Volcanic rock type identification method and device, electronic equipment and storage medium

Abstract

The invention provides a volcanic rock type identification method and device, electronic equipment and a storage medium. The volcanic rock type identification method comprises the steps that sensitive logging data, used for representing volcanic rock types, of a target work area is acquired; according to the sensitive logging data and a pre-generated volcanic rock type identification model, identifying the volcanic rock type of the target work area; wherein the volcanic rock type identification model is generated based on a BP neural network and historical data of the sensitive logging data in the target work area. Logging, logging and seismic data are comprehensively utilized, volcanic rock types and distribution characteristics are identified by coupling a BP neural network and wave impedance inversion, and the precision is further improved.

Description

Technical Field

[0001] The present application belongs to the technical field of oil and gas field seismic exploration, and specifically relates to a method and device for identifying volcanic rock types. Background Art

[0002] When drilling fractured volcanic rocks, igneous rock thickness is a crucial factor in pre-drilling engineering and geological design. In particular, mud density is closely linked to wellbore stability during drilling. High-quality volcanic rock also represents a favorable reservoir in the study area, facilitating the discovery of oil and gas reservoirs. Accurately identifying volcanic rock type and thickness is crucial for economical and efficient drilling and oil and gas discovery. Traditionally, volcanic rock type and thickness identification relies on qualitative methods such as seismic analysis, which cannot quantitatively identify volcanic rock thickness before drilling within the constraints of well completion. Summary of the Invention

[0003] One purpose of the present invention is to provide a volcanic rock type identification method that comprehensively utilizes well logging, mud logging, and seismic data to identify the type and thickness of volcanic rocks by coupling BP neural network and wave impedance inversion, thereby further improving the accuracy.

[0004] Another object of the present invention is to provide a device for identifying volcanic rock types. Another object of the present invention is to provide an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned method for identifying volcanic rock types are implemented. Another object of the present invention is to provide a readable medium storing the computer program, and when the processor executes the computer program, the steps of the above-mentioned method for identifying volcanic rock types are implemented.

[0005] In order to solve the technical problems in the background technology of this application, the present invention provides the following technical solutions:

[0006] In a first aspect, the present invention provides a method for identifying volcanic rock types, comprising:

[0007] Obtain sensitive logging data for characterizing volcanic rock types in the target area;

[0008] The volcanic rock type of the target work area is identified based on the sensitive well logging data and a pre-generated volcanic rock type identification model; wherein the volcanic rock type identification model is generated based on a BP neural network and historical data of the sensitive well logging data in the target work area.

[0009] In some embodiments of the present invention, a method for identifying volcanic rock types further includes:

[0010] performing sparse pulse wave impedance inversion on the identified volcanic rock type to determine a wave impedance threshold range corresponding to the identified volcanic rock type;

[0011] The distribution characteristics corresponding to the identified volcanic rock type are predicted according to the wave impedance threshold range.

[0012] In some embodiments of the present invention, a method for identifying volcanic rock types further includes:

[0013] selecting the sensitive well logging data;

[0014] Selecting the sensitive logging data includes:

[0015] Selecting at least two well logging data from a plurality of well logging data in the target work area for intersection to generate an intersection result;

[0016] The sensitive well logging data is selected from the plurality of well logging data according to the intersection result.

[0017] In some embodiments of the present invention, the step of generating a volcanic rock type identification model includes:

[0018] Initializing the initial weight, threshold, and learning rate of the BP neural network to generate an initial model of the volcanic rock type identification model;

[0019] The initial model is trained using the historical data; wherein, during the training of the initial model, the neuron state of the current layer is generated only by the neuron state of the previous layer, and is propagated layer by layer to the hidden layer; and

[0020] The input signal of the hidden layer is equal to the weighted sum of the input signals of the input layer;

[0021] The output signal of the current output layer is equal to the output value of the previous hidden layer; wherein the output value is generated by mapping the output signal of the previous hidden layer through an activation function.

[0022] In some embodiments of the present invention, when the output signal of the output layer does not meet a preset threshold, the method further includes:

[0023] Inputting the error signal into the output layer and propagating it toward the input layer;

[0024] During the propagation to the input layer, the connection weights of each input layer, hidden layer and output layer and the bias values ​​of the neurons are adjusted.

[0025] In some embodiments of the present invention, performing sparse pulse wave impedance inversion on the identified volcanic rock type to determine a wave impedance threshold range corresponding to the identified volcanic rock type includes:

[0026] obtaining seismic data of the identified volcanic rock type;

[0027] performing synthetic record calibration on the seismic data to generate a calibration result;

[0028] Establishing a low-frequency model based on the pre-selected seismic wavelet and the calibration result;

[0029] The wave impedance threshold range is determined according to the low-frequency model.

[0030] In some embodiments of the present invention, before performing sparse pulse wave impedance inversion on the identified volcanic rock type to determine the wave impedance threshold range corresponding to the identified volcanic rock type, the method further includes:

[0031] The identified volcanic rock types are corrected based on the well logging data of the standard layer and the logging data of the bottom layer outside the standard layer.

[0032] In a second aspect, the present invention provides a volcanic rock type identification device, the device comprising:

[0033] A sensitive logging data acquisition module is used to obtain sensitive logging data for characterizing volcanic rock types in the target work area;

[0034] A volcanic rock type identification module is used to identify the volcanic rock type of the target work area based on the sensitive well logging data and a pre-generated volcanic rock type identification model; wherein the volcanic rock type identification model is generated based on a BP neural network and historical data of the sensitive well logging data in the target work area.

[0035] In some embodiments of the present invention, a volcanic rock type identification device further includes:

[0036] a wave impedance threshold determination module, configured to perform sparse pulse wave impedance inversion on the identified volcanic rock type to determine a wave impedance threshold range corresponding to the identified volcanic rock type;

[0037] A distribution feature prediction module is used to predict the distribution feature corresponding to the identified volcanic rock type according to the wave impedance threshold range.

[0038] In some embodiments of the present invention, a volcanic rock type identification device further includes:

[0039] A sensitive data selection module, used for selecting the sensitive well logging data;

[0040] The sensitive data selection module includes:

[0041] An intersection result generating unit, configured to select at least two well logging data from the plurality of well logging data in the target work area for intersection, so as to generate an intersection result;

[0042] A sensitive data selection unit is used to select the sensitive well logging data from the multiple well logging data according to the intersection result.

