Seismic intelligent inversion method and system driven by geological modeling and seismic forward modeling

Abstract

The present invention relates to a seismic intelligent inversion method and system driven by geological modeling and seismic forward modeling, which comprises: taking a region without wells or few wells as a study area, selecting a sedimentary evolution mode that conforms to the sedimentary phase type of the study area based on the sedimentary phase type of the study area; simulating the deposition and evolution process of the geological body based on the sedimentary evolution mode, establishing a three-dimensional sedimentary phase model of the study area, and constructing a three-dimensional wave impedance model based on the three-dimensional sedimentary phase model, performing seismic forward modeling simulation, and obtaining forward seismic data; setting a virtual well, performing deep learning with the forward seismic data body, obtaining a tested learning model, inputting the actual seismic data body into the tested learning model, solving the actual seismic data body, and outputting a seismic intelligent inversion result. The present invention can be applied to seismic intelligent inversion under conditions of no wells and wells at the same time, and can make full use of the advantages of deep learning to improve the accuracy of seismic inversion; it can be applied in the fields of seismic inversion technology and oil and gas reservoir prediction.

Description

Technical Field

[0001] The present invention relates to the field of seismic inversion technology, and in particular to a seismic intelligent inversion method and system jointly driven by geological modeling and seismic forward modeling. Background Art

[0002] Seismic intelligent inversion is a seismic inversion method based on artificial intelligence algorithms. It establishes a nonlinear mapping relationship between seismic and well logging data through a supervised learning algorithm, and then applies the mapping relationship between wells to achieve three-dimensional seismic inversion. This inversion method is a high-resolution and high-efficiency seismic inversion method. Seismic intelligent inversion has been widely used in oil and gas reservoir prediction and has achieved good results. Compared with traditional inversion methods (such as geological model inversion and sparse pulse inversion), artificial intelligence inversion does not require seismic wavelets and has the advantages of simple operation, easy promotion, and high resolution. However, since this method is based on a supervised learning algorithm, it requires a lot of well logging data as supervised learning data and cannot be applied to areas with few or no wells, which greatly limits the application and promotion of this method. In addition, for most work areas, the number of well logging is small, and the existing seismic intelligent inversion usually stays at "shallow learning", such as using support vector machines or neural networks with no more than 3 layers.

[0003] The modeling method of sedimentary evolution process, that is, under the guidance of sedimentary mode and evolution mechanism, simulates the deposition and evolution process of sedimentary phase in the study area, reproduces the geometric morphology and spatial distribution of geological bodies, and establishes a three-dimensional phase model of the study area. Taking the fluvial phase reservoir as an example, by simulating the formation and evolution process of the river channel, the geometric morphology and spatial distribution of the river channel are reproduced, so as to establish a three-dimensional phase model of the fluvial phase reservoir. The advantages of geological modeling based on sedimentary evolution process are that it fully reflects the geometric morphology and stacking relationship of the geological body, and is highly consistent with the sedimentary mode of the geological body; the disadvantages are that the conditional data fitting is difficult (generally only a small amount of well data can be fitted); the modeling results are highly random, and usually only one possibility of reservoir distribution can be obtained, which cannot be used as the final reservoir distribution. Seismic forward modeling is the process of obtaining the corresponding seismic response with a known geological model. This method has been widely used in seismic interpretation. That is, by conducting seismic forward modeling of different stratigraphic structures and sand body distribution characteristics, the seismic response characteristics of the rock residual sand body distribution characteristics of various stratigraphic structures are clarified, providing an interpretation basis for seismic geological interpretation.

[0004] In summary, the existing methods have the following defects: (1) In the field of sand body prediction or reservoir prediction, the existing seismic intelligent inversion method requires a large amount of well logging data as supervised learning data, that is, it requires a large number of wells and cannot be applied to areas with few wells. (2) For most work areas, the number of well logging is small, and the existing seismic intelligent inversion usually stays at "shallow learning", such as using support vector machine algorithms or neural network algorithms with no more than 3 layers. Summary of the invention

[0005] In response to the above problems, the purpose of the present invention is to provide a seismic intelligent inversion method and system jointly driven by geological modeling and seismic forward modeling, which can be applied to seismic intelligent inversion both with and without wells, and can give full play to the advantages of deep learning to improve the accuracy of seismic inversion.

