CN113406695A - Seismic inversion method and system based on interval velocity seismic geological model - Google Patents
Seismic inversion method and system based on interval velocity seismic geological model Download PDFInfo
- Publication number
- CN113406695A CN113406695A CN202110604309.1A CN202110604309A CN113406695A CN 113406695 A CN113406695 A CN 113406695A CN 202110604309 A CN202110604309 A CN 202110604309A CN 113406695 A CN113406695 A CN 113406695A
- Authority
- CN
- China
- Prior art keywords
- seismic
- velocity
- inversion
- model
- layer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
- G01V1/28—Processing seismic data, e.g. analysis, for interpretation, for correction
- G01V1/282—Application of seismic models, synthetic seismograms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The invention relates to a seismic inversion method and a system based on a layer velocity seismic geological model, which comprises the steps of constructing the layer velocity seismic geological model; carrying out seismic forward modeling on the layer velocity seismic geological model to obtain a seismic record and an original layer velocity set; based on the interval velocity seismic geological model, carrying out seismic inversion simulation to obtain an inversion result, extracting a target interval velocity set from the inversion result, and taking the target interval velocity set as a label; constructing a training set data set based on the seismic records and the labels; inputting the training data set into a preset full convolution neural network for migration learning to obtain a layer velocity label inversion initial model, performing self-checking label iterative training on the layer velocity label inversion initial model by using the seismic record and the original layer velocity set to obtain a layer velocity label inversion target model, and performing seismic inversion simulation on the layer velocity label inversion target model. The high-precision seismic interval velocity inversion method is easier for label inversion.
Description
Technical Field
The application relates to the technical field of seismic exploration, in particular to a seismic inversion method and system based on a layer velocity seismic geological model.
Background
The interval velocity inversion is an important technology for providing technical support for reservoir prediction of oil field exploration and development by using technical means such as earthquake, well logging, geology and the like, and the seismic inversion modeling precision is improved, so that the reliability of seismic reservoir prediction is improved. At present, in oil field exploration and development, the determination of a geological model is based on well logging and earthquake, the geological model is continuously updated according to experience and development practice, and an inversion model is further modified and perfected, so that the multi-solution of the geological model is reduced, and the multi-solution of an inversion result is further reduced.
Because geological knowledge in the layer velocity model construction is highly dependent on expert experience, the acquisition difficulty is high, the model quality is dependent on known logging data and seismic data, and the fine layer velocity model is dependent on high-precision seismic small layer interpretation, the interpretation of the seismic small layer is limited, and the problems can limit samples for deep learning and seriously affect the effect of training an inversion model.
Seismic interval velocity inversion models often need a large amount of correct marked data and mainly depend on drilling data of a research area, interval velocity geological models interpolated by the logging data of the research area are not enough to represent underground real conditions, the acquisition of the geological models also depends on the mastering conditions of various data, and part of data can be polluted due to errors caused by factors such as instruments. In addition, the fitting problem can be caused by the density degree of well data, the space popularization of a local reliable geological model is limited, and the existing deep learning model is difficult to express a data set, so that the model is poor in performance on a test set, and the problem cannot be well solved by a common label acquisition method.
Disclosure of Invention
In order to solve the problem of label acquisition during the training of the conventional seismic interval velocity inversion model, the application provides a seismic inversion method and a seismic inversion system based on an interval velocity seismic geological model.
In a first aspect, the present application provides a seismic inversion method based on a interval velocity seismic geological model, the method comprising:
constructing a layer velocity seismic geological model;
performing seismic forward modeling on the layer velocity seismic geological model to obtain a seismic record and an original layer velocity set;
based on the interval velocity seismic geological model, carrying out seismic inversion simulation to obtain an inversion result, obtaining a target interval velocity set from the inversion result, and taking the target interval velocity set as a tag data set;
constructing a training set data set based on the seismic records and the tag data set;
inputting the training data set into a preset full convolution neural network for migration learning to obtain a layer velocity label inversion initial model, and performing label self-checking iterative training on the layer velocity label inversion initial model by using the seismic record and the original layer velocity set to obtain a layer velocity label inversion target model;
and performing seismic inversion simulation on the target model based on the interval velocity label inversion.