[0043] In some embodiments of the present invention, a volcanic rock type identification device further includes:

[0044] A model generation module is used to generate a volcanic rock type identification model, and the model generation module includes:

[0045] an initial model generating unit, configured to initialize the initial weight, threshold, and learning rate of the BP neural network to generate an initial model of the volcanic rock type identification model;

[0046] An initial model training unit, configured to train the initial model using the historical data; wherein, during the training of the initial model, the neuron state of the current layer is generated only by the neuron state of the previous layer and propagated layer by layer to the hidden layer; and

[0047] The input signal of the hidden layer is equal to the weighted sum of the input signals of the input layer;

[0048] The output signal of the current output layer is equal to the output value of the previous hidden layer; wherein the output value is generated by mapping the output signal of the previous hidden layer through an activation function.

[0049] In some embodiments of the present invention, when the output signal of the output layer does not meet a preset threshold, a volcanic rock type identification device further includes:

[0050] an error signal input module, configured to input the error signal into the output layer and propagate it toward the input layer;

[0051] The numerical adjustment module is used to adjust the connection weights of each input layer, hidden layer and output layer and the bias value of the neuron during the process of propagation to the input layer.

[0052] In some embodiments of the present invention, the wave impedance threshold determination module includes:

[0053] a seismic data acquisition unit, configured to acquire seismic data of the identified volcanic rock type;

[0054] a record calibration unit, configured to perform synthetic record calibration on the seismic data to generate a calibration result;

[0055] A low-frequency model building unit, configured to build a low-frequency model based on the pre-selected seismic wavelet and the calibration result;

[0056] The wave impedance threshold determination unit is used to determine the wave impedance threshold range according to the low-frequency model.

[0057] In some embodiments of the present invention, a volcanic rock type identification device further includes:

[0058] The type correction module is used to correct the identified volcanic rock type based on the well logging data of the standard layer and the logging data of the bottom layer outside the standard layer.

[0059] In a third aspect, the present invention provides a computer program product, comprising a computer program / instruction, which implements the steps of a volcanic rock type identification method when executed by a processor.

[0060] In a fourth aspect, the present invention provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, steps of a method for identifying volcanic rock types are implemented.

[0061] In a fifth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a method for identifying volcanic rock types.

[0062] As can be seen from the foregoing description, embodiments of the present invention provide a method and apparatus for identifying volcanic rock types. The corresponding volcanic rock type identification method includes: first, obtaining sensitive well logging data for characterizing the volcanic rock type in a target work area; then, identifying the volcanic rock type in the target work area based on the sensitive well logging data and a pre-generated volcanic rock type identification model; wherein the volcanic rock type identification model is generated based on a BP neural network and historical data of the sensitive well logging data in the target work area.

[0063] The corresponding volcanic rock type identification device includes: a sensitive logging data acquisition module for acquiring sensitive logging data for characterizing the volcanic rock type in the target work area; and a volcanic rock type identification module for identifying the volcanic rock type in the target work area based on the sensitive logging data and a pre-generated volcanic rock type identification model. The volcanic rock type identification model is generated based on a BP neural network and historical data of sensitive logging data in the target work area.

[0064] The present invention comprehensively utilizes logging, mud recording, and seismic data to accurately predict the type and distribution characteristics of volcanic rocks based on the clear physical response characteristics of sensitive volcanic rocks, by coupling BP neural network and wave impedance inversion. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0066] Figure 1 A schematic flow chart of a method for identifying volcanic rock types in an embodiment of the present invention;

[0067] Figure 2 Another schematic flow chart of a method for identifying volcanic rock types according to an embodiment of the present invention;

[0068] Figure 3 This is a third flow chart of a method for identifying volcanic rock types in an embodiment of the present invention;

[0069] Figure 4 This is a flow chart of step 500 of a method for identifying volcanic rock types in an embodiment of the present invention;

[0070] Figure 5 This is a fourth flow chart of a method for identifying volcanic rock types in an embodiment of the present invention;

[0071] Figure 6 A schematic flow chart of step 600 of a method for identifying volcanic rock types in an embodiment of the present invention;

[0072] Figure 7 This is another flowchart diagram of step 600 of a method for identifying volcanic rock types in an embodiment of the present invention;

[0073] Figure 8 This is a flow chart of step 300 of a method for identifying volcanic rock types in an embodiment of the present invention;

[0074] Figure 9 This is a schematic flow chart of a method for identifying volcanic rock types in a specific embodiment of the present invention;

[0075] Figure 10 This is a cross-plot of volcanic rock logging curves (DEN-GR) in the study area in a specific embodiment of the present invention;

[0076] Figure 11 It is the cross plot of volcanic rock logging curves (CNL-SP) in the study area in a specific embodiment of the present invention;

[0077] Figure 12 The frequency histogram (CNL) of the standard layer logging curve in the specific embodiment of the present invention;

[0078] Figure 13 The frequency histogram (SP) of the standard layer logging curve in the specific embodiment of the present invention;

[0079] Figure 14 The frequency histogram (DEN) of the standard layer logging curve in the specific embodiment of the present invention;

[0080] Figure 15 The frequency histogram (GR) of the standard layer logging curve in the specific embodiment of the present invention;

[0081] Figure 16 A schematic diagram of the BP network operation structure in a specific embodiment of the present invention;

[0082] Figure 17 BP network algorithm flow chart in a specific embodiment of the present invention;

[0083] Figure 18 This is a model training optimization curve diagram in a specific embodiment of the present invention;

[0084] Figure 19 This is a comparison chart of validation set loss prediction results in a specific embodiment of the present invention;

[0085] Figure 20 This is a time-thickness distribution diagram of dacite in the study area in a specific embodiment of the present invention;

[0086] Figure 21 is a block diagram of a volcanic rock type identification device in an embodiment of the present invention;

[0087] Figure 22 Schematic diagram of the structure of an electronic device in an embodiment of the present invention. DETAILED DESCRIPTION

[0088] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0089] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0090] It should be noted that the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or are inherent to these processes, methods, products or devices. The embodiments in this application and the features described in the embodiments may be combined with each other unless there is a conflict. The present application will be described in detail below with reference to the drawings and in conjunction with the embodiments.

[0091] The acquisition, storage, use, and processing of data in the technical solution of this application comply with relevant laws and regulations.

[0092] In the prior art, the application number is CN201611142635.0, a method for predicting volcanic rock facies, which relates to the technical field of oil development and reservoir engineering, and includes the following steps: (1) model forward modeling: based on the calibration of seismic geological layers, a profile model is forward modeled; (2) seismic attribute extraction: for the target layer, seismic attributes along the layer are extracted and the attribute time window range is determined; (3) attribute optimization: seismic attributes that reflect the distribution of volcanic rocks are optimized and preprocessed; (4) volcanic rock facies classification: volcanic rock facies are classified according to drilling results; (5) multi-attribute neural network pattern recognition; (6) facies prediction: multiple seismic attributes are fused according to the trained neural network parameters to predict the distribution and change law of volcanic rock facies in unknown areas. The prediction method of the invention integrates multiple seismic attributes and drilling information, overcomes the disadvantage that a single seismic attribute can only partially reflect the difference in lithologic combinations, and can effectively predict the spatial distribution of volcanic rock facies.