[0006] To achieve the above-mentioned purpose, the present invention adopts the following technical scheme: a seismic intelligent inversion method jointly driven by geological modeling and seismic forward modeling, which comprises: taking an area with no wells or few wells as a study area, and selecting a sedimentary evolution model that conforms to the sedimentary phase type of the study area based on the sedimentary phase type of the study area; simulating the deposition and evolution process of the geological body based on the sedimentary evolution model, establishing a three-dimensional sedimentary phase model of the study area, and constructing a three-dimensional wave impedance model based on the three-dimensional sedimentary phase model, performing seismic forward modeling simulation, and obtaining a forward seismic data body of the study area; setting a virtual well, performing deep learning on the virtual well and the forward seismic data body to obtain a tested learning model, inputting an actual seismic data body into the tested learning model, solving the actual seismic data body, and outputting a seismic intelligent inversion result.

[0007] Furthermore, based on the sedimentary evolution model, the sedimentation and evolution process of the geological body is simulated to establish a three-dimensional sedimentary facies model of the study area, including:

[0008] Restore the paleo-geomorphic features of the study area by seismic horizon interpretation alone or by seismic horizon interpretation and well logging stratification data, and select seismic attributes that can reflect the distribution of sedimentary facies;

[0009] Using seismic attributes alone or seismic attributes and well logging data as conditional data, and paleo-geomorphological data as the basis for the sedimentary direction and evolutionary process of the geological body, under the guidance of the sedimentary evolution model and the limitation of the stratigraphic framework, the sedimentary and evolutionary process of the geological body is simulated layer by layer, and then the geometric form and spatial distribution of the geological body are simulated to establish a spatial distribution model of the geological body;

[0010] According to the geometric shape and lithological characteristics of the geological body, the sedimentary facies type of each geological body is clarified, so as to transform the spatial distribution model of the geological body into a three-dimensional sedimentary facies model.

[0011] Furthermore, the method for restoring the paleo-geomorphological features of the study area includes:

[0012] If there are a small amount of well data in the study area, an isochronous stratigraphic framework of the study area is established based on seismic horizon interpretation and well logging layering data;

[0013] On the basis of the isochronous stratigraphic framework, the methods of marker layer flattening and sediment thickness correction were adopted to restore the paleo-geomorphic features of the study area.

[0014] Further, constructing a three-dimensional wave impedance model based on the three-dimensional sedimentary phase model includes:

[0015] According to the outcrop deposits with similar sedimentary facies characteristics to the study area or the actual logging data in the study area, the range of wave impedance values ​​of each sedimentary microfacies is clarified;

[0016] Under the limitation of the wave impedance value domain, in the sedimentary phase model after wave impedance calibration, if there are a small number of wells in the study area, the three-dimensional wave impedance model of the study area is established by interpolating well data; if there are no wells in the study area, the three-dimensional wave impedance model of the study area is established by assignment.

[0017] Furthermore, the seismic forward modeling is performed to obtain the forward seismic data volume of the study area, including:

[0018] Extracting the seismic statistical wavelet of the target layer segment from the actual seismic data volume of the study area as the actual seismic wavelet of the study area;

[0019] The three-dimensional seismic forward modeling method is used to perform three-dimensional seismic forward simulation on the three-dimensional wave impedance model of the study area through the actual seismic wavelet to obtain a three-dimensional forward seismic data volume of the study area.

[0020] Furthermore, when there is no well condition, the virtual well is set, and deep learning is performed with the forward seismic data volume to obtain a tested learning model, which includes:

[0021] Setting the virtual well in the forward modeling seismic data volume;

[0022] Using the geological data of the virtual well and the well bypass forward modeling seismic data as a forward modeling learning data set;

[0023] Randomly selecting a portion of the data from the forward learning data set as a training data set and another portion as a test data set, and using a deep learning algorithm to perform deep learning on the training data set to establish an initial learning model;

[0024] The initial learning model is tested using the test data set, and after passing the test, the tested learning model is obtained.