Optionally, the constructing a layer velocity seismic geological model includes:
acquiring logging information of a work area according to a logging sublayer division method;
according to the logging information, a geological frame model is constructed, and a single-well initial layer speed set is obtained;
and acquiring a layer velocity seismic geologic model by using a velocity space topological method based on the geologic frame model and the single-well initial layer velocity set.
Optionally, the logging information includes logging horizon information and acoustic logging curves, according to the logging information, a geologic framework model is constructed, and a single-well-horizon velocity set is obtained, including:
constructing a geological frame model according to the logging horizon information;
according to the acoustic logging curve, the inter-layer speed is obtained, and a single-well initial layer speed set is formed;
based on the geological frame model and the single-well interval velocity set, a velocity space topological method is utilized to obtain an interval velocity seismic geological model, which comprises the following steps:
carrying out earthquake forward modeling based on the single-well initial layer velocity set to obtain single-well earthquake records;
extracting geological reflection coefficient data and a single-well initial layer velocity set from the single-well seismic records;
and carrying out speed space topology on the geological frame model based on geological reflection coefficient data and the single-well initial layer speed set to obtain a layer speed seismic geological model.
Optionally, after performing seismic forward modeling based on the single-well interval velocity set to obtain a single-well seismic record, the method further includes:
and performing Hilbert transform on the single-well seismic records to obtain a phase section, and performing seismic small-layer interpretation on the geological framework model by using the phase section.
Optionally, performing a seismic forward modeling on the interval velocity seismic geological model to obtain a seismic record and an original interval velocity set, including:
and under the condition of changing the phase and frequency of the seismic wavelets, performing seismic forward modeling on the layer velocity seismic geological model to obtain a seismic record, wherein the seismic record comprises a single-well seismic record, extracting a single-well initial velocity and an initial layer velocity set constrained by seismic logging from the seismic record, and taking the single-well initial velocity and the initial layer velocity set constrained by the seismic logging as an initial layer velocity set.
Optionally, the target interval velocity set includes an extracted interval velocity set and a seismic inversion interval velocity set, and the extracted interval velocity set includes a single-well extracted interval velocity set and a seismic logging constrained extracted interval velocity set.
Optionally, inputting the training data set to a preset full convolution neural network for migration learning, to obtain a layer velocity label inversion initial model, including:
the full convolution neural network comprises three convolution layers, three deconvolution layers, three dropout layers, two pooling layers, six relu activation layers and one full convolution layer;
and training the interval velocity of the seismic inversion on a training set to obtain an interval velocity label inversion initial model.
Optionally, performing label self-checking iterative training on the interval velocity label inversion initial model by using the seismic record and the original interval velocity set, and obtaining an interval velocity label inversion target model, including:
inputting the seismic record and the original layer velocity set into the layer velocity tag inversion initial model, and outputting the layer velocity tag inversion initial model as a layer velocity prediction set corresponding to the original layer velocity set;
comparing the layer velocity prediction set with the original layer velocity set, correcting the data with inconsistent labels in the layer velocity prediction set by using seismic record calibration, and performing iterative training until the layer velocity prediction set is consistent with the original layer velocity set.
In a second aspect, the present application provides a seismic inversion system based on a interval velocity seismic geological model, the system comprising:
the model building module is used for building a layer velocity seismic geological model;
the forward modeling module is used for carrying out seismic forward modeling on the interval velocity seismic geological model to obtain a seismic record and an original interval velocity set;
the inversion module is used for developing seismic inversion simulation based on the interval velocity seismic geological model, obtaining inversion results, and extracting a target interval velocity set from the inversion results, wherein the target interval velocity set comprises a single-well extraction interval velocity set, a seismic logging constrained extraction interval velocity set and a seismic inversion interval velocity set; taking the target layer velocity set as a label;
the data construction module is used for constructing a training set data set based on the seismic records and the labels;
the training module is used for inputting the training data set to a preset full convolution neural network for migration learning to obtain a layer velocity label inversion initial model, performing self-checking label iterative training on the layer velocity label inversion initial model by using the seismic record and the original layer velocity set, and finishing the training to obtain a layer velocity label inversion target model;
and the tag inversion model is used for inverting the target model to perform seismic inversion simulation based on the interval velocity tag.