[0093] This invention uses seismic attributes to divide volcanic rock phases through a neural network method. However, the identification of volcanic rocks through seismic attributes still has certain limitations. It is not precise enough, the boundaries of rock type distinction are not clear, and it is easily affected by objective factors, such as software calculation methods and human subjective factors. Therefore, a combination of multiple methods is required for identification.

[0094] Example 1:

[0095] Based on the above reasons, the embodiment of the present invention provides a specific implementation method of a volcanic rock type identification method, see Figure 1 , specifically including the following contents:

[0096] Step 100: Acquire sensitive logging data for characterizing volcanic rock types in a target work area;

[0097] Step 200: Identify the volcanic rock type of the target work area based on the sensitive well logging data and a pre-generated volcanic rock type identification model; wherein the volcanic rock type identification model is generated based on a BP neural network and historical data of the sensitive well logging data in the target work area.

[0098] As can be seen from the foregoing description, an embodiment of the present invention provides a method for identifying volcanic rock types, comprising: first, obtaining sensitive well logging data for characterizing the volcanic rock type in a target work area; then, identifying the volcanic rock type in the target work area based on the sensitive well logging data and a pre-generated volcanic rock type identification model; wherein the volcanic rock type identification model is generated based on a BP neural network and historical data of the sensitive well logging data in the target work area.

[0099] The present invention provides a volcanic rock type identification method based on coupled BP neural network and wave impedance inversion, which effectively solves the difficulties in pre-drilling drilling engineering design and geological design, thereby avoiding the problem of mud loss caused by volcanic rocks and the problem of predicting high-quality volcanic rock reservoirs.

[0100] Example 2:

[0101] Regarding step 100, volcanic rock is a type of rock formed by magma ejected from a volcano and condensed. It is a type of igneous rock. It can be divided into two categories: surface volcanic rock and deep volcanic rock based on its formation location and silica content:

[0102] Epivolcanic rocks: These rocks form at the Earth's surface and include: Basalt: Formed primarily from the cooling of magma containing 45% to 54% silica, it can range in color from dark gray to nearly black. Andesite: Formed from the cooling of magma containing 54% to 62% silica, it is typically medium gray. Dacite: Formed from the cooling of magma containing 62% to 70% silica, it ranges in color from light brown to gray. Rhyolite: Formed from the cooling of magma containing 70% to 78% silica, it is typically gray.

[0103] Deep volcanic rocks: These rocks form when magma solidifies underground. They include: Gabbro, which solidifies from magma containing 45% to 54% silica. Diorite, which solidifies from magma containing 54% to 62% silica. Granodiorite, which solidifies from magma containing 62% to 70% silica. Granite, which solidifies from magma containing 70% to 78% silica.

[0104] The BP neural network in step 200 is a multi-layer feedforward neural network trained by error back propagation. In this application, the BP neural network learning process includes two stages: forward propagation of signals and back propagation of errors.

[0105] During the forward propagation phase, the input is passed from the input layer through the hidden layers to the output layer. The difference between the actual output of the output layer and the expected output (i.e., the labels in the training dataset) is then used to calculate the error value.

[0106] During the backpropagation phase, the error propagates backwards from the output layer to the input layer, layer by layer, while adjusting the weights and biases of each neuron. Backpropagation of error is based on the chain rule, which calculates the partial derivative (gradient) of the error with respect to each weight, thereby achieving layer-by-layer error propagation.

[0107] In this application, the process of training a BP neural network includes the following steps: All connection weights and biases in the network are randomly initialized. Input samples are propagated forward through the network until an output is generated. The error between the network output and the expected output is calculated. The error is propagated back from the output layer to the input layer, and the error contribution of each node in each layer is calculated. Using gradient descent or other optimization algorithms, the weights and biases are updated based on the error contribution. These steps are repeated until a stopping condition is met (e.g., a preset number of iterations is reached or the error falls to a lower limit).

[0108] In addition, during the training of BP application network, optimization can be performed in the following three ways:

[0109] The loss function preferably includes mean square error and cross entropy error.

[0110] The learning rate is an important hyperparameter that controls the magnitude of weight updates. Setting it too high may lead to unstable learning, while setting it too low may result in slow learning.

[0111] Momentum is a method used to accelerate the learning process, help avoid local minima and improve training stability.

[0112] In some embodiments of the present invention, see Figure 2 , a volcanic rock type identification method, further comprising:

[0113] Step 300: performing sparse pulse wave impedance inversion on the identified volcanic rock type to determine a wave impedance threshold range corresponding to the identified volcanic rock type;

[0114] The sparse pulse wave impedance inversion here mainly includes maximum likelihood deconvolution, L1 norm deconvolution and minimum entropy deconvolution:

[0115] Maximum likelihood deconvolution: Using state space and system identification methods, the seismic wavelet is described by the autoregressive-sliding average model, and the reflection coefficient is described by the Gaussian-Bernoulli sequence. The quadratic objective function (likelihood function) is derived from this, and then the nonlinear optimization method is used to gradually maximize it in the iterative process.

[0116] Minimum Entropy Deconvolution: The output is a sparse pulse train, resulting in a simple and sparse output shape. It places no restrictions on the phase of the wavelet, resulting in strong adaptability. The minimum entropy criterion is only a reasonable objective function or constraint and does not fully control the deconvolution quality. The criterion is very sensitive to strong reflectors. If unconstrained, it may cause multiple adjacent reflection coefficients to be compressed into a single large reflection coefficient, reducing resolution.

[0117] Step 400: predicting the distribution characteristics corresponding to the identified volcanic rock type according to the wave impedance threshold range.

[0118] It should be noted that the distribution characteristics in step 500 include identifying the depth, thickness, distribution range and extension trend of the volcanic rock formation corresponding to the volcanic rock type.

[0119] In some embodiments of the present invention, see Figure 3 , a volcanic rock type identification method, further comprising:

[0120] Step 500: Select the sensitive logging data; then, see Figure 4 , step 500 includes:

[0121] Well logging data can reflect the characteristics of different lithologies and horizons, and then determine the specific lithology and horizon based on the obtained curve. Generally, well logging data includes the following:

[0122] The natural potential curve is symmetrical about the center of the permeable formation when the formation and mud are uniform and the upper and lower surrounding rocks have the same lithology. The natural potential of the permeable layer varies the most at the interface between the top and bottom of the formation. When the formation thickness is greater than four times the wellbore diameter, the formation interface can be determined by the curve's half-amplitude point. The natural potential of a permeable formation can deflect to the left or right relative to the mudstone baseline, depending primarily on the relative salinity of the formation water and mud filtrate. The natural potential curve is influenced by lithology, the ratio of formation water salinity to mud filtrate salinity, formation thickness, wellbore diameter, formation resistivity, mud resistivity, surrounding rock resistivity, and mud intrusion zones.