[0025] Furthermore, when well conditions are available, the method of setting virtual wells, performing deep learning on all wells and all seismic data, and obtaining a tested learning model includes:

[0026] Setting the virtual well in the forward modeling seismic data volume;

[0027] The geological data of the virtual well and the seismic data of the well bypass of the virtual well are used as a forward learning data set, and the well logging interpretation data of the actual well and the seismic data of the well bypass of the actual well are used as an actual learning data set, and the forward learning data set and the actual learning data set constitute a total learning data set;

[0028] Randomly select a portion of the data from the total learning data set as a training data set and another portion as a test data set, and use a deep learning algorithm to perform supervised learning on the training data set, using a learning network of no less than 5 layers in the learning process to establish an initial learning model;

[0029] The initial learning model is tested using the test data set, and after passing the test, the tested learning model is obtained.

[0030] A seismic intelligent inversion system jointly driven by geological modeling and seismic forward modeling comprises: a first processing module, which takes an area with no wells or few wells as a study area, and selects a sedimentary evolution model that conforms to the sedimentary phase type of the study area based on the sedimentary phase type of the study area; a second processing module, which simulates the deposition and evolution process of the geological body based on the sedimentary evolution model, establishes a three-dimensional sedimentary phase model of the study area, and constructs a three-dimensional wave impedance model based on the three-dimensional sedimentary phase model, performs seismic forward modeling, and obtains a three-dimensional forward seismic data body of the study area; and an intelligent inversion output module, which sets up virtual wells, performs deep learning with the forward seismic data body, obtains a tested learning model, inputs an actual seismic data body into the tested learning model, solves the actual seismic data body, and outputs a seismic intelligent inversion result.

[0031] A computer-readable storage medium storing one or more programs, wherein the one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any of the above methods.

[0032] A computing device comprises: one or more processors, a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any of the above methods.

[0033] The present invention adopts the above technical solution, which has the following advantages:

[0034] 1. The present invention solves the problem that seismic intelligent inversion cannot be carried out under the condition of no wells (or few wells).

[0035] 2. The present invention organically combines geological modeling, seismic forward modeling and seismic intelligent inversion, thereby improving the accuracy of seismic inversion under conditions of few wells. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a schematic diagram of the overall process of a seismic intelligent inversion method jointly driven by geological modeling and seismic forward modeling in one embodiment of the present invention;

[0037] Figure 2 It is a detailed flow chart of a seismic intelligent inversion method jointly driven by geological modeling and seismic forward modeling in one embodiment of the present invention;

[0038] Figure 3 is a flow chart of constructing a three-dimensional sedimentary facies model in one embodiment of the present invention;

[0039] Figure 4 is a flow chart of constructing a three-dimensional wave impedance model in one embodiment of the present invention;

[0040] Figure 5 It is a flow chart of intelligent learning of virtual well setting and forward modeling seismic data volume under well-free conditions in one embodiment of the present invention;

[0041] Figure 6 It is a flow chart of intelligent learning of virtual well setting and forward modeling seismic data volume under well conditions in one embodiment of the present invention. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0043] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0044] Abbreviations and key terms are defined as:

[0045] Sedimentary evolution geological modeling: also known as sedimentary evolution modeling, is a geological modeling method based on the sedimentation and evolution process of geological bodies;

[0046] Sedimentary microfacies: smaller sedimentary units within a sedimentary facies;

[0047] Earthquake forward modeling: the process of obtaining the corresponding earthquake response given a known geological model;

[0048] Seismic inversion: The process of converting seismic data into a form that can be directly compared with well log data (or geological bodies);

[0049] Seismic intelligent inversion: seismic inversion method based on artificial intelligence algorithm.

[0050] The seismic intelligent inversion method and system jointly driven by geological modeling and seismic forward modeling provided by the present invention include: taking a well-free or well-less area as a study area, selecting a sedimentary evolution mode that conforms to the sedimentary phase type of the study area based on the sedimentary phase type of the study area; simulating the deposition and evolution process of the geological body based on the sedimentary evolution mode, establishing a three-dimensional sedimentary phase model of the study area, and constructing a three-dimensional wave impedance model based on the three-dimensional sedimentary phase model, performing seismic forward modeling simulation, and obtaining a forward seismic data body of the study area; setting a virtual well, performing deep learning with the forward seismic data, obtaining a tested learning model, inputting the actual seismic data body into the tested learning model, solving the actual seismic data body, and outputting a seismic intelligent inversion result. The present invention can be applied to seismic intelligent inversion under both well-free and well-containing conditions, and can give full play to the advantages of deep learning and improve the accuracy of seismic inversion.