In a third aspect, the present application provides a computer device, which adopts the following technical solution:
a computer apparatus comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the method for seismic inversion based on a interval velocity seismic geological model when executing the computer program.
In a fourth aspect, the present application provides a computer-readable storage medium, which adopts the following technical solutions:
a computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the method for seismic inversion based on a interval velocity seismic geological model.
The application has the following beneficial technical effects:
the method comprises the steps of constructing a layer velocity seismic geological model, carrying out forward modeling on the model to obtain a seismic record and an original layer velocity set, and carrying out subsequent training and label prediction; then, a target layer velocity set is obtained through inversion simulation, and the target layer velocity set is used as a tag data set, so that a high-resolution layer velocity tag inversion model can be conveniently constructed subsequently; the method comprises the steps of constructing a training data set based on seismic records and a label data set, inputting the training data set into a full convolution neural network for supervised transfer learning to obtain a layer velocity label inversion initial model, and performing frequent label self-checking iterative training on the layer velocity label inversion initial model by using the seismic records and the original layer velocity set to obtain a layer velocity label inversion target model, so that the construction and prediction of the layer velocity label inversion target model are self-consistent by a deep learning method, the limitation of production dynamic check of a geological model and hypothesis inference which are purely dependent on expert experience and geological cognition can be avoided, and the high-precision seismic layer velocity inversion method based on the geological model and not fully dependent on the geological model is easy to obtain labels, low in training difficulty, economical, reliable and high in efficiency.
Drawings
FIG. 1 is a flow chart of a method of an embodiment of a method for seismic inversion based on a interval velocity seismic geological model provided by the present invention;
FIG. 2 is a schematic block diagram of an embodiment of a seismic inversion system based on a interval velocity seismic geological model provided by the invention.
Detailed Description
The present application is described in further detail below with reference to figures 1-2.
The embodiment of the application discloses a seismic inversion method based on a layer velocity seismic geological model, and with reference to fig. 1, the method comprises the following steps:
and S1, constructing a layer velocity seismic geological model.
Specifically, logging information of a work area is obtained according to a logging sublayer division method; according to the logging information, a geological frame model is constructed, and a single-well initial layer speed set is obtained; and acquiring a layer velocity seismic geologic model by using a velocity space topological method based on the geologic frame model and the single-well initial layer velocity set.
The method includes the steps that logging sub-layer division is conducted on a work area, logging information is obtained, and the logging information comprises work area logging position information and an acoustic wave logging curve; further, according to the logging horizon information, a one-dimensional geological frame model of a logging work area is constructed; the method comprises the following steps of obtaining the layer speed between layers according to an acoustic logging curve to form a single-well initial layer speed set, wherein a specific calculation formula is as follows:
τkis the average value of the k-th layer propagation time, VkIs the layer velocity of the k-th layer.
Further, based on the single-well initial layer velocity set, carrying out earthquake forward modeling to obtain a single-well earthquake record, wherein the specific calculation formula is as follows: s (t) ═ w (t) × r (t), where s (t) denotes forward recordings and w (t) denotes wavelets; r (t) represents the geological reflection coefficient. In the embodiment, Hilbert transform can be performed on single-well seismic records to obtain a phase section, and the geological frame model is subjected to seismic small-layer interpretation by using the phase section, wherein the seismic small-layer interpretation is based on logging calibration and seismic wavelet phase characteristics, so that the seismic frame model with complete data can be conveniently obtained. And further, extracting geological reflection coefficient data and a single-well initial layer velocity set from the single-well seismic records, and providing a layer velocity space topological structure for the seismic frame model by the geological reflection coefficient data and the single-well initial layer velocity set so as to obtain a fine layer velocity seismic geological model.
And S2, performing seismic forward modeling on the layer velocity seismic geological model to obtain a seismic record and an original layer velocity set.
