An underwater stratum data deduction method

Abstract

The application discloses an underwater stratum data deduction method, which comprises the following steps: acquiring multi-source data of an underwater stratum, pre-processing the acquired multi-source data, and constructing a multi-modal data set; acquiring a hierarchical space adjacency graph and a multi-scale time adjacency graph, and constructing a multi-scale space-time coupling graph; fusing and extracting multi-scale space-time features based on a multi-scale space-time convolutional neural network, embedding a physical field model in the multi-scale space-time convolutional neural network in the form of a model loss function, and constructing an underwater stratum space-time deduction model; training model parameters based on a joint training strategy combining self-supervised learning and supervised learning; inputting current underwater stratum state and marine environment prediction data for a future period of time into the optimized underwater stratum space-time deduction model, and outputting stratum evolution prediction results at different time scales. The underwater stratum data deduction method provided in the application embodiment can realize high-precision space-time evolution prediction of an underwater stratum.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of underwater stratum dynamic simulation, and more particularly to an underwater stratum data deduction method, device, equipment and storage medium. BACKGROUND

[0002] The prediction of spatio-temporal evolution of underwater strata is crucial for the development of marine resources, the construction and operation of cross-sea facilities, and the study of marine environment. The seabed is rich in resources such as oil and natural gas, and the construction and operation of cross-sea facilities require accurate assessment of the stability of the seabed strata. However, the evolution of underwater strata is influenced by a variety of complex factors, including sedimentation, tectonic movement, and changes in marine environment. Traditional methods of stratum evolution prediction mainly rely on seismic data and simple mathematical models, which have obvious limitations in dealing with complex stratum structures and long-term evolution.

[0003] In the prior art, some studies analyze strata by combining seismic data with simple mathematical models, but this method cannot effectively integrate multi-source data and is difficult to reflect the complex changes of strata in space and time. Although some methods based on machine learning improve the prediction accuracy to some extent, they still have many shortcomings. First, existing methods are inefficient in data processing, and data from different sources have different scales and formats, making it difficult to directly fuse and analyze. Second, existing methods lack in-depth consideration of physical mechanisms, resulting in insufficient rationality and reliability of the prediction results. In addition, existing methods also have shortcomings in multi-scale analysis, making it difficult to comprehensively and accurately simulate the evolution process of strata at different time and spatial scales. These shortcomings limit the effectiveness of existing technology in practical applications, so a new method is needed to overcome these problems and improve the accuracy and reliability of spatio-temporal evolution prediction of underwater strata. SUMMARY

[0004] In view of at least one defect or improvement demand of the prior art, the present application provides an underwater stratum data deduction method, which can solve at least one of the problems in the background art.

[0005] To achieve the above-mentioned purpose, according to the first aspect of the present application, an underwater stratum data deduction method is provided, which comprises:

[0006] Obtaining multi-source data of underwater strata, preprocessing the obtained multi-source data, and constructing a multi-modal data set;

[0007] Based on the multi-modal data set, a hierarchical spatial adjacency graph and a multi-scale temporal adjacency graph are obtained, and a multi-scale spatio-temporal coupling graph is constructed;

[0008] The multi-scale space-time features of the multi-scale space-time coupled graph are fused and extracted based on a multi-scale space-time convolutional neural network, a physical field model is embedded in the multi-scale space-time convolutional neural network in the form of a model loss function, and an underwater formation space-time deduction model is constructed.

[0009] A joint training strategy combining self-supervised learning and supervised learning is used to optimize parameters of the underwater formation space-time deduction model.

[0010] Current underwater formation state and marine environment prediction data for a period of time in the future are input into the optimized underwater formation space-time deduction model, and formation evolution prediction results at different time scales are output.

[0011] Further, the above-mentioned underwater formation data deduction method, the multi-source data of the underwater formation is obtained, the obtained multi-source data is preprocessed, and a multi-modal data set is constructed, specifically including:

[0012] The multi-source data of the underwater formation includes marine seismic data, logging while drilling time series data, seabed topography scanning data, seabed formation scanning data, and marine environment monitoring data.

[0013] The multi-source data of the underwater formation is unified to the same space-time coordinate system using a space-time alignment algorithm.

[0014] The missing data is processed using a completion algorithm combining a space-time variational autoencoder and a generative adversarial network to construct a multi-modal data set.

[0015] Further, the above-mentioned underwater formation data deduction method, the multi-modal data set is used to obtain a hierarchical spatial adjacency graph and a multi-scale time adjacency graph, and a multi-scale space-time coupled graph is constructed, specifically including:

[0016] The underwater formation space is divided into different scale voxel units, the node connection and edge weight are determined based on regional geological structure characteristics, formation property similarity and distance, and a hierarchical spatial adjacency graph is constructed.

[0017] The same spatial position nodes at different periods are connected based on time slices, and the time edge weight is adjusted according to the time interval and the marine environment parameter change rate, and a multi-scale time adjacency graph is constructed.

[0018] The multi-scale space-time coupled graph is obtained by fusing the spatial and time adjacency graphs.

[0019] Further, the above-mentioned underwater formation data deduction method, the physical field model is embedded in the multi-scale space-time convolutional neural network in the form of a model loss function, and an underwater formation space-time deduction model is constructed, specifically including:

[0020] Based on a multi-scale spatio-temporal convolutional neural network, spatial features are extracted through different size convolution kernels, and a recurrent neural network variant is combined with a multi-scale time adjacency graph to learn the time variation pattern.

[0021] A physical field model is embedded in the multi-scale spatio-temporal convolutional neural network, and a constraint term is added to the loss function to ensure that the model prediction conforms to the physical law.