[0051] In one embodiment of the present invention, a seismic intelligent inversion method driven by geological modeling and seismic forward modeling is provided. In this embodiment, for reservoir prediction in areas with no wells or few wells (such as an average well spacing greater than 2 kilometers or the number of wells drilled is less than 10), seismic intelligent inversion is performed, such as Figure 1 , Figure 2 As shown, the method comprises the following steps:

[0052] 1) Take the area with no wells or few wells as the study area, and select the sedimentary evolution model that conforms to the sedimentary facies type of the study area based on the sedimentary facies type of the study area;

[0053] 2) Based on the sedimentary evolution model, simulate the sedimentation and evolution process of the geological body and establish a three-dimensional sedimentary phase model (phase model for short) of the study area; and construct a three-dimensional wave impedance model based on the three-dimensional sedimentary phase model to perform seismic forward modeling and obtain the forward seismic data volume of the study area;

[0054] 3) Set up a virtual well, conduct deep learning on the virtual well and the forward seismic data volume to obtain a tested learning model, input the actual seismic data volume into the tested learning model, solve the actual seismic data volume, and output the seismic intelligent inversion results.

[0055] In the above step 2), based on the sedimentary evolution model, the deposition and evolution process of the geological body is simulated, the geometric shape, spatial distribution and superposition relationship of the geological body are reproduced, and a three-dimensional sedimentary facies model of the study area is established, such as Figure 3 As shown, the following steps are included:

[0056] 2.1.1) Restore the paleo-geomorphic features of the study area by seismic horizon interpretation alone or by combining seismic horizon interpretation with well logging stratification data, and select seismic attributes that can reflect the distribution of sedimentary facies;

[0057] 2.1.2) Using seismic attributes alone or seismic attributes and well logging data as conditional data, and paleo-geomorphological data as the basis for the sedimentary direction and evolutionary process of the geological body, under the guidance of the sedimentary evolution model and within the constraints of the stratigraphic framework, simulate the sedimentary and evolutionary process of the geological body layer by layer, and then simulate the geometric morphology and spatial distribution of the geological body to establish a spatial distribution model of the geological body;

[0058] Specifically:

[0059] If there is no well data in the study area, the seismic attributes are used as conditional data, and the paleo-geomorphological data are used as the basis for the sedimentary direction and evolution process of the geological body. Under the guidance of the sedimentary evolution model and the limitation of the stratigraphic framework, the sedimentary and evolution process of the geological body is simulated layer by layer, and then the geometric morphology and spatial distribution of the geological body are simulated to establish a spatial distribution model of the geological body.

[0060] If there are a small amount of well data in the study area, the seismic attributes and well logging data are used as conditional data, and the paleo-geomorphological data are used as the basis for the sedimentary direction and evolution process of the geological body. Under the guidance of the sedimentary evolution model and the limitation of the stratigraphic framework, the sedimentary and evolutionary process of the geological body is simulated layer by layer, and then the geometric morphology and spatial distribution of the geological body are simulated to establish a spatial distribution model of the geological body.

[0061] 2.1.3) According to the geometric shape and lithological characteristics of the geological body, the sedimentary facies type of each geological body is clarified. For example, in the context of river deposition, the strip-shaped thick sand body is the river channel facies, thereby converting the spatial distribution model of the geological body into a three-dimensional sedimentary facies model.

[0062] In the above embodiment, the modeling result of the three-dimensional sedimentary facies model is only a random model that conforms to the sedimentary pattern of the study area, and cannot represent the actual sedimentary facies distribution of the study area.

[0063] In step 2.1.1) above, the method for restoring the paleo-geomorphological features of the study area includes the following steps:

[0064] 2.1.1.1) Based on the interpretation of seismic horizons or in combination with well logging data (if available), establish a high-precision isochronous stratigraphic framework for the study area;

[0065] Specifically: if there is no well data in the study area, an isochronous stratigraphic framework of the study area is established based on seismic horizon interpretation;

[0066] If there are a small amount of well data in the study area, an isochronous stratigraphic framework of the study area is established based on seismic horizon interpretation and well logging layering data;

[0067] 2.1.1.2) Based on the isochronous stratigraphic framework, the methods of marker layer flattening and sediment thickness correction are used to restore the paleo-geomorphological characteristics of the study area.