[0123] Acoustic transit time curve: In sandstone and mudstone sections, sandstone generally has a very high velocity, resulting in a lower transit time curve value. The magnitude of the transit time is influenced by the properties, type, and content of the sandstone cement. Siliceous and calcareous cements generally have lower transit times than mudstone cements. The transit time decreases with increasing calcium content and increases with increasing mud content. Mudstone has a high transit time, while siltstone and shale fall between these two groups. Conglomerate generally has a low transit time. Shallow gas-bearing formations exhibit frequency jumps or increased transit time. This is primarily influenced by factors such as wellbore diameter, formation thickness, and frequency jumps.

[0124] Microelectrode curve: Mudstone, the microelectrode curve has a low amplitude, no amplitude difference or a very small positive or negative irregular amplitude difference, and the curve is straight. The microelectrode curve amplitude of dense sandstone or calcareous sandstone is particularly high, often jagged or bayonet-shaped, with positive or negative amplitude differences of varying sizes. The microelectrode amplitude of biogenic limestone is very high and the positive amplitude difference is large. The amplitude value of siltstone is low, with a smaller positive amplitude difference. The amplitude value of porous limestone is much lower than that of dense limestone, and generally has an obvious positive amplitude difference.

[0125] Natural gamma ray curve: In a sandstone-mudstone profile, pure sandstone has the lowest GR, clay has the highest, argillaceous sandstone has a lower GR, and argillaceous siltstone and sandy mudstone have higher GR. In other words, the natural gamma ray value increases with increasing argillaceous content. This is mainly affected by formation thickness, borehole, radioactivity fluctuation error, and velocity measurement.

[0126] It is understandable that some logging curves are more sensitive to certain types of volcanic rocks, so it is extremely necessary to select appropriate logging curves as sensitive logging data to characterize volcanic rock types.

[0127] Step 501: selecting at least two well logging data from a plurality of well logging data of the target work area for intersection to generate an intersection result;

[0128] Step 502: Select the sensitive logging data from the plurality of logging data according to the intersection result.

[0129] In steps 501 and 502, the petrophysical response characteristics of the volcanic rock are clarified by analyzing the logging and geological data of the target block. The crossplot method is used to optimize the sensitive logging curves of the volcanic rock, and a variable set of four input parameters is determined: natural gamma ray logging (GR), density logging (DEN), spontaneous potential logging (SP), and neutron logging (CNL).

[0130] In some embodiments of the present invention, see Figure 5 , a volcanic rock type identification method, further comprising:

[0131] Step 600: Generate a volcanic rock type identification model, then refer to Figure 6, step 600 includes:

[0132] Step 601: Initializing the initial weight, threshold, and learning rate of the BP neural network to generate an initial model of the volcanic rock type identification model;

[0133] Initialize the original variable parameters, set the initial weights, thresholds, and learning rates to generate the initial model.

[0134] Step 602: training the initial model using the historical data; wherein, during the training of the initial model, the neuron state of the current layer is generated only by the neuron state of the previous layer, and propagates to the hidden layer layer by layer; and

[0135] The input signal of the hidden layer is equal to the weighted sum of the input signals of the input layer;

[0136] The output signal of the current output layer is equal to the output value of the previous hidden layer; wherein the output value is generated by mapping the output signal of the previous hidden layer through an activation function.

[0137] In step 602, a given sample is input and the input and output values ​​of each layer are calculated through the sigmoid action function. The state of each layer of neurons only affects the state of the neurons in the next layer and gradually propagates to the hidden layer. The input of the hidden layer is equal to the weighted sum of the input layer signals, and the output layer is equal to the output value of the output of the previous hidden layer after being mapped by the excitation function, generating an output signal at the output end.

[0138] In some embodiments of the present invention, see Figure 7 When the output signal of the output layer does not meet the preset threshold, step 600 further includes:

[0139] Step 603: Input the error signal to the output layer and propagate it toward the input layer;

[0140] Step 604: During the propagation process to the input layer, the connection weights of the input layer, the hidden layer, and the output layer and the bias values ​​of the neurons are adjusted.

[0141] In steps 603 and 604, if the ideal output value cannot be obtained in the output layer, the system enters the error signal back propagation process. The error signal propagates from the output layer to the input layer and adjusts the connection weights between each layer and the bias values ​​of the neurons in each layer along the way. The error function gradient descent strategy is executed to continuously reduce the error signal, and the weights are continuously adjusted to minimize the network error function.

[0142] Furthermore, based on the gap between the predicted value of volcanic rock type and the measured value of volcanic rock type, backpropagation is performed to update the weights of the neural network parameters accordingly, and the neural network is modified according to the weights. The model prediction value approaches the target value during continuous modification until a certain accuracy rate or number of training times is reached.

[0143] In some embodiments of the present invention, see Figure 8 , step 300 includes:

[0144] Step 301: Acquire seismic data of the identified volcanic rock type;

[0145] Step 302: performing synthetic record calibration on the seismic data to generate a calibration result;

[0146] Specifically, the synthetic seismograms are compared and matched with actual acquired seismic data. This process is intended to ensure the accuracy of seismic interpretation, provide precise identification of geological horizons, and improve the processing and interpretation of seismic data.

[0147] First, acoustic and density logging data are acquired to calculate acoustic impedance. Next, a series of reflection coefficients is generated: changes in acoustic impedance generate reflections at formation interfaces. By calculating the differences in acoustic impedance between adjacent layers, a series of reflection coefficients is generated, representing the reflection strength of seismic waves at different formation interfaces.

[0148] A synthetic seismogram is generated by convolving the reflection coefficient sequence with a seismic waveform (a source waveform, such as a Ricker wavelet). The source waveform represents the shape of the seismic wave as it propagates through the strata from the source. By comparing the synthetic seismogram with the actual seismogram, the time-depth relationship is adjusted to align the reflection events on the synthetic seismogram with the events of the actual seismic data (i.e., reflection events of the same geological age). If necessary, the synthetic seismogram is corrected to match the amplitude, frequency content, and phase characteristics of the actual seismogram.