Specifically, under the condition of changing the phase and frequency of seismic wavelets, performing seismic forward modeling on a layer velocity seismic geological model to obtain a seismic record, wherein the seismic record has larger volume compared with the single well, a single-well initial velocity and an initial layer velocity set constrained by seismic logging are extracted from the seismic record, and the single-well initial velocity and the initial layer velocity set constrained by the seismic logging are used as the initial layer velocity set.
S3, carrying out seismic inversion simulation based on the interval velocity seismic geological model, obtaining inversion results, obtaining a target interval velocity set from the inversion results, and taking the target interval velocity set as a tag data set.
In this embodiment, inversion simulation is performed on the layer velocity seismic geological model by using logging horizon interpolation control to obtain an inversion result, and a target layer velocity set is obtained from the inversion result, where the target layer velocity set includes an extraction layer velocity set and a seismic inversion layer velocity set, and the extraction layer velocity set includes a single-well extraction layer velocity set and a seismic logging constrained extraction layer velocity set.
And S4, constructing a training set data set based on the seismic records and the label data set.
Specifically, the label data set includes the single-well extraction interval velocity set, the extraction interval velocity set constrained by seismic logging and the interval velocity set of seismic inversion, in this embodiment, the single-well extraction interval velocity set can be divided into a training set, a verification set and a test set, and the data volume proportion of the three is 1: 1: 1; the seismic logging constrained extraction interval velocity set is divided into a training set, a verification set and a testing set, and the data volume proportion of the training set, the verification set and the testing set is 2: 1: 1; dividing a seismic inversion interval velocity set into a training set, a verification set and a test set, wherein the data volume proportion of the training set, the verification set and the test set is 2: 1: 1; the final data of the single-well extraction interval velocity set, the seismic logging constrained extraction interval velocity set and the seismic inversion interval velocity set are divided into a training set, a verification set and a test set, and the data volume proportion of the three sets is 5: 3: 3. in this embodiment, for the problem of unbalanced data samples, data enhancement may be performed on a target data set by using random methods, or more samples may be generated by using a method of changing the phase and frequency of seismic wavelets, so as to improve the application effect of a small data scene.
S5, inputting the training data set to a preset full convolution neural network for migration learning to obtain a layer velocity label inversion initial model, and performing label self-checking iterative training on the layer velocity label inversion initial model by using the seismic record and the original layer velocity set to obtain a layer velocity label inversion target model.
In this embodiment, the full convolution neural network includes three convolution layers, three deconvolution layers, three dropout layers, two pooling layers, six relu activation layers, and one full convolution layer; specifically, a FCN _ inv _ d for image pixel set classification is adopted, and the last full connection layer of the FCN _ inv _ d network is removed, so that the FCN _ inv network is obtained. It should be noted that the purpose of building the convolutional layer is to extract data features, and the convolutional layer contains relative position information in addition to feature values; the purpose of the pooling layer is to compress convolutional layer information, reduce the number of feature-mapped neural units to reduce data dimensionality; the purpose of the Dropout layer is to improve the generalization capability of the model, perform overfitting on a complex data set and improve the calculation efficiency; the main purpose of the activation function is to construct a non-linear mapping relationship to establish the relationship between the input and the output.
Further, training a minimum loss function on the training set according to the interval velocity of seismic inversion to obtain an interval velocity label inversion initial model; in this embodiment, a gradient descent technique is used to train the model parameters, and a set of model parameters with relatively small loss functions is obtained.
Further, inputting the seismic record and the original layer velocity set into a layer velocity label inversion initial model, and outputting the layer velocity label inversion initial model as a layer velocity prediction set corresponding to the original layer velocity set; comparing the layer velocity prediction set with the original layer velocity set, correcting data with inconsistent labels in the layer velocity prediction set by using seismic record calibration, and performing iterative training until the layer velocity prediction set is consistent with the original layer velocity set, and ending the training to obtain a layer velocity label inversion target model.
It should be noted that after the target model inverted by the layer velocity label is trained and evaluated for several times in the training set and the verification set, the target model inverted by the layer velocity label is evaluated in the test set.
And S6, inverting the target model based on the interval velocity label to perform seismic inversion simulation.
It should be noted that the trained interval velocity label is used to invert the target model, and the real-time seismic record is used to perform seismic inversion simulation, so as to obtain the accurate interval velocity.
