The method according to claim 1, wherein The first sample set consists of the logging parameters and the matching irreducible water saturation; the second sample set consists of the logging parameters and the matching core permeability; the input layer of the first neural network is the logging parameters of the first sample set; the output layer is the irreducible water saturation of the first sample set; the input layer of the second constrained neural network is the logging parameters of the second sample set; the output layer is the permeability of the second sample set.

6. The method according to claim 1, characterized in that, The types of the first neural network and the second constrained neural network include: single-well model, regional model, and general model; the type is determined according to the quantity size and / or distribution range of the sample set.

7. A permeability calculation device based on a rock physics constrained neural network, characterized in that the device includes: A sample construction module, adapted to extract corresponding depth-point logging parameters based on core samples, and preprocess the logging parameters and corresponding core data to construct a first sample set and a second sample set; A first training module, adapted to construct and train a first neural network according to the first sample set; The second training module is adapted to construct and train a second constrained neural network according to a second sample set; wherein, the second constrained neural network is composed of an input layer, a constraint layer, a hidden layer and an output layer, and the constraint layer is a network layer where constraint elements are specified to be added; the constraint elements are calculated according to the output result of the first neural network, input to the constraint layer of the second constrained neural network and used to constrain the calculation of the second constrained neural network; the constraint elements of the constraint layer are in a connection state with the output layer of the first neural network and in a non-connection state with the neurons of the input layer of the second constrained neural network, and the neurons of the constraint layer that are not constraint elements are in a fully connected state with the neurons of the input layer of the second constrained neural network; the output of the constraint elements is fully connected to the hidden layer, and the weights and biases of the constraint elements are jointly adjusted with other neurons. The prediction module is adapted to obtain the logging parameters of a well to be predicted and input them into the first neural network and the second constrained neural network respectively to obtain the permeability of the well to be predicted.

8. A computing device, comprising: A processor, a memory, a communication interface and a communication bus, where the processor, the memory and the communication interface complete communication with each other through the communication bus. The memory is used to store at least one executable instruction, and the executable instruction causes the processor to perform the operations corresponding to the permeability calculation method based on the rock physics constrained neural network according to any one of claims 1-6.

9. A computer storage medium, where at least one executable instruction is stored in the storage medium, and the executable instruction causes the processor to perform the operations corresponding to the permeability calculation method based on the rock physics constrained neural network according to any one of claims 1-6.

Images