[0149] Step 303: establishing a low-frequency model based on the pre-selected seismic wavelet and the calibration result;

[0150] A seismic wavelet is a segment of a seismic wave with a definite start time and finite energy. It is generated by a seismic source, propagates underground, and is ultimately received by detectors on the surface or in wells. Seismic wavelets are non-periodic in nature, so their dynamic parameters differ from those of periodic vibrations, such as amplitude, frequency, and phase. Here, they are preferably described using amplitude and phase spectra.

[0151] In practical applications, seismic wavelets are closely related to the properties of the formation rock, so their waveforms vary depending on the region, geological structure, and formation lithology. Seismic wavelets play a crucial role in seismic recordings, not only containing information about the propagation of seismic energy but also relating to the frequency components of the seismic waves and the output signals of the detectors.

[0152] In seismic data processing and inversion, a subsurface velocity or impedance model that includes only lower-frequency components. Seismic data often lack low-frequency information due to the ineffective excitation of low-frequency waves by the seismic source, as well as the effects of attenuation and noise in the subsurface. Low-frequency models are constructed to supplement the missing low-frequency information in seismic data and provide a stable background model for more accurate predictions of the physical properties of the subsurface during seismic inversion.

[0153] Step 304: Determine the wave impedance threshold range according to the low-frequency model.

[0154] In some embodiments of the present invention, before step 300, a volcanic rock type identification method further includes:

[0155] The identified volcanic rock types are corrected based on the well logging data of the standard layer and the logging data of the bottom layer outside the standard layer.

[0156] Specifically, based on the well logging curve, appropriate correction is made to the logging lithology of the target layer other than the standard layer;

[0157] As can be seen from the above description, the present invention provides a method for identifying volcanic rock types. Based on neural network and inversion techniques, it combines seismic and well logging data. First, the crossplot method is used to screen well logging curves sensitive to volcanic rock identification. Then, a neural network model is established using marker intervals. This eliminates erroneous data caused by well logging, making the established neural network model more accurate. This model then corrects the lithology of the well logging data. The inversion results generated using this method predict more accurate thresholds for the wave impedance of different volcanic rock types, resulting in more precise identification. This method also leverages the strengths and weaknesses of both seismic and well logging data.

[0158] Example 3:

[0159] To further illustrate the solution, the present invention also provides a specific implementation method for determining the continuity of the main displacement zone of a strike-slip fault, see Figure 9 , specifically including the following steps.

[0160] S1. Collect logging, seismic and geological data in the target area, determine the type of volcanic rock after drilling, clarify the rock physical response characteristics of the volcanic rock, and use the crossplot method to optimize the sensitive logging curve of the volcanic rock.

[0161] The collected volcanic rock types in this block are taken as true values, with tuff as value 1, dacite as value 2, and basalt as value 3 as standard values ​​for model training.

[0162] Intersection diagram Figure 10 as well as Figure 11 As shown, Figure 10 This is the intersection diagram of GR and DEN logging curves of volcanic rocks in the study area. Figure 11 It is the intersection diagram of the SP and CNL logging curves of the volcanic rocks in the study area. By combining these four logging curves in pairs to form an intersection diagram, different volcanic rocks can be clearly and quickly divided, indicating that there are clear boundaries between different types of volcanic rocks. Therefore, these four logging curves are determined to be sensitive curves, which are the input parameters of the neural network model.

[0163] S2. Standardize the sensitive logging curves;

[0164] Here, the histogram method is used to standardize sensitive logging curves. Specifically, the AC, CNL, DEN, and SP logging curves are standardized using the histogram method. The specific steps are: determining the relative standard layer; drawing the histogram; obtaining the standard value of the sensitive logging curve in the study area; and obtaining the correction value of the single well curve.

[0165] In determining the relative standard layer, the basalt sections of each single well in the study area are used as the relative standard layer. The basalt rock physical response characteristics are obvious and are widely distributed in the study area. A histogram is drawn, and the standard layer data of each single well are extracted. The logging response frequency distribution histogram is fitted by the Gaussian normal distribution function to obtain the peak reading of each logging curve.

[0166] Obtain the standard values ​​for the study area's curves. Take the logging response values ​​for each individual well and plot frequency distribution histograms for the entire area's natural gamma ray (GR), density (DEN), spontaneous potential (SP), and neutron log (CNL). Determine the peak value of the histogram for the entire area, which is the standard value for each logging curve. To obtain correction values ​​for individual well curves, calculate correction value = standard value for the entire area - peak value for the individual well.

[0167] refer to Figure 12 、 Figure 13 、 Figure 14 as well as Figure 15 This is a frequency histogram of the standard layer logging curve provided by an embodiment of the present invention. In step S2, the sensitive logging curves are normalized. Because the logging data for each well block are measured at different times, at different depths, in different downhole environments, and using different instrumentation, and because the later standardization calibration methods may not be identical, errors may occur between the logging data for different blocks. When applying this logging data, it must be normalized.

[0168] The histogram method is used to standardize the AC, CNL, DEN, and SP logging curves. The main steps are: determining the relative standard layer; drawing the histogram; obtaining the standard value of the sensitive logging curve in the study area; and obtaining the correction value of the single well curve.

[0169] To determine the relative standard layer, the basalt intervals of each well in the study area were used as the relative standard layer. Basalt rock physical response characteristics are distinct and widely distributed in the study area. Histograms were drawn, and standard layer data for each well was extracted. The logging response frequency distribution histogram was fitted using a Gaussian normal distribution function to obtain the peak readings of each logging curve. To obtain the standard values ​​of the study area curves, the logging response values ​​of each well were used to draw frequency distribution histograms of natural gamma ray logging (GR), density logging (DEN), spontaneous potential logging (SP), and neutron logging (CNL) for the entire area. The peak values ​​of the histograms for the entire area were obtained, which are the standard values ​​of each logging curve. To obtain the correction values ​​for the individual well curves, the correction value = the standard value for the entire area - the peak value of the individual well. Figure 12 as well as Figure 13 They are the SP and CNL frequency histograms of the standard layer in the study area, Figure 14 as well as Figure 15 They are the DEN and GR frequency histograms of the standard layer in the study area, which can clearly obtain the peak readings of each logging curve and perform correction analysis on each logging curve through the correction formula.

[0170] S3. Using the pre-processed standard layer sensitivity curve data as input, the volcanic rock lithology category as output, and the actual lithology category as the standard value, supervised training and optimization are performed to obtain a volcanic rock prediction neural network model;

[0171] Specifically, the input parameter InputData of the training set is set, and the four collected and organized well logging curve data are used as the data for neural network training; the output standard value OutputTarget of the training set is set, and the standard value is the rock type at the corresponding depth of the block measured; a BP neural network is established, the number of nodes in the input layer, hidden layer, and output layer is set, the transfer function is set, and the batch gradient descent algorithm is used to perform unconstrained nonlinear optimization on the model.