[0068] In the above step 2), a three-dimensional wave impedance model is constructed based on the three-dimensional sedimentary phase model, such as Figure 4 As shown, the following steps are included:

[0069] 2.2.1) Based on outcrop deposits with similar sedimentary facies characteristics to the study area or actual well logging data in the study area, clarify the range of wave impedance values ​​of each sedimentary microfacies;

[0070] If there is no well data in the study area, the wave impedance range of each sedimentary microfacies shall be clarified (calibrated) based on outcrop deposits with similar sedimentary facies characteristics to those in the study area;

[0071] If there are a small amount of well data in the study area, the wave impedance value range of each sedimentary microfacies can be clarified (calibrated) based on the actual well logging data (such as acoustic wave and density logging) in the study area;

[0072] 2.2.2) Within the range of wave impedance values, in the sedimentary facies model after wave impedance calibration, if there are a small number of wells in the study area, a three-dimensional wave impedance model is established by interpolating well data; if there are no wells in the study area, a three-dimensional wave impedance model is established by assigning values.

[0073] In the above embodiment, since the sedimentary phase model conforms to the sedimentary pattern (ie, distribution law) of the study area, the wave impedance model also conforms to the wave impedance distribution law of the geological body in the study area.

[0074] In the above step 2), the method of performing seismic forward modeling to obtain the forward modeling seismic data volume of the study area includes the following steps:

[0075] 2.3.1) Extract the seismic statistical wavelet of the target layer from the actual seismic data volume of the study area as the actual seismic wavelet of the study area;

[0076] 2.3.2) Through the actual seismic wavelet, the three-dimensional wave impedance model of the study area is simulated by the seismic forward modeling method to obtain the three-dimensional forward seismic data volume of the study area.

[0077] In this embodiment, the seismic forward modeling method may adopt wavefront velocity field scanning or wave equation.

[0078] In the above embodiment, because the three-dimensional wave impedance model conforms to the wave impedance distribution law of the study area and the seismic wavelet comes from the actual seismic data body of the study area, the forward seismic data conforms to the seismic response mode and law of the study area.

[0079] In the above step 3), setting up virtual wells based on well and well-free data and deep learning based on forward seismic data are divided into two technical solutions. Among them:

[0080] like Figure 5 As shown, when there is no well condition, a virtual well is set, and deep learning is performed with the forward seismic data body to obtain a tested learning model, and the method includes the following steps:

[0081] 3.1.1) Set virtual wells in the forward seismic model. The initial value is 100 virtual wells, which can be adjusted;

[0082] 3.1.2) Since the geological model and its forward seismic data volume are known, the geological data of the virtual well and the seismic data of the virtual well's side channel are used as the forward learning data set;

[0083] 3.1.3) Randomly select a part of the data from the forward learning data set as the training data set and the other part as the test data set, and use the deep learning algorithm to perform deep learning on the training data set to establish an initial learning model;

[0084] In this embodiment, 70% of the forward learning data set is randomly selected as the training data set, and 30% of the data is used as the test data set;

[0085] Deep learning algorithms can use existing methods, such as random forest and convolutional neural network algorithms.

[0086] 3.1.4) The initial learning model is tested with the test data set. If the test is qualified, the tested learning model is obtained. If the test is unqualified, the model parameters are optimized and the number of virtual wells is adjusted. The optimization process is iterated until the model is qualified.

[0087] like Figure 6 As shown, when there are well conditions (i.e., actual wells), the number of virtual wells is set, and all wells (including virtual wells and actual wells) and all seismic data (including actual seismic data and forward seismic data bodies) are subjected to deep learning. The method for obtaining a tested learning model includes the following steps:

[0088] 3.2.1) Set virtual wells in the forward seismic data volume. The initial value is 100 virtual wells, which can be adjusted;

[0089] 3.2.2) The geological data of the virtual well and the seismic data of the well bypass of the virtual well are used as the forward learning data set, and the well logging interpretation data of the actual well and the seismic data of the well bypass of the actual well are used as the actual learning data set. The forward learning data set and the actual learning data set constitute the total learning data set;

[0090] 3.2.3) Randomly select a portion of the total learning data set as the training data set and another portion as the test data set, and use a deep learning algorithm to perform supervised learning on the training data set. A learning network with no less than 5 layers is used in the learning process to establish an initial learning model; deep learning in the true sense is achieved through a learning network with no less than 5 layers.