In the embodiment, a layer velocity seismic geological model is constructed, a seismic record and an original layer velocity set are obtained through forward modeling on the basis of the model, and the model is used for subsequent training and label prediction; then, a target layer velocity set is obtained through inversion simulation, and the target layer velocity set is used as a tag data set, so that a high-resolution layer velocity tag inversion model can be conveniently constructed subsequently; the method comprises the steps of constructing a training data set based on seismic records and a label data set, inputting the training data set into a full convolution neural network for supervised transfer learning to obtain a layer velocity label inversion initial model, and performing frequent label self-checking iterative training on the layer velocity label inversion initial model by using the seismic records and the original layer velocity set to obtain a layer velocity label inversion target model, so that the construction and prediction of the layer velocity label inversion target model are self-consistent by a deep learning method, the limitation of production dynamic check of a geological model and hypothesis inference which are purely dependent on expert experience and geological cognition can be avoided, and a high-precision seismic layer velocity inversion method based on the geological model and not completely dependent on the geological model has low training difficulty, is economic and reliable, and has high efficiency.
It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.
The embodiment also provides a seismic inversion system based on the interval velocity seismic geological model, and the seismic inversion system based on the interval velocity seismic geological model corresponds to the seismic inversion method based on the interval velocity seismic geological model in the embodiment one to one. As shown in fig. 2, the seismic inversion system based on the interval velocity seismic geologic model comprises a model building module 201, a forward modeling module 202, an inversion module 203, a data building module 204, a training module 205 and a tag inversion module 206. The functional modules are explained in detail as follows:
a construction model module 201, configured to construct a layer velocity seismic geological model;
the forward modeling module 202 is configured to perform seismic forward modeling on the layer velocity seismic geological model to obtain a seismic record and an original layer velocity set;
the inversion module 203 is used for developing seismic inversion simulation based on the interval velocity seismic geological model, obtaining an inversion result, extracting a target interval velocity set from the inversion result, and taking the target interval velocity set as a label;
the data construction module 204 is used for constructing a training set data set based on the seismic records and the labels;
the training module 205 is configured to input a training data set to a preset full convolution neural network for migration learning, obtain a layer velocity label inversion initial model, perform self-checking label iterative training on the layer velocity label inversion initial model by using the seismic record and the original layer velocity set, and end the training to obtain a layer velocity label inversion target model;
and the tag inversion module 206 is used for inverting the target model based on the interval velocity tag to perform seismic inversion simulation.
For specific limitations of the seismic inversion system based on the interval velocity seismic geological model, reference may be made to the above limitations of the seismic inversion method based on the interval velocity seismic geological model, and details are not repeated here. The various modules in the above-described seismic inversion system based on the interval velocity seismic geological model may be implemented in whole or in part by software, hardware, and combinations thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
The embodiment also provides a computer device, which can be a server. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing seismic records, tag datasets, interval velocity tag inversion target models, and the like. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a seismic inversion method based on a interval velocity seismic geological model, and the processor executes the computer program to implement the following steps:
constructing a layer velocity seismic geological model; carrying out seismic forward modeling on the layer velocity seismic geological model to obtain a seismic record and an original layer velocity set; based on the interval velocity seismic geological model, carrying out seismic inversion simulation to obtain an inversion result, obtaining a target interval velocity set from the inversion result, and taking the target interval velocity set as a tag data set; constructing a training set data set based on the seismic records and the label data set; inputting the training data set into a preset full convolution neural network for migration learning to obtain a layer velocity label inversion initial model, and performing label self-checking iterative training on the layer velocity label inversion initial model by using the seismic record and the original layer velocity set to obtain a layer velocity label inversion target model; and carrying out seismic inversion simulation on the target model based on the interval velocity label inversion.