[0022] An adaptive attention mechanism is introduced to calculate the importance score of the spatio-temporal node features, dynamically adjust the information propagation path and weight, and construct an underwater formation spatio-temporal inference model.

[0023] Further, the above-mentioned underwater formation data inference method, the joint training strategy based on the combination of self-supervised learning and supervised learning optimizes the parameters of the underwater formation spatio-temporal inference model, specifically including:

[0024] A joint training strategy combining self-supervised and supervised learning is adopted, and self-supervised learning enhances the understanding of the spatio-temporal features of the underwater formation spatio-temporal inference model through spatio-temporal contrast learning tasks;

[0025] Supervised learning optimizes the parameters of the underwater formation spatio-temporal inference model using stratigraphic evolution history data, and adjusts the neural network weights and biases using stochastic gradient descent and its variants.

[0026] Further, the above-mentioned underwater formation data inference method, the construction of a hierarchical spatial adjacency graph further includes:

[0027] An improved Gaussian kernel function is used to calculate the edge weight, and the formula is

[0028]

[0029] wherein is the edge weight between node and node , is the distance between nodes, is the scale parameter, is the attribute similarity between node and node .

[0030] Further, the above-mentioned underwater formation data inference method, the physical field model embedded in the multi-scale spatio-temporal convolutional neural network specifically includes:

[0031] In terms of deposition, the deposition rate is calculated through a formula according to the flow velocity, sediment particle size and density parameters, and is added to the model loss function as a constraint term;

[0032] In terms of tectonic movement, by using elastic mechanics and plate motion theory, a stress difference penalty term based on physical model calculation is added in the loss function by simulating the stress field distribution.

[0033] According to a second aspect of the present application, an underwater stratum data deduction device is also provided, comprising:

[0034] A data acquisition module is configured to acquire multi-source data of an underwater stratum, pre-process the acquired multi-source data, and construct a multi-modal data set.

[0035] A feature extraction module is configured to acquire a hierarchical spatial adjacency graph and a multi-scale time adjacency graph based on the multi-modal data set, and construct a multi-scale space-time coupled graph.

[0036] A model construction module is configured to fuse and extract multi-scale space-time features of the multi-scale space-time coupled graph based on a multi-scale space-time convolutional neural network, embed a physical field model in the multi-scale space-time convolutional neural network in the form of a model loss function, and construct an underwater stratum space-time deduction model.

[0037] A parameter optimization module is configured to optimize parameters of the underwater stratum space-time deduction model based on a joint training strategy combining self-supervised learning and supervised learning.

[0038] A prediction module is configured to input current underwater stratum state and future marine environment prediction data into the optimized underwater stratum space-time deduction model, and output stratum evolution prediction results at different time scales.

[0039] According to a third aspect of the present application, an underwater stratum data deduction device is also provided, comprising at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program which, when executed by the processing unit, causes the processing unit to perform the steps of any of the above-mentioned methods.

[0040] According to a fourth aspect of the present application, a storage medium is also provided, which stores a computer program executable by an underwater stratum data deduction device, and when the computer program runs on the underwater stratum data deduction device, causes the underwater stratum data deduction device to perform the steps of any of the above-mentioned methods.

[0041] Overall, the above technical solutions conceived by the present application can achieve the following beneficial effects compared with the prior art:

[0042] The underwater stratum data deduction method provided by the embodiments of the present application can realize high-precision spatiotemporal evolution prediction of the underwater stratum through technical means such as multi-source data fusion, multi-scale spatiotemporal graph construction, physical field constraint embedding, and joint training strategy. The multi-source data fusion preprocessing constructs a multi-modal data set, providing rich and accurate data basis for model training; the multi-scale spatiotemporal coupling graph construction comprehensively expresses the stratum spatiotemporal correlation; the physical field model is embedded in the multi-scale spatiotemporal convolutional neural network and an adaptive attention mechanism is introduced, ensuring that the prediction result conforms to the geological physical law and focuses on the key area and time period; and a joint training strategy combining self-supervised and supervised learning is used to optimize the model parameters. The final output prediction result is reliable and can provide strong technical support for marine resource development, cross-sea facility construction and the like. BRIEF DESCRIPTION OF DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort on the basis of these drawings.

[0044] Figure 1 A flowchart of an underwater stratum data deduction method provided by the embodiments of the present application is shown in DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solutions and advantages of the present application more clear, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application. In addition, the technical features involved in the various embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0046] The terms "first", "second", "third", etc. in the specification and claims of the present application and the above drawings are used to distinguish different objects, and are not used to describe a specific order. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but can optionally include steps or units not listed, or can optionally include other steps or units inherent to the process, method, product or device.

[0047] Figure 1 A flowchart of an underwater stratum data deduction method provided by the embodiments of the present application is shown in Figure 1 As shown in the flowchart, the underwater stratum data deduction method provided by the embodiments of the present application comprises the following steps.

[0048] acquire multi-source data of the underwater formation, preprocess the acquired multi-source data, and construct a multi-modal data set;

[0049] acquire a hierarchical spatial adjacency graph and a multi-scale temporal adjacency graph based on the multi-modal data set, and construct a multi-scale spatio-temporal coupling graph;

[0050] fuse and extract multi-scale spatio-temporal features of the multi-scale spatio-temporal coupling graph based on a multi-scale spatio-temporal convolutional neural network, embed a physical field model in the multi-scale spatio-temporal convolutional neural network in the form of a model loss function, and construct an underwater formation spatio-temporal inference model;

[0051] optimize the parameters of the underwater formation spatio-temporal inference model based on a joint training strategy combining self-supervised learning and supervised learning;

[0052] input the current underwater formation state and future marine environment prediction data into the optimized underwater formation spatio-temporal inference model, and output formation evolution prediction results at different time scales.