[0091] In this embodiment, 70% of the total learning data set is randomly selected as the training data set, and 30% of the data is used as the test data set;

[0092] Deep learning algorithms can use existing methods, such as random forest and convolutional neural network algorithms.

[0093] 3.2.4) The initial learning model is tested with the test data set. If the test model is qualified, the tested learning model is obtained. If the test model is unqualified, the model parameters are optimized and the number of virtual wells is adjusted. The optimization process is iterated until the model is qualified.

[0094] In summary, the present invention is suitable for seismic intelligent inversion under no well conditions (or few well conditions), and has good application prospects and significant industrial value for oil fields that lack well data (such as offshore oil fields, onshore oil fields, and blocks in the early stages of exploration or development). Through sedimentary evolution geological modeling and seismic forward simulation, a large amount of virtual well data is provided for the study area, thereby providing sufficient learning samples for deep learning. Sufficient learning samples can apply multi-layer (5 layers and above) learning networks, fully reflecting the advantages of deep learning (the present invention selects random forest or convolutional neural network algorithms), better explore the mapping relationship between geological bodies and seismic responses; ultimately improve the reliability and recognition accuracy of seismic inversion results.

[0095] In one embodiment of the present invention, a seismic intelligent inversion system driven jointly by geological modeling and seismic forward modeling is provided, which comprises:

[0096] The first processing module takes the area without wells or with few wells as the study area, and selects the sedimentary evolution model that meets the sedimentary facies type of the study area based on the sedimentary facies type of the study area;

[0097] The second processing module simulates the deposition and evolution process of the geological body based on the sedimentary evolution model, establishes a three-dimensional sedimentary phase model of the study area, and constructs a three-dimensional wave impedance model based on the three-dimensional sedimentary phase model to perform seismic forward modeling and obtain the forward seismic data body of the study area;

[0098] The intelligent inversion output module sets up a virtual well and conducts deep learning with the forward seismic data body to obtain a tested learning model. The actual seismic data body is input into the tested learning model, the actual seismic data body is solved, and the seismic intelligent inversion results are output.

[0099] The system provided in this embodiment is used to execute the above-mentioned method embodiments. Please refer to the above-mentioned embodiments for specific processes and detailed contents, which will not be repeated here.

[0100] The computing device structure provided in one embodiment of the present invention may be a terminal, which may include: a processor, a communications interface, a memory, a display screen and an input device. Among them, the processor, the communications interface and the memory communicate with each other through a communication bus. The processor is used to provide computing and control capabilities. The memory includes a non-volatile storage medium and an internal memory, wherein the non-volatile storage medium stores an operating system and a computer program, and when the computer program is executed by the processor, a method for intelligent seismic inversion is implemented; the internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be implemented through WIFI, a management network, NFC (near field communication) or other technologies. The display screen may be a liquid crystal display screen or an electronic ink display screen, and the input device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computing device, or an external keyboard, touchpad or mouse, etc. The processor can call the logic instructions in the memory to execute the following method: take the area with no wells or few wells as the study area, and based on the sedimentary phase type of the study area, select the sedimentary evolution model that conforms to the sedimentary phase type of the study area; based on the sedimentary evolution model, simulate the deposition and evolution process of the geological body, establish a three-dimensional sedimentary phase model of the study area, and construct a three-dimensional wave impedance model based on the three-dimensional sedimentary phase model, perform seismic forward simulation, and obtain the forward seismic data body of the study area; set up virtual wells, perform deep learning with the forward seismic data body, obtain a tested learning model, input the actual seismic data body into the tested learning model, solve the actual seismic data body, and output the seismic intelligent inversion result.

[0101] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0102] Those skilled in the art will understand that the above-mentioned computing device structure is only a partial structure related to the scheme of the present application, and does not constitute a limitation on the computing device to which the scheme of the present application is applied. The specific computing device may include more or fewer components, or combine certain components, or have a different arrangement of components.