The present embodiments also provide a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of:
constructing a layer velocity seismic geological model; carrying out seismic forward modeling on the layer velocity seismic geological model to obtain a seismic record and an original layer velocity set; based on the interval velocity seismic geological model, carrying out seismic inversion simulation to obtain an inversion result, obtaining a target interval velocity set from the inversion result, and taking the target interval velocity set as a tag data set; constructing a training set data set based on the seismic records and the label data set; inputting the training data set into a preset full convolution neural network for migration learning to obtain a layer velocity label inversion initial model, and performing label self-checking iterative training on the layer velocity label inversion initial model by using the seismic record and the original layer velocity set to obtain a layer velocity label inversion target model; and carrying out seismic inversion simulation on the target model based on the interval velocity label inversion.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-mentioned functions. The above embodiments are preferred embodiments of the present application, and the protection scope of the present application is not limited by the above embodiments, so: all equivalent changes made according to the structure, shape and principle of the present application shall be covered by the protection scope of the present application.
Claims (10)
1. A seismic inversion method based on a layer velocity seismic geological model is characterized by comprising the following steps: the method comprises the following steps:
constructing a layer velocity seismic geological model;
performing seismic forward modeling on the layer velocity seismic geological model to obtain a seismic record and an original layer velocity set;
based on the interval velocity seismic geological model, carrying out seismic inversion simulation to obtain an inversion result, obtaining a target interval velocity set from the inversion result, and taking the target interval velocity set as a tag data set;
constructing a training set data set based on the seismic records and the tag data set;
inputting the training data set into a preset full convolution neural network for migration learning to obtain a layer velocity label inversion initial model, and performing label self-checking iterative training on the layer velocity label inversion initial model by using the seismic record and the original layer velocity set to obtain a layer velocity label inversion target model;
and performing seismic inversion simulation on the target model based on the interval velocity label inversion.
2. The seismic inversion method based on the interval velocity seismic geological model according to claim 1, characterized in that: the constructing of the interval velocity seismic geological model comprises the following steps:
acquiring logging information of a work area according to a logging sublayer division method;
according to the logging information, a geological frame model is constructed, and a single-well initial layer speed set is obtained;
and acquiring a layer velocity seismic geologic model by using a velocity space topological method based on the geologic frame model and the single-well initial layer velocity set.
3. The seismic inversion method based on the interval velocity seismic geological model according to claim 1, characterized in that: the logging information includes logging horizon information and acoustic logging curve, according to the logging information, constructs the geological framework model to obtain single well way layer speed set, include:
constructing a geological frame model according to the logging horizon information;
according to the acoustic logging curve, the inter-layer speed is obtained, and a single-well initial layer speed set is formed;
based on the geological frame model and the single-well interval velocity set, a velocity space topological method is utilized to obtain an interval velocity seismic geological model, which comprises the following steps:
carrying out earthquake forward modeling based on the single-well initial layer velocity set to obtain single-well earthquake records;
extracting geological reflection coefficient data and a single-well initial layer velocity set from the single-well seismic records;
and carrying out speed space topology on the geological frame model based on geological reflection coefficient data and the single-well initial layer speed set to obtain a layer speed seismic geological model.
4. The seismic inversion method based on the interval velocity seismic geological model according to claim 2, characterized in that: based on the single-well layer velocity set, carrying out earthquake forward modeling, and after obtaining the single-well earthquake record, the method further comprises the following steps:
and performing Hilbert transform on the single-well seismic records to obtain a phase section, and performing seismic small-layer interpretation on the geological framework model by using the phase section.
5. The seismic inversion method based on the interval velocity seismic geological model according to claim 3, characterized in that: and carrying out seismic forward modeling on the interval velocity seismic geological model to obtain a seismic record and an original interval velocity set, wherein the method comprises the following steps:
and under the condition of changing the phase and frequency of the seismic wavelets, performing seismic forward modeling on the layer velocity seismic geological model to obtain a seismic record, wherein the seismic record comprises a single-well seismic record, extracting a single-well initial velocity and an initial layer velocity set constrained by seismic logging from the seismic record, and taking the single-well initial velocity and the initial layer velocity set constrained by the seismic logging as an initial layer velocity set.
6. The seismic inversion method based on the interval velocity seismic geological model according to claim 1, characterized in that: the target interval velocity set comprises an extraction interval velocity set and a seismic inversion interval velocity set, and the extraction interval velocity set comprises a single-well extraction interval velocity set and a seismic logging constrained extraction interval velocity set.