[0172] refer to Figure 16 This is a schematic diagram of the BP network operation structure in the technical solution of the present invention, which consists of Figure 16 As you can see, a BP neural network, also known as an error signal feedback network, is a multi-stage feedback network that considers both forward signal propagation and backward error propagation. A BP neural network is primarily a feedforward network with three or more network layers. The first and last layers are typically called the input and output layers, while the intermediate layers are called hidden layers. There is no feedback between networks, and they are not connected at the network level. This is largely determined by the complexity of the system, which dynamically changes the connections and strengths between internal network nodes to better reflect the system structure.

[0173] Figure 17 This is a flow chart of the BP network algorithm in the technical solution of the present invention. The specific process of supervised training and optimization of the neural network model for predicting pre-drilling volcanic rock types is as follows: first, the original variable parameters are standardized in sequence, and the initial weights, thresholds, and learning rates are set. Then, a given sample is input and the input and output values ​​of each layer are calculated through the sigmoid action function. The state of each layer of neurons only affects the state of the neurons in the next layer and gradually propagates to the hidden layer. The input of the hidden layer is equal to the weighted sum of the input layer signals, and the output layer is equal to the output value of the output of the previous hidden layer after being mapped by the activation function, and an output signal is generated at the output end. If the ideal output value cannot be obtained in the output layer, the system enters the error signal back propagation process. The error signal propagates from the output layer to the input layer and adjusts the connection weights between each layer and the bias value of each layer neuron along the way. The error function gradient descent strategy is executed to continuously reduce the error signal, and the weights are continuously adjusted to minimize the network error function. The neural network selects the Tansig excitation function and the trainlm learning algorithm, sets the learning rate to 0.01, the minimum target error to 0.001, and the maximum number of iterations to 1000. Other settings are saved as default settings. The number of hidden layer nodes is 12, the input layer is 5, and the output layer is 1. The following formula is used to calculate the number of hidden layer nodes k:

[0174]

[0175] Where m is the number of hidden layer nodes; n is the number of input layer nodes; l is the number of output layer nodes; α is an integer between 1 and 10, so k is 4 to 13. Different numbers of hidden layer nodes are introduced and simulated training of BP neural network is performed. Finally, the optimal number of hidden layer nodes is 12. At this time, the number of training iterations is small and the result accuracy is the highest.

[0176] In step S3, the input parameter InputData of the training set is set, and the four collected well logging curve data are used as the data for neural network training. The output standard value OutputTarget of the training set is set, and the standard value is the volcanic rock type at the corresponding depth of the block. A BP neural network is established, the number of nodes in the input layer, hidden layer, and output layer is set, the transfer function is set, and the batch gradient descent algorithm is used to perform unconstrained nonlinear optimization on the model.

[0177] To ensure automatic initialization of weights and thresholds during model runtime, the newff function is used as the generator function for the network model. The first step in training a feedforward network is to create a network object, essentially passing it as a parameter to the newff function. This command creates the network object and initializes the network weights and biases, allowing the network to be trained. During training, batch gradient descent with momentum is triggered using the training function traingdm; the hyperbolic tangent function tansig is used as the transfer function between the input and hidden layers; logsigz is used as the transfer function between the hidden and output layers; and the traingdx function is used as the training function for the "back propagation" process.

[0178] In step S3, supervised training using data such as the predicted and standard values ​​of volcanic rock types is performed as follows: based on the gap between the predicted and measured values ​​of volcanic rock types, the backpropagation algorithm updates the weights of the neural network parameters accordingly and modifies the neural network based on the weights. The model prediction value approaches the target value during continuous modification until a certain accuracy rate or number of training times is reached.

[0179] refer to Figure 18 This is the model training optimization curve. As can be seen from the figure, after the neural network model completed 60 training runs, the results were less than 0.001 for 6 consecutive times, and the final maximum system error was 0.000562, which is less than the given accuracy (1×10 -3 ), and successfully established a neural network training model.

[0180] refer to Figure 19 This is a comparison chart of the volcanic rock type prediction results from the validation set. As can be seen, the well logging points, centered around basalt, achieved a high prediction accuracy of 0.957. The well logging points with higher prediction deviations were located in dacite, with an accuracy of 0.855. The well logging points with intermediate prediction deviations were located in tuff, with an accuracy of 0.906. The main causes of prediction deviations may be the geological and logging conditions in different regions, and varying logging interpretations cannot be completely ruled out. Furthermore, the selected samples may also be insufficiently representative. However, the overall prediction accuracy is above 90%, demonstrating the effectiveness of using these well logging data to identify volcanic bodies. It also indicates that the BP neural network is suitable for determining the lithology of volcanic rocks in this region. After training, it has achieved certain practical significance and is suitable for prediction models. To determine the lithology of volcanic rocks, the weights used can be used to obtain practical results in the output layer.

[0181] S4. Input the well logging data of the target depth in the study area. The model automatically identifies the volcanic rock type corresponding to each depth in this area and makes appropriate corrections to the logging lithology of the target layer other than the standard layer based on the well logging curve.

[0182] S5. Using the corrected lithology, sparse pulse wave impedance inversion is used to determine the wave impedance threshold range of different types of volcanic rocks, so as to predict the thickness distribution of igneous rocks before drilling.

[0183] Specifically, the corrected data is imported into Jason software, synthetic record calibration is performed, high-quality seismic wavelets are selected, a low-frequency model is established, and appropriate parameters are input for sparse pulse wave impedance inversion to determine the wave impedance threshold range of different types of volcanic rocks and realize the pre-drilling thickness distribution prediction of different types of igneous rocks.

[0184] Example effect display:

[0185] The prediction accuracy of the well logging points, centered around basalt, was as high as 0.957. The highest prediction deviation rates were observed in dacite, with an accuracy of 0.855. The prediction deviation rates were moderate in tuff, with an accuracy of 0.906. The main causes of the prediction deviations may be due to the different geological and logging conditions in different regions, and different logging interpretations cannot be completely ruled out. Furthermore, the selected samples may not be representative enough. However, the overall prediction accuracy is above 90%, demonstrating the effectiveness of using these well logging data to identify volcanic bodies. It was also noted that the BP neural network is suitable for determining the lithology of volcanic rocks in this region. After training, it has achieved certain practical significance and is suitable for prediction models. The weights used to obtain practical results in the output layer can be used to determine the lithology of volcanic rocks.