[0103] In one embodiment of the present invention, a computer program product is provided, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the methods provided by the above-mentioned method embodiments, for example, including: taking an area with no wells or few wells as a study area, and selecting a sedimentary evolution model that conforms to the sedimentary phase type of the study area based on the sedimentary phase type of the study area; simulating the deposition and evolution process of the geological body based on the sedimentary evolution model, establishing a three-dimensional sedimentary phase model of the study area, and constructing a three-dimensional wave impedance model based on the three-dimensional sedimentary phase model, performing seismic forward modeling, and obtaining a forward seismic data body of the study area; setting up virtual wells, performing deep learning with the forward seismic data body, obtaining a tested learning model, inputting an actual seismic data body into the tested learning model, solving the actual seismic data body, and outputting a seismic intelligent inversion result.

[0104] In one embodiment of the present invention, a non-transitory computer-readable storage medium is provided, which stores server instructions, and the computer instructions enable a computer to execute the methods provided in the above embodiments, for example, including: taking an area with no wells or few wells as a study area, and selecting a sedimentary evolution model that conforms to the sedimentary phase type of the study area based on the sedimentary phase type of the study area; simulating the deposition and evolution process of the geological body based on the sedimentary evolution model, establishing a three-dimensional sedimentary phase model of the study area, and constructing a three-dimensional wave impedance model based on the three-dimensional sedimentary phase model, performing seismic forward modeling, and obtaining a forward seismic data body of the study area; setting up virtual wells, performing deep learning with the forward seismic data body, obtaining a tested learning model, inputting the actual seismic data body into the tested learning model, solving the actual seismic data body, and outputting seismic intelligent inversion results.

[0105] The above embodiment provides a computer-readable storage medium, whose implementation principle and technical effect are similar to those of the above method embodiment, and will not be repeated here.

[0106] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0107] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0108] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0109] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A seismic intelligent inversion method driven jointly by geological modeling and seismic forward modeling, characterized in that: include: The area without wells or with few wells is taken as the study area, and based on the sedimentary facies type of the study area, the sedimentary evolution model that conforms to the sedimentary facies type of the study area is selected; Based on the sedimentary evolution model, the sedimentation and evolution process of the geological body is simulated to establish a three-dimensional sedimentary phase model of the study area, and a three-dimensional wave impedance model is constructed based on the three-dimensional sedimentary phase model to perform seismic forward modeling to obtain a forward seismic data body of the study area; Setting a virtual well, performing deep learning on the virtual well and the forward modeling seismic data volume to obtain a tested learning model, inputting an actual seismic data volume into the tested learning model, solving the actual seismic data volume, and outputting a seismic intelligent inversion result; The method of simulating the deposition and evolution process of the geological body based on the deposition evolution model to establish a three-dimensional depositional facies model of the study area includes: Restore the paleo-geomorphic features of the study area by seismic horizon interpretation alone or by seismic horizon interpretation and well logging stratification data, and select seismic attributes that can reflect the distribution of sedimentary facies; Using seismic attributes alone or seismic attributes and well logging data as conditional data, and paleo-geomorphological data as the basis for the sedimentary direction and evolutionary process of the geological body, under the guidance of the sedimentary evolution model and the limitation of the stratigraphic framework, the sedimentary and evolutionary process of the geological body is simulated layer by layer, and then the geometric form and spatial distribution of the geological body are simulated to establish a spatial distribution model of the geological body; According to the geometric shape and lithological characteristics of the geological body, the sedimentary facies type of each geological body is clarified, so as to transform the spatial distribution model of the geological body into a three-dimensional sedimentary facies model.

2. The seismic intelligent inversion method driven by geological modeling and seismic forward modeling as claimed in claim 1, characterized in that: The method for restoring the paleo-geomorphic features of the study area comprises: If there are a small amount of well data in the study area, an isochronous stratigraphic framework of the study area is established based on seismic horizon interpretation and well logging layering data; On the basis of the isochronous stratigraphic framework, the methods of marker layer flattening and sediment thickness correction were adopted to restore the paleo-geomorphic features of the study area.

3. The seismic intelligent inversion method driven by geological modeling and seismic forward modeling as claimed in claim 1, characterized in that: The constructing of a three-dimensional wave impedance model based on the three-dimensional sedimentary phase model comprises: According to the outcrop deposits with similar sedimentary facies characteristics to the study area or the actual logging data in the study area, the range of wave impedance values ​​of each sedimentary microfacies is clarified; Under the limitation of the wave impedance value domain, in the sedimentary phase model after wave impedance calibration, if there are a small number of wells in the study area, the three-dimensional wave impedance model of the study area is established by interpolating well data; if there are no wells in the study area, the three-dimensional wave impedance model of the study area is established by assignment.

4. The seismic intelligent inversion method driven by geological modeling and seismic forward modeling as claimed in claim 1, characterized in that: The forward modeling of earthquakes is performed to obtain the forward modeling earthquake data volume of the study area, including: Extracting the seismic statistical wavelet of the target layer segment from the actual seismic data volume of the study area as the actual seismic wavelet of the study area; The three-dimensional seismic forward modeling method is used to perform three-dimensional seismic forward simulation on the three-dimensional wave impedance model of the study area through the actual seismic wavelet to obtain a three-dimensional forward seismic data volume of the study area.

5. The seismic intelligent inversion method driven by geological modeling and seismic forward modeling as claimed in claim 1, characterized in that: When there is no well condition, the virtual well is set, and deep learning is performed with the forward seismic data volume to obtain a tested learning model, which includes: Setting the virtual well in the forward modeling seismic data volume; Using the geological data of the virtual well and the well bypass forward modeling seismic data as a forward modeling learning data set; Randomly selecting a portion of the data from the forward learning data set as a training data set and another portion as a test data set, and using a deep learning algorithm to perform deep learning on the training data set to establish an initial learning model; The initial learning model is tested using the test data set, and after passing the test, the tested learning model is obtained.

6. The seismic intelligent inversion method driven jointly by geological modeling and seismic forward modeling as claimed in claim 1, characterized in that: When well conditions are available, the method of setting virtual wells, performing deep learning on all wells and all seismic data, and obtaining a tested learning model includes: Setting the virtual well in the forward modeling seismic data volume; The geological data of the virtual well and the seismic data of the well bypass of the virtual well are used as a forward learning data set, and the well logging interpretation data of the actual well and the seismic data of the well bypass of the actual well are used as an actual learning data set, and the forward learning data set and the actual learning data set constitute a total learning data set; Randomly select a portion of the data from the total learning data set as a training data set and another portion as a test data set, and use a deep learning algorithm to perform supervised learning on the training data set, using a learning network of no less than 5 layers in the learning process to establish an initial learning model; The initial learning model is tested using the test data set, and after passing the test, the tested learning model is obtained.

7. A seismic intelligent inversion system driven jointly by geological modeling and seismic forward modeling, characterized in that: include: The first processing module takes an area without wells or with few wells as a study area, and selects a sedimentary evolution model that conforms to the sedimentary facies type of the study area based on the sedimentary facies type of the study area; The second processing module simulates the deposition and evolution process of the geological body based on the deposition evolution model, establishes a three-dimensional depositional phase model of the study area, and constructs a three-dimensional wave impedance model based on the three-dimensional depositional phase model to perform seismic forward modeling to obtain a three-dimensional forward seismic data body of the study area; The intelligent inversion output module sets a virtual well, performs deep learning with the forward seismic data volume, obtains a tested learning model, inputs the actual seismic data volume into the tested learning model, solves the actual seismic data volume, and outputs the seismic intelligent inversion result; The method of simulating the deposition and evolution process of the geological body based on the deposition evolution model to establish a three-dimensional depositional facies model of the study area includes: Restore the paleo-geomorphic features of the study area by seismic horizon interpretation alone or by seismic horizon interpretation and well logging stratification data, and select seismic attributes that can reflect the distribution of sedimentary facies; Using seismic attributes alone or seismic attributes and well logging data as conditional data, and paleo-geomorphological data as the basis for the sedimentary direction and evolutionary process of the geological body, under the guidance of the sedimentary evolution model and the limitation of the stratigraphic framework, the sedimentary and evolutionary process of the geological body is simulated layer by layer, and then the geometric form and spatial distribution of the geological body are simulated to establish a spatial distribution model of the geological body; According to the geometric shape and lithological characteristics of the geological body, the sedimentary facies type of each geological body is clarified, so as to transform the spatial distribution model of the geological body into a three-dimensional sedimentary facies model.

8. A computer-readable storage medium storing one or more programs, characterized in that: The one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any one of the methods of claims 1 to 6.

9. A computing device, characterized in that include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any one of the methods described in claims 1 to 6.

Images