[0053] Specifically, first, multi-source data of the underwater formation is acquired. These data are widely sourced, including but not limited to marine seismic data, logging while drilling time series data, seafloor topography scanning data, seafloor formation scanning data, and marine environment monitoring data. Marine seismic data can provide macro information of the formation structure, and logging while drilling data contain physical property parameters of the formation, such as resistivity and natural gamma. Seafloor topography scanning data and formation scanning data provide morphological information of the seafloor surface and the formation interior, respectively. Marine environment monitoring data cover dynamic parameters such as current velocity, temperature, and salinity, which have important influences on the deposition and erosion processes of the formation.

[0054] The acquired multi-source data is preprocessed, and the preprocessing process includes data cleaning, correction, and fusion. Data cleaning aims to remove noise and outliers. For example, wavelet transform filtering technology can be used to remove noise from seismic data, and a phase correction algorithm can be used to improve data resolution. Logging data is normalized to eliminate bias caused by different measurement devices. Data fusion unifies data from different sources into the same spatio-temporal coordinate system using a spatio-temporal alignment algorithm. To address the problem of data missing, a completion algorithm combining a variational autoencoder (VAE) and a generative adversarial network (GAN) is used to generate reasonable pseudo data based on the spatio-temporal correlation of the data, thereby constructing a complete and accurate multi-modal data set.

[0055] Based on the pre-processed multi-modal data set, a hierarchical spatial adjacency graph and a multi-scale temporal adjacency graph are constructed, and then a multi-scale spatio-temporal coupling graph is formed. In the spatial dimension, the subsurface formation space is divided into different scale voxel units. For example, macro-scale can divide larger voxel units to capture large-scale structural features of the formation, and micro-scale can divide smaller voxel units to focus on the details of the formation. The connection relationship and edge weight between nodes are determined according to the geological structure characteristics, formation property similarity and spatial distance of different scale voxel units. The improved Gaussian kernel function can be applied to the calculation of edge weight to highlight the connection strength between nodes with high similarity and short distance.

[0056] In the time dimension, the nodes at the same spatial position in different periods are connected with a set time interval as a time slice. The determination of the time edge weight is dynamically adjusted according to the length of the time interval and the change rate of the marine environmental parameters. For example, if the current speed changes dramatically in a certain time period, it may have a greater erosion or deposition effect on the formation, and at this time the weight of the corresponding time edge will be reduced accordingly to reflect the increase in uncertainty of the formation change in the time interval. By fusing the spatial adjacency graph and the time adjacency graph, a multi-scale spatio-temporal coupling graph that can comprehensively express the spatio-temporal correlation of the formation is finally constructed, providing a data basis for subsequent feature extraction and model training.

[0057] Based on the multi-scale spatio-temporal convolutional neural network, the multi-scale spatio-temporal features of the multi-scale spatio-temporal coupling graph are fused and extracted. In the spatial dimension, different size convolution kernels are designed to act on different scale spatial adjacency graphs to extract the spatial structure features of the formation from macro to micro. Large convolution kernels can capture the overall trend of regional formation distribution, while small convolution kernels can focus on the details of local formation changes. In the time dimension, the recurrent neural network (RNN) variant is combined with the time adjacency graph to learn the dynamic change pattern of the formation over time, thereby realizing the deep fusion and extraction of multi-scale spatio-temporal features.

[0058] In order to embed the physical field model into the multi-scale spatio-temporal convolutional neural network, the soft constraint form is used to integrate it into the loss function of the model. For example, in the aspect of deposition, according to the parameters such as flow velocity, sediment particle size and density, the deposition rate is calculated through physical formula, and the calculated deposition rate is added as a constraint term to the loss function. For tectonic movement, the stress field distribution is simulated by using the theory of elasticity and plate movement, and then the difference between the stress calculated based on the physical model and the stress predicted by the model is added as a penalty term in the loss function. In this way, during the model training process, by optimizing the loss function, it can ensure that the predicted formation evolution result of the model conforms to the physical laws such as sediment dynamics and structural geology.

[0059] Meanwhile, an adaptive attention mechanism is introduced, allowing the model to automatically focus on key areas and time periods that have a greater impact on stratigraphic evolution. During the analysis process, the model calculates the importance scores of features at different spatiotemporal nodes, dynamically adjusting the paths and weights of information propagation based on the scores. For example, in areas with active tectonics or during periods of rapid deposition, the model increases the attention weight of the corresponding regional and temporal slice features, thereby improving the model's ability to extract complex stratigraphic evolution characteristics and prediction accuracy.

[0060] A joint training strategy combining self-supervised learning and supervised learning is used to optimize the parameters of the underwater stratigraphic spatiotemporal inference model. In the self-supervised learning phase, a spatiotemporal contrast learning task is designed. For example, given the states of the same stratigraphic region at different times, the model needs to determine whether these states belong to the same evolution sequence. By maximizing the similarity between positive sample pairs (i.e., true evolution states of the same region at different times) and minimizing the similarity between negative sample pairs (i.e., randomly selected unrelated states), the model can better understand the spatiotemporal characteristics and evolution rules of the stratigraphy.