[0186] Figure 20 The time thickness distribution diagram of the dacite in the research area obtained by the volcanic rock type identification method provided by the specific embodiment of the present invention is as follows: Figure 20 As can be seen, based on the wave impedance threshold values ​​of dacite and non-dacite, a planar distribution map of dacite thickness predictions was obtained for the study area. The lighter-colored areas in the map indicate areas with thicker dacite. The dacite thickness predictions were relatively good, with the maximum thickness located in the southeast and northeast of the study area, and the thickness of the dacite locally increasing in the western part of the study area. Since data confirm that dacite is prone to leakage, the dacite thickness distribution, combined with logging and seismic data to identify volcanic lithology and assemblages, can provide a reference for well placement and exploration, preventing the occurrence of complex well conditions and contributing to the drilling of high-yield wells in the study area. Previous studies have shown that high-quality tuff reservoirs are visible in some areas. Therefore, the prediction of tuff thickness can provide a basis for identifying favorable reservoirs for drilling.

[0187] The volcanic rock type identification method based on a coupled BP neural network and wave impedance inversion, provided in a specific embodiment of the present invention, uses the predicted volcanic rock type as a basis for determining well location and development, providing drilling technicians and construction personnel with more accurate and effective decision-making. This improves the probability of avoiding complex well conditions and predicting favorable reservoirs, increasing drilling efficiency, avoiding duplication of work, and saving engineering time. It can also quickly and accurately predict pre-drilling lost circulation, providing support for drilling decisions and increasing the probability of avoiding lost circulation and reducing drilling time.

[0188] Example 4:

[0189] Based on the same inventive concept, the embodiments of the present application also provide a volcanic rock type identification device that can be used to implement the methods described in the above embodiments, as shown in the following examples. Since the principles of the volcanic rock type identification device are similar to those of the volcanic rock type identification method, the implementation of the volcanic rock type identification device can be referred to as the implementation of the volcanic rock type identification method, and any repetitions will not be repeated. As used below, the terms "unit" or "module" can refer to a combination of software and / or hardware that implements a predetermined function. Although the systems described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0190] The embodiment of the present invention provides a specific implementation of a volcanic rock type identification device capable of implementing a volcanic rock type identification method, see Figure 21 , a volcanic rock type identification device, comprising:

[0191] A sensitive logging data acquisition module 10 is used to acquire sensitive logging data for characterizing volcanic rock types in a target work area;

[0192] The volcanic rock type identification module 20 is used to identify the volcanic rock type of the target work area based on the sensitive well logging data and a pre-generated volcanic rock type identification model; wherein the volcanic rock type identification model is generated based on a BP neural network and historical data of the sensitive well logging data in the target work area.

[0193] In some embodiments of the present invention, a volcanic rock type identification device further includes:

[0194] a wave impedance threshold determination module, configured to perform sparse pulse wave impedance inversion on the identified volcanic rock type to determine a wave impedance threshold range corresponding to the identified volcanic rock type;

[0195] A distribution feature prediction module is used to predict the distribution feature corresponding to the identified volcanic rock type according to the wave impedance threshold range.

[0196] In some embodiments of the present invention, a volcanic rock type identification device further includes:

[0197] A sensitive data selection module, used for selecting the sensitive well logging data;

[0198] The sensitive data selection module includes:

[0199] An intersection result generating unit, configured to select at least two well logging data from the plurality of well logging data in the target work area for intersection, so as to generate an intersection result;

[0200] A sensitive data selection unit is used to select the sensitive well logging data from the multiple well logging data according to the intersection result.

[0201] In some embodiments of the present invention, a volcanic rock type identification device further includes:

[0202] A model generation module is used to generate a volcanic rock type identification model, and the model generation module includes:

[0203] an initial model generating unit, configured to initialize the initial weight, threshold, and learning rate of the BP neural network to generate an initial model of the volcanic rock type identification model;

[0204] An initial model training unit, configured to train the initial model using the historical data; wherein, during the training of the initial model, the neuron state of the current layer is generated only by the neuron state of the previous layer and propagated layer by layer to the hidden layer; and

[0205] The input signal of the hidden layer is equal to the weighted sum of the input signals of the input layer;

[0206] The output signal of the current output layer is equal to the output value of the previous hidden layer; wherein the output value is generated by mapping the output signal of the previous hidden layer through an activation function.

[0207] In some embodiments of the present invention, when the output signal of the output layer does not meet a preset threshold, a volcanic rock type identification device further includes:

[0208] an error signal input module, configured to input the error signal into the output layer and propagate it toward the input layer;

[0209] The numerical adjustment module is used to adjust the connection weights of each input layer, hidden layer and output layer and the bias value of the neuron during the process of propagation to the input layer.

[0210] In some embodiments of the present invention, the wave impedance threshold determination module includes:

[0211] a seismic data acquisition unit, configured to acquire seismic data of the identified volcanic rock type;

[0212] a record calibration unit, configured to perform synthetic record calibration on the seismic data to generate a calibration result;

[0213] A low-frequency model building unit, configured to build a low-frequency model based on the pre-selected seismic wavelet and the calibration result;

[0214] The wave impedance threshold determination unit is used to determine the wave impedance threshold range according to the low-frequency model.

[0215] In some embodiments of the present invention, a volcanic rock type identification device further includes:

[0216] The type correction module is used to correct the identified volcanic rock type based on the well logging data of the standard layer and the logging data of the bottom layer outside the standard layer.

[0217] As can be seen from the above description, an embodiment of the present invention provides a volcanic rock type identification device that focuses on pre-drilling prediction of volcanic rocks in a study area. Sensitivity curves for identifying volcanic rocks are determined through well logging intersection plot analysis. Pre-processed standard layer sensitivity curve data is used as input, volcanic rock lithology is output, and the actual lithology category is used as the standard value. A volcanic rock prediction neural network model is trained and optimized through supervision. Well logging data at the target depth of the study area is input, and the model automatically identifies the volcanic rock type corresponding to each depth in this area. Appropriate corrections are made to the logged lithology based on the well logging curves. Sparse pulse wave impedance inversion is used using the corrected lithology to determine the wave impedance range of different types of volcanic rocks, enabling pre-drilling prediction of igneous rock thickness distribution. This provides a reference for predicting high-quality volcanic rock reservoirs or pre-drilling drilling engineering and geological design, preventing wellbore instability and well leakage.

[0218] Embodiment 5:

[0219] The embodiments of the present application also provide a specific implementation of an electronic device capable of implementing all steps of a volcanic rock type identification method in the above embodiment, see Figure 22 , electronic equipment specifically includes the following:

[0220] Processor 1201, memory 1202, communications interface 1203, and bus 1204;

[0221] The processor 1201, the memory 1202, and the communication interface 1203 communicate with each other via the bus 1204; the communication interface 1203 is used to implement information transmission between the server device and the client device and other related devices;

[0222] The processor 1201 is configured to call the computer program in the memory 1202. When the processor executes the computer program, all steps of the volcanic rock type identification method in the above embodiment are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0223] Obtain sensitive logging data for characterizing volcanic rock types in the target area;

[0224] The volcanic rock type of the target work area is identified based on the sensitive well logging data and a pre-generated volcanic rock type identification model; wherein the volcanic rock type identification model is generated based on a BP neural network and historical data of the sensitive well logging data in the target work area.

[0225] Example 6:

[0226] The present application also provides a computer-readable storage medium capable of implementing all steps of the volcanic rock type identification method in the above embodiment. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the computer program implements all steps of the volcanic rock type identification method in the above embodiment. For example, when the processor executes the computer program, the following steps are implemented:

[0227] Obtain sensitive logging data for characterizing volcanic rock types in the target area;

[0228] The volcanic rock type of the target work area is identified based on the sensitive well logging data and a pre-generated volcanic rock type identification model; wherein the volcanic rock type identification model is generated based on a BP neural network and historical data of the sensitive well logging data in the target work area.

[0229] The various embodiments in this specification are described in a progressive manner. Similar portions between the various embodiments can be referenced to each other. Each embodiment focuses on the differences between the other embodiments. In particular, the hardware + program embodiments are generally similar to the method embodiments, so their description is relatively simple. For relevant portions, refer to the description of the method embodiments.

[0230] The foregoing description of this specification describes specific embodiments. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0231] Although the present application provides method operation steps such as embodiments or flowcharts, more or fewer operation steps may be included based on conventional or non-creative work. The order of steps listed in the embodiments is only one way of executing the steps among many steps and does not represent the only execution order. When an actual device or client product is executed, it can be executed in the order shown in the embodiments or the drawings or in parallel (for example, in a parallel processor or multi-threaded processing environment).

[0232] For the convenience of description, the above devices are described in terms of functions divided into various modules. Of course, when implementing the embodiments of this specification, the functions of each module can be implemented in the same or multiple software and / or hardware, or the modules that implement the same function can be implemented by a combination of multiple sub-modules or sub-units, etc. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0233] Those skilled in the art will also appreciate that, in addition to implementing the controller in pure computer-readable program code, it is entirely possible to implement the same functionality by logically programming the method steps in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, embedded microcontrollers, and the like. Therefore, such a controller can be considered a hardware component, and the devices included therein for implementing various functions can also be considered structures within the hardware component. Alternatively, the devices for implementing various functions can be considered both software modules implementing the method and structures within the hardware component.

[0234] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0235] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0236] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between the various embodiments can be referenced across them. Each embodiment focuses on the differences from the other embodiments. In particular, since the system embodiments are generally similar to the method embodiments, their description is relatively simple. For relevant parts, reference can be made to the description of the method embodiments. Throughout this specification, reference to the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the embodiments in this specification. In this specification, the schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, those skilled in the art may combine and integrate the different embodiments or examples, and features of different embodiments or examples, described in this specification, without conflict.

[0237] The above description is merely an example of the embodiments of this specification and is not intended to limit the embodiments of this specification. For those skilled in the art, various modifications and variations of the embodiments of this specification are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of this specification shall be included within the scope of the claims of the embodiments of this specification.

Claims

1. A method for identifying volcanic rock types, characterized in that: include: Obtain sensitive logging data for characterizing volcanic rock types in the target area; The volcanic rock type of the target work area is identified based on the sensitive well logging data and a pre-generated volcanic rock type identification model; wherein the volcanic rock type identification model is generated based on a BP neural network and historical data of the sensitive well logging data in the target work area.

2. The volcanic rock type identification method according to claim 1, characterized in that: Also includes: performing sparse pulse wave impedance inversion on the identified volcanic rock type to determine a wave impedance threshold range corresponding to the identified volcanic rock type; The distribution characteristics corresponding to the identified volcanic rock type are predicted according to the wave impedance threshold range.

3. The volcanic rock type identification method according to claim 1, characterized in that: Also includes: selecting the sensitive well logging data; Selecting the sensitive logging data includes: Selecting at least two well logging data from a plurality of well logging data in the target work area for intersection to generate an intersection result; The sensitive well logging data is selected from the plurality of well logging data according to the intersection result.

4. The volcanic rock type identification method according to any one of claims 1 to 3, characterized in that: The steps to generate a volcanic rock type identification model include: Initializing the initial weight, threshold, and learning rate of the BP neural network to generate an initial model of the volcanic rock type identification model; The initial model is trained using the historical data; wherein, during the training of the initial model, the neuron state of the current layer is generated only by the neuron state of the previous layer, and is propagated layer by layer to the hidden layer; and The input signal of the hidden layer is equal to the weighted sum of the input signals of the input layer; The output signal of the current output layer is equal to the output value of the previous hidden layer; wherein the output value is generated by mapping the output signal of the previous hidden layer through an activation function.

5. The volcanic rock type identification method according to claim 4, characterized in that: When the output signal of the output layer does not meet the preset threshold, the method further includes: Inputting the error signal into the output layer and propagating it toward the input layer; During the propagation to the input layer, the connection weights of each input layer, hidden layer and output layer and the bias values ​​of the neurons are adjusted.

6. The volcanic rock type identification method according to claim 2, characterized in that: The performing sparse pulse wave impedance inversion on the identified volcanic rock type to determine a wave impedance threshold range corresponding to the identified volcanic rock type includes: obtaining seismic data of the identified volcanic rock type; performing synthetic record calibration on the seismic data to generate a calibration result; Establishing a low-frequency model based on the pre-selected seismic wavelet and the calibration result; The wave impedance threshold range is determined according to the low-frequency model.

7. The volcanic rock type identification method according to claim 2, characterized in that: Before performing sparse pulse wave impedance inversion on the identified volcanic rock type to determine the wave impedance threshold range corresponding to the identified volcanic rock type, the method further includes: The identified volcanic rock types are corrected based on the well logging data of the standard layer and the logging data of the bottom layer outside the standard layer.

8. A volcanic rock type identification device, characterized in that: include: A sensitive logging data acquisition module is used to obtain sensitive logging data for characterizing volcanic rock types in the target work area; A volcanic rock type identification module is used to identify the volcanic rock type of the target work area based on the sensitive well logging data and a pre-generated volcanic rock type identification model; wherein the volcanic rock type identification model is generated based on a BP neural network and historical data of the sensitive well logging data in the target work area.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the volcanic rock type identification method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the volcanic rock type identification method according to any one of claims 1 to 7 are implemented.

Images