7. The seismic inversion method based on the interval velocity seismic geological model according to claim 1, characterized in that: inputting the training data set into a preset full convolution neural network for migration learning to obtain a layer velocity label inversion initial model, wherein the method comprises the following steps:
the full convolution neural network comprises three convolution layers, three deconvolution layers, three dropout layers, two pooling layers, six relu activation layers and one full convolution layer;
and training the interval velocity of the seismic inversion on a training set to obtain an interval velocity label inversion initial model.
8. The seismic inversion method based on the interval velocity seismic geological model according to claim 3, characterized in that: performing label self-checking iterative training on the interval velocity label inversion initial model by using the seismic record and the original interval velocity set to obtain an interval velocity label inversion target model, wherein the method comprises the following steps:
inputting the seismic record and the original layer velocity set into the layer velocity tag inversion initial model, and outputting the layer velocity tag inversion initial model as a layer velocity prediction set corresponding to the original layer velocity set;
comparing the layer velocity prediction set with the original layer velocity set, correcting the data with inconsistent labels in the layer velocity prediction set by using seismic record calibration, and performing iterative training until the layer velocity prediction set is consistent with the original layer velocity set.
9. A seismic inversion system based on a layer velocity seismic geological model is characterized in that: the system comprises:
the model building module is used for building a layer velocity seismic geological model;
the forward modeling module is used for carrying out seismic forward modeling on the interval velocity seismic geological model to obtain a seismic record and an original interval velocity set;
the inversion module is used for developing seismic inversion simulation based on the interval velocity seismic geological model, obtaining inversion results, and extracting a target interval velocity set from the inversion results, wherein the target interval velocity set comprises a single-well extraction interval velocity set, a seismic logging constrained extraction interval velocity set and a seismic inversion interval velocity set; taking the target layer velocity set as a label;
the data construction module is used for constructing a training set data set based on the seismic records and the labels;
the training module is used for inputting the training data set to a preset full convolution neural network for migration learning to obtain a layer velocity label inversion initial model, performing self-checking label iterative training on the layer velocity label inversion initial model by using the seismic record and the original layer velocity set, and finishing the training to obtain a layer velocity label inversion target model;
and the tag inversion module is used for inverting the target model based on the interval velocity tag to perform seismic inversion simulation.
10. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor when executing the computer program implements the steps of the method for seismic inversion based on a interval velocity seismic geological model according to any of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110604309.1A CN113406695B (en) | 2021-05-31 | 2021-05-31 | Seismic inversion method and system based on interval velocity seismic geological model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110604309.1A CN113406695B (en) | 2021-05-31 | 2021-05-31 | Seismic inversion method and system based on interval velocity seismic geological model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113406695A true CN113406695A (en) | 2021-09-17 |
| CN113406695B CN113406695B (en) | 2022-12-13 |
Family
ID=77675539
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110604309.1A Active CN113406695B (en) | 2021-05-31 | 2021-05-31 | Seismic inversion method and system based on interval velocity seismic geological model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113406695B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114706119A (en) * | 2022-04-06 | 2022-07-05 | 中国科学院地质与地球物理研究所 | Reflection waveform inversion method and system based on convolutional neural network deep learning |
| CN115576011A (en) * | 2022-09-30 | 2023-01-06 | 北京福瑞宝能源科技有限公司 | Phased reservoir prediction method, device and system based on geological model |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190064389A1 (en) * | 2017-08-25 | 2019-02-28 | Huseyin Denli | Geophysical Inversion with Convolutional Neural Networks |