[0061] In the supervised learning phase, known stratigraphic evolution history data is used as labels, which can be obtained from historical seismic data, drilling core analysis results, etc. By calculating the error between the model's prediction results and the true labels, and using stochastic gradient descent (SGD) and its variants to adjust the weights and biases of the neural network, the model's parameters are further optimized. The joint training strategy takes full advantage of both self-supervised learning and supervised learning, enabling the model to effectively learn the characteristics and patterns of stratigraphic evolution even in the absence of large amounts of labeled data, and continuously improving the accuracy of predictions.

[0062] The current underwater stratigraphic state and future marine environment prediction data for a certain period of time are input into the underwater stratigraphic spatiotemporal inference model that has been optimized. Marine environment prediction data can be obtained based on global climate change models, including information such as sea level rise rate, changes in ocean current patterns, etc. Based on the input data, the model outputs stratigraphic evolution prediction results at different time scales in the future, covering changes in stratigraphic interfaces, changes in lithology distribution, fault activity, and other aspects.

[0063] To evaluate the reliability of the prediction results, multiple indicators can be used for evaluation. The root mean square error can be used to measure the error in stratigraphic thickness prediction, and the intersection over union ratio can be used to evaluate the accuracy of lithology distribution prediction. In addition, qualitative evaluation of the prediction results is based on physical reasonableness, checking whether they conform to geological common sense and known physical laws. By comprehensively using these evaluation indicators, the performance of the model and the credibility of the prediction results can be verified comprehensively and objectively, providing reliable basis for practical applications such as marine resource development and cross-sea facility construction.

[0064] The underwater stratum data deduction method provided by the embodiments of the present application can realize high-precision spatiotemporal evolution prediction of the underwater stratum through technical means such as multi-source data fusion, multi-scale spatiotemporal graph construction, physical field constraint embedding and joint training strategy. The multi-source data fusion preprocessing constructs a multi-modal data set, providing a rich and accurate data basis for model training; the multi-scale spatiotemporal coupling graph construction comprehensively expresses the spatiotemporal correlation of the stratum; the physical field model is embedded in the multi-scale spatiotemporal convolutional neural network and an adaptive attention mechanism is introduced, ensuring that the prediction result conforms to the geological physical law and focuses on the key area and time period; and a joint training strategy combining self-supervised and supervised learning is used to optimize the model parameters. The final output prediction result is reliable and can provide strong technical support for marine resource development, cross-sea facility construction and the like.

[0065] Optionally, the underwater stratum data deduction method provided by the embodiments of the present application comprises the following steps.

[0066] The multi-source data of the underwater stratum is obtained, and the obtained multi-source data is preprocessed to construct a multi-modal data set, specifically comprising:

[0067] The multi-source data of the underwater stratum is unified to the same spatiotemporal coordinate system by using a spatiotemporal alignment algorithm.

[0068] The missing data is processed by using a completion algorithm based on the combination of a spatiotemporal variational autoencoder and a generative adversarial network to construct a multi-modal data set.

[0069] Specifically, the multi-source data of the underwater stratum is collected, which includes marine seismic data, logging while drilling time series data, seafloor topography scanning data, seafloor stratum scanning data and marine environment monitoring data. The marine seismic data can provide macro-structure information of the stratum, and the logging while drilling data contains physical property parameters of the stratum, such as resistivity, natural gamma and the like. The seafloor topography scanning data and the stratum scanning data provide morphological information of the seafloor surface and the stratum interior, respectively, and the marine environment monitoring data covers dynamic parameters such as current velocity, temperature, salinity and the like, which have important influence on the deposition and erosion process of the stratum.

[0070] Next, the spatiotemporal alignment algorithm is used to unify the multi-source data to the same spatiotemporal coordinate system. The spatiotemporal alignment algorithm can process data of different sources and different scales, ensuring the consistency of all data in time and space, and providing a basis for subsequent data fusion and analysis.

[0071] Finally, to address the potential missing parts in the data, a completion algorithm based on the combination of spatio-temporal variational autoencoder and generative adversarial network is adopted. This algorithm utilizes the spatio-temporal correlation of the data to generate reasonable pseudo-data to fill in the missing parts, thereby constructing a complete and accurate multi-modal data set. In this way, the integrity and accuracy of the data can be ensured, providing a high-quality data foundation for subsequent multi-scale spatio-temporal graph construction and model training.

[0072] Optionally, the underwater stratum data deduction method provided by the embodiments of the present application comprises:

[0073] The underwater stratum space is divided into different scale voxel units, and node connections and edge weights are determined based on regional geological structure characteristics, stratum attribute similarity, and distance to construct a layered space adjacency graph.

[0074] The same space position nodes in different periods are connected based on time slices, and the time edge weight is adjusted according to the time interval and the change rate of marine environmental parameters to construct a multi-scale time adjacency graph.

[0075] The multi-scale spatio-temporal coupling graph is obtained by fusing the space and time adjacency graphs.

[0076] Specifically, the underwater stratum space is divided into different scale voxel units, such as macro, meso, and micro scales. For each scale, the connection relationship and edge weight between nodes are determined according to the regional geological structure characteristics, stratum attribute similarity, and spatial distance. The edge weight is calculated by an improved Gaussian kernel function to highlight the connection strength between nodes with high similarity and short distance. At the macro scale, nodes are divided according to the macro similarity of regional stratum distribution; at the meso scale, node connections are determined in combination with lithology and stratum structure characteristics; at the micro scale, adjacency relationships are constructed according to micro attribute similarity such as lithology and porosity and spatial distance, and finally a layered space adjacency graph is constructed.

[0077] The same space position nodes in different periods are connected based on time slices, and the time edge weight is adjusted according to the time interval and the change rate of marine environmental parameters to construct a multi-scale time adjacency graph.

[0078] By fusing the spatial adjacency graph and the temporal adjacency graph, a multi-scale space-time coupled graph is formed. The graph can comprehensively express the change characteristics of the stratum at different spatial scales and time scales, each node contains rich stratum attributes (lithology, permeability, stratum thickness, etc.) and time stamp information, and the edges clearly express the space-time association at different scales, providing a data basis for detailed analysis in the spatial and time dimensions.

[0079] Optionally, the underwater stratum data deduction method provided in the embodiments of the present application, the physical field model is embedded in the multi-scale space-time convolutional neural network in the form of a model loss function, and an underwater stratum space-time deduction model is constructed, specifically comprising:

[0080] Based on the multi-scale space-time convolutional neural network, spatial features are extracted by different size convolution kernels, and a recurrent neural network variant is combined with a multi-scale time adjacency graph to learn a time change pattern;

[0081] The physical field model is embedded in the multi-scale space-time convolutional neural network, and a constraint term is added in the loss function to ensure that the model prediction conforms to the physical law;

[0082] An adaptive attention mechanism is introduced to calculate a space-time node feature importance score, dynamically adjust the information propagation path and weight, and construct an underwater stratum space-time deduction model.

[0083] Specifically, the multi-scale space-time convolutional neural network is used to process the multi-scale space-time coupled graph data. In the spatial dimension, different size convolution kernels are used to extract spatial features at each scale, large convolution kernels grasp the overall trend of the stratum, and small convolution kernels pay attention to details and capture the spatial structure characteristics of the stratum from macro to micro; in the time dimension, a recurrent neural network variant is combined with a multi-scale time adjacency graph to learn a dynamic change pattern of the stratum, realizing deep fusion and extraction of multi-scale space-time features.

[0084] The sedimentation dynamics, tectonic geology and other physical field models are embedded in the neural network in the form of soft constraint, and a constraint term is added in the loss function. In terms of deposition, the deposition rate is calculated based on the flow velocity, sediment particle size density and other parameters and set as a constraint, the deposition rate is calculated by a physical formula, and is added as a constraint term to the model loss function, to ensure that the predicted deposition layer thickness change of the model conforms to the actual physical law; in terms of tectonic movement, the stress field distribution is simulated by using the theory of elasticity and plate motion, and the displacement and deformation prediction of the stratum near the fault is constrained. A penalty term based on the difference between the stress calculated by the physical model and the stress predicted by the model is added in the loss function, so that the model continuously adjusts the parameters during the learning process to meet the physical field constraint.

[0085] An adaptive attention mechanism is introduced, and the model automatically focuses on key areas and time periods that have a greater impact on stratigraphic evolution when processing multi-scale spatiotemporal data. When analyzing active tectonic areas, the model increases the attention weight of the spatiotemporal features of the area through the attention mechanism; when studying rapid deposition periods, the features of the corresponding time slices are highlighted. By calculating the importance scores of different spatiotemporal nodes, the information propagation path and weight are dynamically adjusted to improve the model's ability to extract complex stratigraphic evolution features. The finally constructed underwater stratigraphic spatiotemporal inference model can accurately predict stratigraphic evolution.

[0086] Optionally, the underwater stratigraphic data inference method provided by the embodiment of the application, the joint training strategy based on the combination of self-supervised learning and supervised learning optimizes the underwater stratigraphic spatiotemporal inference model parameters, specifically including:

[0087] The joint training strategy combining self-supervised learning and supervised learning enhances the understanding of the underwater stratigraphic spatiotemporal inference model for stratigraphic spatiotemporal features through spatiotemporal contrast learning tasks;

[0088] Supervised learning optimizes the parameters of the underwater stratigraphic spatiotemporal inference model using stratigraphic evolution history data, and adjusts the neural network weights and biases using stochastic gradient descent and its variants.

[0089] Specifically, in the self-supervised learning phase, a spatiotemporal contrast learning task is designed, and data pairs of different periods of the same stratigraphic area are input. The model needs to determine whether they belong to the same evolution sequence. By maximizing the similarity of positive sample pairs (true evolution sequence) and minimizing the similarity of negative sample pairs (random unrelated sequence), the model's understanding of stratigraphic spatiotemporal features is enhanced. This process does not rely on labeled data and can fully utilize a large amount of unlabeled data to learn spatiotemporal features.

[0090] In the supervised learning phase, known stratigraphic evolution history data are used as labels, which come from historical seismic data and drilling core analysis results, etc. By calculating the error between the model's prediction results and the true labels, the weights and biases of the neural network are adjusted using stochastic gradient descent and its variants (such as Adagrad, Adadelta, Adam, etc.). This supervised learning method can further refine the model's understanding of stratigraphic evolution and improve the accuracy of the model's prediction.

[0091] During the entire joint training process, self-supervised learning and supervised learning complement each other. Self-supervised learning extracts spatiotemporal features from unlabeled data, and supervised learning uses labeled data to correct and optimize model parameters. This combination can make full use of rich unlabeled data and to some extent avoid overfitting, so that the model continuously converges during the training process and ultimately improves the prediction accuracy.

[0092] Optionally, the method for inferring underwater formation data provided by the embodiments of the present application comprises constructing a hierarchical spatial adjacency graph, and the method further comprises:

[0093] The edge weight is calculated by using the improved Gaussian kernel function, and the formula is

[0094]

[0095] wherein d is the distance between nodes, is the edge weight between node and node , is the distance between nodes, is a scale parameter, is the attribute similarity between node and node .

[0096] Specifically, in the embodiments of the present application, the improved Gaussian kernel function is used to calculate the edge weight when the hierarchical spatial adjacency graph is constructed, and the specific formula is:

[0097]

[0098] wherein d is the distance between nodes, is the edge weight between node and node , is the spatial distance between node and node , is a scale parameter, used to control the influence range of distance on weight, is the attribute similarity between node and node , reflecting the similarity of the two in formation attributes (such as lithology, porosity, etc.).

[0099] In the calculation process, the spatial distance between nodes is first determined , which can be obtained by using Euclidean distance or other appropriate spatial distance measurement methods. The attribute similarity is calculated based on the formation attributes of the nodes, and the more similar the attributes are, the higher the value is. The scale parameter is selected according to the formation characteristics and analysis scale of the specific research area, so as to balance the contribution of distance and attribute similarity to weight.

[0100] The edge weight calculated by the formula is , which can comprehensively reflect the spatial proximity and attribute similarity between nodes. In the construction of the hierarchical spatial adjacency graph, the connection between nodes with larger edge weights is closer, which helps to more accurately capture the spatial structure characteristics and attribute distribution rules of the stratum in the subsequent model training and deduction process, thereby improving the prediction ability of the model for stratum evolution.

[0101] Optionally, the underwater stratum data deduction method provided by the embodiment of the application comprises the following steps:

[0102] In terms of deposition, the deposition rate is calculated by formula according to the flow velocity, sediment particle size and density parameters, and is added as a constraint term to the model loss function;

[0103] In terms of tectonic movement, the stress field distribution is simulated by using the theory of elasticity and plate motion, and a penalty term based on the difference between the stress calculated by the physical model and the stress predicted by the model is added to the loss function.

[0104] In one specific embodiment, the underwater stratum data deduction method provided by the embodiment of the application is used, and the specific execution steps are as follows:

[0105] Data acquisition and preprocessing: In a target sea area of the Hong Kong-Zhuhai-Macao Bridge, past cross-period seismic data with a resolution of 10m x 10m x 5m covering different geological periods are collected; logging-while-drilling data of 5 exploration wells are selected, including resistivity, natural gamma and other parameters; detailed seabed topographic data of the region are obtained by using a multi-beam depth sounding device; stratum profile data of the region are obtained by using a shallow stratum profiler; long-term marine environmental monitoring data of the region are collected, including sea current velocity, temperature, salinity, etc. Through a spatio-temporal alignment algorithm, these data are unified to the same spatio-temporal coordinate system. For seismic data, wavelet transform filtering is used to remove noise, and a phase correction algorithm is used to improve resolution; logging data are normalized. The topographic data and stratum data are fused, and the water depth value of the topographic data is used to correct the water depth value of the stratum data. For the problem of missing part of the stratum profile data, a completion algorithm combining a spatio-temporal variational autoencoder with a generative adversarial network is used to generate reasonable pseudo data, and a complete multi-modal data set is constructed.

[0106] Multi-scale spatio-temporal graph construction: In the spatial dimension, the regional strata are divided into macro (1 km x 1 km x 100 m), meso (100 m x 100 m x 10 m), and micro (10 m x 10 m x 1 m) scale voxel units. At the macro scale, based on the regional geological structure characteristics, the region with similar sedimentary facies is divided into the same node, and the connection between nodes is based on the macro similarity of regional strata distribution; at the meso scale, the node connection is determined by combining lithology and strata structure characteristics; at the micro scale, the adjacency relationship is constructed according to the similarity of micro properties such as lithology and porosity and spatial distance. The weight of the edge is calculated using an improved Gaussian kernel function, highlighting the connection strength of nodes with high similarity and short distance. In the time dimension, 1 year is taken as the time slice, and nodes at the same spatial location in different periods are connected. The time edge weight is dynamically adjusted according to the change rate of marine environmental parameters, such as when the current speed changes more than a certain threshold in a certain time period, the corresponding time edge weight is reduced. By fusing multi-scale spatial adjacency graph and time adjacency graph, a multi-scale spatio-temporal coupled graph is formed.

[0107] Multi-physical field constrained neural network modeling: A multi-scale spatio-temporal convolutional neural network is constructed. In the spatial dimension, different size convolution kernels are used to extract features from the macro, meso, and micro scale spatial adjacency graphs, capturing the spatial structure characteristics of the strata from the regional to the local. In the time dimension, the gated recurrent unit (GRU) is used to learn the dynamic change pattern of the strata over time in combination with the time adjacency graph. The sediment dynamics and tectonic geology physical field model are embedded in the neural network in the form of soft constraint, and a penalty term based on the difference between the calculated sedimentation rate and tectonic stress and the model prediction results is added to the loss function, ensuring that the model prediction conforms to the physical law. An adaptive attention mechanism is introduced to allow the model to automatically focus on key areas and time periods that have a greater impact on strata evolution, such as areas near active faults and rapid deposition periods.

[0108] Model training and optimization: A joint training strategy combining self-supervised and supervised learning is adopted. The spatio-temporal contrast learning task is designed, such as given the state of the same strata region at different times, the model needs to determine whether it belongs to the same evolution sequence, by maximizing the similarity of positive samples and minimizing the similarity of negative samples, the model's understanding of strata spatio-temporal features is enhanced. Supervised learning uses the known strata evolution history data (obtained through historical earthquake data and drilling core analysis results) of the region as labels, and further optimizes the model parameters by minimizing the error between the predicted results and the true labels. The Adam optimization algorithm is used to adjust the weights and biases of the neural network, allowing the model to continuously converge during training.

[0109] Stratigraphic evolution deduction and result evaluation: input the trained model into the current stratigraphic state of the region and the future 10-year marine environment prediction data (sea level rise rate predicted by global climate change model, sea current mode change, etc.), and the model outputs the stratigraphic evolution prediction results at different time scales, including stratigraphic interface changes, lithology distribution changes, fault activity, etc. The root mean square error is used to measure the stratigraphic thickness prediction error, and the intersection ratio is used to evaluate the accuracy of the lithology distribution prediction. At the same time, geology experts are invited to qualitatively evaluate the prediction results based on physical rationality. Compared with the actual observation data and expert evaluation results, the reliability of the method in the prediction of stratigraphic evolution in the region is improved by 53% compared with the traditional single prediction method.

[0110] Optionally, the embodiment of the application also provides an underwater stratigraphic data deduction device, comprising:

[0111] A data acquisition module is configured to acquire multi-source data of an underwater stratum, pre-process the acquired multi-source data, and construct a multi-modal data set.

[0112] A feature extraction module is configured to acquire a hierarchical spatial adjacency graph and a multi-scale time adjacency graph based on the multi-modal data set, and construct a multi-scale spatio-temporal coupling graph.

[0113] A model construction module is configured to fuse and extract multi-scale spatio-temporal features of the multi-scale spatio-temporal coupling graph based on a multi-scale spatio-temporal convolutional neural network, embed a physical field model in the form of a model loss function into the multi-scale spatio-temporal convolutional neural network, and construct an underwater stratigraphic spatio-temporal deduction model.

[0114] A parameter optimization module is configured to optimize parameters of the underwater stratigraphic spatio-temporal deduction model based on a joint training strategy combining self-supervised learning and supervised learning.

[0115] A prediction module is configured to input a current underwater stratigraphic state and marine environment prediction data for a future period of time into the optimized underwater stratigraphic spatio-temporal deduction model, and output stratigraphic evolution prediction results at different time scales.

[0116] The application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above method. The computer-readable storage medium can include but is not limited to any type of disk, including floppy disks, optical disks, DVDs, CD-ROMs, micro-drives, and magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic or optical cards, nanosystems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.

[0117] It should be noted that, for the foregoing method embodiments, the sequences of the described actions can be changed, and the actions can be performed in other sequences or concurrently. Additionally, it should be understood that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0118] In the above embodiments, the description of each embodiment is focused on different aspects, and the parts not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0119] In the several embodiments provided by the present application, it should be understood that the disclosed apparatus can be implemented in other manners. For example, the described apparatus embodiments are merely schematic. For example, the division of the units is only a logical function division. There can be another division manner for the actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections can be indirect couplings or communication connections through some interfaces, devices or units, and can be in electrical, mechanical or other forms.

[0120] The units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place or distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purposes of the embodiments.

[0121] In addition, each functional unit in the embodiments of the present application can be integrated in a processing unit, or each unit can exist physically as a separate unit, or two or more units can be integrated in one unit. The integrated unit can be implemented in the form of hardware or software functional units.

[0122] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable memory. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a memory and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in the embodiments of the present application. The aforementioned memory includes: a U disk, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0123] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by a program instructing relevant hardware, and the program can be stored in a computer readable memory, which can include a flash disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0124] The above is only exemplary embodiments of the present disclosure, and cannot limit the scope of the present disclosure. That is, any equivalent changes and modifications made in accordance with the teachings of the present disclosure are still within the scope of the present disclosure. Those skilled in the art will easily think of embodiments of the present disclosure after considering the specification and practicing the disclosure herein. The present application is intended to cover any variations, uses or adaptive changes of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or conventional technical means in the technical field not described in the present disclosure. The specification and examples are only considered as exemplary, and the scope and spirit of the present disclosure are defined by the claims.

[0125] The technical features of the above embodiments can be combined in any way. To make the description concise, not all possible combinations of the technical features in the above embodiments are described, however, as long as the combinations of the technical features do not contradict, they should be considered as within the scope of the present disclosure.

[0126] Those skilled in the art readily understand that the above only describes preferred embodiments of the present application and is not intended to limit the present application. Any modification, equivalent replacement and improvement made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for extrapolating underwater stratigraphic data, characterized in that, include: Acquire multi-source data of underwater strata, preprocess the acquired multi-source data, and construct a multimodal dataset; Based on the multimodal dataset, hierarchical spatial adjacency graphs and multi-scale temporal adjacency graphs are obtained, and multi-scale spatiotemporal coupling graphs are constructed. Multi-scale spatiotemporal features of multi-scale spatiotemporal coupling maps are fused and extracted based on multi-scale spatiotemporal convolutional neural networks, and physical field models are embedded into the multi-scale spatiotemporal convolutional neural networks in the form of model loss functions to construct a spatiotemporal inference model of underwater strata. The parameters of the underwater strata spatiotemporal extrapolation model are optimized based on a joint training strategy that combines self-supervised learning and supervised learning. Input the current underwater stratigraphic state and the marine environment prediction data for a period of time into the optimized underwater stratigraphic spatiotemporal extrapolation model, and output the stratigraphic evolution prediction results at different time scales; Specifically, the step of obtaining a hierarchical spatial adjacency graph and a multi-scale temporal adjacency graph based on the multimodal dataset, and constructing a multi-scale spatiotemporal coupled graph, includes: The underwater strata space is divided into voxel units of different scales. Based on regional geological structural characteristics, strata attribute similarity and distance, node connections and edge weights are determined to construct a hierarchical spatial adjacency graph. By connecting nodes at different times in the same spatial location based on time slices, and adjusting the weights of time edges according to time intervals and the rate of change of marine environmental parameters, a multi-scale time adjacency graph is constructed. By fusing spatial and temporal adjacency graphs, a multi-scale spatiotemporal coupling graph is obtained.

2. The underwater stratigraphic data extrapolation method as described in claim 1, characterized in that, The acquisition of multi-source data from underwater strata, followed by preprocessing of the acquired multi-source data to construct a multimodal dataset, specifically includes: Acquire multi-source data of underwater strata, including marine seismic data, logging-while-drilling time-series data, seafloor topography scanning data, seafloor strata scanning data, and marine environmental monitoring data; The multi-source data of the underwater strata were unified into the same spatiotemporal coordinate system using a spatiotemporal alignment algorithm. A completion algorithm combining spatiotemporal variational autoencoders and generative adversarial networks is used to process missing data and construct a multimodal dataset.

3. The underwater stratigraphic data extrapolation method as described in claim 1, characterized in that, The step of embedding the physical field model into the multi-scale spatiotemporal convolutional neural network in the form of a model loss function to construct a spatiotemporal inference model of underwater strata specifically includes: Based on a multi-scale spatiotemporal convolutional neural network, spatial features are extracted by convolutional kernels of different sizes, and temporal variation patterns are learned by combining a variant of a recurrent neural network with a multi-scale temporal adjacency graph. Embedding a physics field model into a multi-scale spatiotemporal convolutional neural network, and adding a constraint term to the loss function to ensure that the model predictions conform to physical laws; An adaptive attention mechanism is introduced to calculate the importance score of spatiotemporal node features, dynamically adjust the information propagation path and weight, and construct a spatiotemporal inference model of underwater strata.

4. The underwater stratigraphic data extrapolation method as described in claim 1, characterized in that, The joint training strategy based on a combination of self-supervised learning and supervised learning optimizes the parameters of the underwater strata spatiotemporal extrapolation model, specifically including: A joint training strategy combining self-supervised and supervised learning is adopted. Self-supervised learning enhances the underwater stratigraphic spatiotemporal inference model’s understanding of stratigraphic spatiotemporal characteristics through a spatiotemporal comparison learning task. Supervised learning utilizes historical stratigraphic data to optimize the parameters of underwater stratigraphic spatiotemporal extrapolation models, employing stochastic gradient descent and its variants to adjust the neural network weights and biases.

5. The underwater stratigraphic data extrapolation method as described in claim 1, characterized in that, The construction of the hierarchical spatial adjacency graph shown also includes: The edge weights are calculated using an improved Gaussian kernel function, as shown in the formula below. in For nodes With nodes edge weights, The distance between nodes. For scale parameters, For nodes With nodes Attribute similarity.

6. The underwater stratigraphic data extrapolation method as described in claim 3, characterized in that, The embedding of the physical field model into a multi-scale spatiotemporal convolutional neural network specifically includes: In terms of sedimentation, the sedimentation rate is calculated using formulas based on parameters such as water flow velocity, sediment grain size, and density, and is then incorporated as a constraint term into the model loss function. In terms of structural motion, by using elasticity mechanics and plate tectonics theory, and by simulating the stress field distribution, a penalty term is added to the loss function to address the difference between the stress calculated based on the physical model and the stress predicted by the model.

7. An underwater stratigraphic data extrapolation device, characterized in that, include: The data acquisition module is used to acquire multi-source data of underwater strata, preprocess the acquired multi-source data, and construct a multimodal dataset. The feature extraction module is used to obtain a hierarchical spatial adjacency graph and a multi-scale temporal adjacency graph based on the multimodal dataset, and to construct a multi-scale spatiotemporal coupled graph. The model building module is used to fuse and extract the multi-scale spatiotemporal features of the multi-scale spatiotemporal coupling map based on the multi-scale spatiotemporal convolutional neural network, and to embed the physical field model into the multi-scale spatiotemporal convolutional neural network in the form of a model loss function to build a spatiotemporal inference model of underwater strata. The parameter optimization module is used to optimize the parameters of the underwater stratum spatiotemporal extrapolation model based on a joint training strategy that combines self-supervised learning and supervised learning. The prediction module is used to input the current underwater strata state and marine environmental prediction data for a future period into the optimized underwater strata spatiotemporal extrapolation model, and output the strata evolution prediction results at different time scales; Specifically, the step of obtaining a hierarchical spatial adjacency graph and a multi-scale temporal adjacency graph based on the multimodal dataset, and constructing a multi-scale spatiotemporal coupled graph, includes: The underwater strata space is divided into voxel units of different scales. Based on regional geological structural characteristics, strata attribute similarity and distance, node connections and edge weights are determined to construct a hierarchical spatial adjacency graph. By connecting nodes at different times in the same spatial location based on time slices, and adjusting the weights of time edges according to time intervals and the rate of change of marine environmental parameters, a multi-scale time adjacency graph is constructed. By fusing spatial and temporal adjacency graphs, a multi-scale spatiotemporal coupling graph is obtained.

8. An underwater stratigraphic data extrapolation device, characterized in that, It includes at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program that, when executed by the processing unit, causes the processing unit to perform the steps of the method according to any one of claims 1 to 6.

9. A storage medium, characterized in that, It stores a computer program that can be executed by an underwater stratigraphic data extrapolation device. When the computer program is run on the underwater stratigraphic data extrapolation device, it causes the underwater stratigraphic data extrapolation device to perform the steps of the method according to any one of claims 1 to 6.

Images