| CN111239802A (en) * | 2020-01-19 | 2020-06-05 | 中国海洋大学 | Deep learning speed modeling method based on seismic reflection waveform and velocity spectrum |
| CN112213771A (en) * | 2019-07-10 | 2021-01-12 | 中国石油天然气股份有限公司 | Seismic wave impedance inversion method and device |
| US20210063591A1 (en) * | 2019-08-29 | 2021-03-04 | Advanced Geophysical Technology Inc. | Methods and systems for obtaining reconstructed low-frequency seismic data for determining a subsurface feature |
| CN112444850A (en) * | 2019-08-29 | 2021-03-05 | 中国石油化工股份有限公司 | Seismic data velocity modeling method, storage medium, and computing device |
-
2021
- 2021-05-31 CN CN202110604309.1A patent/CN113406695B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190064389A1 (en) * | 2017-08-25 | 2019-02-28 | Huseyin Denli | Geophysical Inversion with Convolutional Neural Networks |
| CN112213771A (en) * | 2019-07-10 | 2021-01-12 | 中国石油天然气股份有限公司 | Seismic wave impedance inversion method and device |
| US20210063591A1 (en) * | 2019-08-29 | 2021-03-04 | Advanced Geophysical Technology Inc. | Methods and systems for obtaining reconstructed low-frequency seismic data for determining a subsurface feature |
| CN112444850A (en) * | 2019-08-29 | 2021-03-05 | 中国石油化工股份有限公司 | Seismic data velocity modeling method, storage medium, and computing device |
| CN111239802A (en) * | 2020-01-19 | 2020-06-05 | 中国海洋大学 | Deep learning speed modeling method based on seismic reflection waveform and velocity spectrum |
Non-Patent Citations (1)
| Title |
|---|
| 张兵: "基于卷积神经网络和叠加速度谱的地震层速度自动建模方法", 《石油物探》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114706119A (en) * | 2022-04-06 | 2022-07-05 | 中国科学院地质与地球物理研究所 | Reflection waveform inversion method and system based on convolutional neural network deep learning |
| CN114706119B (en) * | 2022-04-06 | 2022-09-02 | 中国科学院地质与地球物理研究所 | Reflection waveform inversion method and system based on convolutional neural network deep learning |
| CN115576011A (en) * | 2022-09-30 | 2023-01-06 | 北京福瑞宝能源科技有限公司 | Phased reservoir prediction method, device and system based on geological model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113406695B (en) | 2022-12-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113406695B (en) | Seismic inversion method and system based on interval velocity seismic geological model | |
| CN114492211B (en) | Residual oil distribution prediction method based on autoregressive network model | |
| CN114154427B (en) | Volume fracturing fracture expansion prediction method and system based on deep learning | |
| CN113109874B (en) | Wave impedance inversion method using neural network and neural network system | |
| CN109613623B (en) | Lithology prediction method based on residual error network | |
| CN113919219A (en) | Stratum evaluation method and system based on logging big data | |
| CN112213771A (en) | Seismic wave impedance inversion method and device | |
| CN112861721B (en) | Method and device for automatically extracting background noise dispersion curve | |
| CN114492213A (en) | Wavelet neural operator network model-based residual oil saturation and pressure prediction method | |
| Kumar et al. | Ensemble-based assimilation of nonlinearly related dynamic data in reservoir models exhibiting non-Gaussian characteristics | |
| CN116047583A (en) | Adaptive wave impedance inversion method and system based on depth convolution neural network | |
| CN115455828A (en) | Interpretable well logging curve completion method | |
| CN111027249B (en) | Machine learning-based inter-well connectivity evaluation method | |
| CN113642675B (en) | Underground rock stratum distribution imaging acquisition method, system, terminal and readable storage medium based on full waveform inversion and convolutional neural network | |
| CN113325480A (en) | Seismic lithology identification method based on integrated deep learning | |
| Razak et al. | Embedding physical flow functions into deep learning predictive models for improved production forecasting | |
| CN113625336A (en) | Seismic wave impedance thin layer inversion method based on full convolution neural network | |
| CN112444850B (en) | Seismic data velocity modeling method, storage medium and computing device | |
| CN113608258B (en) | Self-consistent deep learning method for constructing high-resolution wave impedance inversion label | |
| CN115079257A (en) | Q value estimation and seismic attenuation compensation method based on fusion network | |
| CN111949945B (en) | Method, device, equipment and system for determining total organic carbon content of source rock | |
| CN115130529A (en) | Pre-stack seismic inversion method and device | |
| CN113296150A (en) | High-dimensional closed-loop network seismic inversion method under logging constraint | |
| CN116430448B (en) | Inversion method and device of seismic data and electronic equipment | |
| CN116299690A (en) | Fault detection method, device, equipment and medium based on deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |