Intelligent evaluation and decision-making method for geothermal potential of urban underground space

Abstract

The invention discloses an urban underground space geothermal potential intelligent evaluation and development decision-making method, which comprises the following steps of: obtaining multi-source heterogeneous data related to geothermal potential, and preprocessing to obtain standardized spatio-temporal data with a unified spatio-temporal dimension; performing layer-by-layer feature extraction on spatial data corresponding to each time step in the processed standardized spatio-temporal data to obtain a multi-level spatial feature map corresponding to each time step; combining the bidirectional long-short-term memory network and the graph convolution network to obtain space-time correlation field features; inputting physical parameters and space-time correlation field features in a three-dimensional geological knowledge map constructed by the unstructured data and the structured data into a prediction model fused with a physical mechanism to obtain a grading result of geothermal potential and a real-time potential index in an uncertainty interval of the grading result; and obtaining a target geological disaster risk map according to the scoring result, the real-time potential index, the structured data, the unstructured data, the historical disaster data and a plurality of risk assessment models, and further deducing a development decision of the geothermal potential.

Description

Technical Field

[0001] The present invention relates to the field of earth science and resource environment information technology, and relates to, but is not limited to, an intelligent evaluation and development decision-making method for geothermal potential in urban underground space. Background Art

[0002] With the deepening development of urban underground space (such as geothermal mining and foundation pit construction), the potential assessment of geothermal resources as a clean and renewable energy source requires a comprehensive consideration of geological structure, thermophysical parameters, groundwater flow and geological disaster risks (earthquakes, landslides, etc.).

[0003] Related technologies rely on manual experience or a single model, which makes it difficult to cope with the complexity and spatiotemporal dynamics of multi-source data, resulting in insufficient assessment accuracy and lack of risk coupling, and cannot meet the needs of refined urban planning.

[0004] Therefore, how to efficiently and accurately evaluate the geothermal potential of urban underground space and the corresponding development decisions has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, an embodiment of the present invention provides a method for intelligent assessment of geothermal potential and development decision-making in urban underground space, which at least solves the problem that related technologies cannot efficiently and accurately assess the geothermal potential of urban underground space and corresponding development decisions.

[0006] According to a first aspect of an embodiment of the present invention, a method for intelligently assessing geothermal potential and making development decisions in urban underground spaces is provided, comprising: Acquire multi-source heterogeneous data related to geothermal potential and pre-process the multi-source heterogeneous data to obtain standardized spatiotemporal data with unified spatiotemporal dimensions; the multi-source heterogeneous data includes structured data and unstructured data, the structured data includes geological basic data, geophysical data, borehole data, and layer data, and the unstructured data includes geological survey report data and geological description data; Performing layer-by-layer feature extraction on the spatial data corresponding to each time step in the processed standardized spatiotemporal data to obtain a multi-level spatial feature map corresponding to each time step; and based on the multi-level spatial feature map, performing feature extraction and optimization through a bidirectional long short-term memory network and a graph convolutional network to obtain spatiotemporal correlation field features; constructing a three-dimensional geological knowledge map based on the unstructured data and the structured data; and inputting physical parameters extracted from the three-dimensional geological knowledge map and the spatiotemporal correlation field characteristics into a prediction model integrating physical mechanisms to obtain a geothermal potential scoring result and a real-time potential index within an uncertainty interval; the physical parameters including permeability, fault location, and boundary conditions; Based on the scoring results, the real-time potential index, the structured data, the unstructured data and the historical disaster data, a target geological hazard risk map is obtained through multiple risk assessment models, and a geothermal potential development decision is deduced based on the target geological hazard risk map.

[0007] According to a second aspect of an embodiment of the present invention, there is provided an electronic device comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; the memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform an operation corresponding to the method described in the first aspect.

[0008] According to a third aspect of an embodiment of the present invention, a computer storage medium is provided, on which a computer program is stored. When the program is executed by a processor, the method according to the first aspect is implemented.

[0009] According to the solution provided by the embodiment of the present invention, multi-source heterogeneous data related to geothermal potential is obtained, and the multi-source heterogeneous data is preprocessed to obtain standardized spatiotemporal data with unified spatiotemporal dimensions; the multi-source heterogeneous data includes structured data and unstructured data, the structured data includes geological basic data, geophysical data, drilling data and layer data, and the unstructured data includes geological survey report data and geological description data; the spatial data corresponding to each time step in the processed standardized spatiotemporal data is subjected to layer-by-layer feature extraction to obtain a multi-level spatial feature map corresponding to each time step; and based on the multi-level spatial feature map, feature extraction is performed through a bidirectional long short-term memory network and a graph convolution network. The process extracts and optimizes spatiotemporal correlation field features. A three-dimensional geological knowledge graph is constructed based on the unstructured and structured data. The physical parameters extracted from the 3D geological knowledge graph and the spatiotemporal correlation field features are then input into a prediction model that integrates physical mechanisms to obtain a geothermal potential score and a real-time potential index within its uncertainty interval. The physical parameters include permeability, fault location, and boundary conditions. Based on the score, the real-time potential index, the structured and unstructured data, and historical disaster data, a target geohazard risk map is generated using multiple risk assessment models. Development decisions regarding geothermal potential are then inferred based on the target geohazard risk map. This process integrates structured (geological, geophysical, borehole, and map) and unstructured data (survey reports and geological descriptions) to enhance information coverage and reduce biases introduced by single data sources. Layer-by-layer convolution operations are used to extract spatial feature maps, capturing geothermal geological structural characteristics from local to global perspectives. Combining a bidirectional long short-term memory network with a graph convolutional network, the spatiotemporal correlation field features are constructed, effectively simulating the temporal evolution of geothermal systems and their spatial interactions. Key physical parameters such as lithologic thermal conductivity, permeability, fault location, and boundary conditions are extracted from the constructed three-dimensional geological knowledge map and used as important inputs to the prediction model, making the prediction results more physically meaningful. Introducing physical equations or constraints based on deep learning gives the model both powerful fitting capabilities and conformity to geophysical laws. The output of the scoring results is accompanied by uncertainty intervals, which facilitates risk assessment and resource exploration prioritization. Combining historical disaster data with multiple risk assessment models, a target geological hazard risk map is generated to identify potential high-risk areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive work, among which: Figure 1A flow chart of a method for intelligently assessing and developing geothermal potential in urban underground space provided by an embodiment of the present invention; Figure 2 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0011] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. The following embodiments are used to illustrate the present invention, but are not used to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0012] In the following description, reference is made to “some embodiments”, which describes a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0013] It should be pointed out that the terms "first\second\third" involved in the embodiments of the present invention are only used to distinguish similar objects and do not represent a specific ordering of the objects. It can be understood that "first\second\third" can be interchanged with a specific order or sequence where permitted, so that the embodiments of the present invention described here can be implemented in an order other than that illustrated or described here.

[0014] Those skilled in the art will understand that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by those skilled in the art in the art to which the embodiments of the present invention pertain. It should also be understood that terms such as those defined in common dictionaries should be understood to have meanings consistent with their meanings in the context of the prior art and, unless specifically defined as herein, should not be interpreted in an idealized or overly formal sense.

[0015] Figure 1 A flow chart of a method for intelligently evaluating and making development decisions about the geothermal potential of urban underground space provided by an embodiment of the present invention. The method for intelligently evaluating and making development decisions about the geothermal potential of urban underground space provided by an embodiment of the present invention can be executed by electronic devices, such as computers, servers, etc.

[0016] like Figure 1 As shown, a method for intelligent assessment and development decision-making of geothermal potential in urban underground space includes: S101. Acquire multi-source heterogeneous data related to geothermal potential, and pre-process the multi-source heterogeneous data to obtain standardized spatiotemporal data with unified spatiotemporal dimensions; the multi-source heterogeneous data include structured data and unstructured data, the structured data include geological basic data, geophysical data, drilling data and layer data, and the unstructured data include geological survey report data and geological description data.

[0017] In embodiments of the present invention, geothermal potential refers to the total amount or capacity of geothermal energy that can be developed and utilized in a specific region. Multi-source heterogeneous data related to geothermal potential is obtained. Multi-source heterogeneous data includes structured data and unstructured data. Structured data includes basic geological data and geophysical data. Basic geological data includes data such as stratum lithology, thermal conductivity, permeability, and geological structure. Geophysical data includes data such as gravity anomalies and resistivity profiles. Structured data also includes borehole data and layer data. Unstructured data includes geological survey report data and geological description data.

[0018] Furthermore, the multi-source heterogeneous data sets are cleaned, format converted, and feature engineered to obtain standardized spatiotemporal data with unified spatiotemporal dimensions.

[0019] S102. Perform layer-by-layer feature extraction on the spatial data corresponding to each time step in the processed standardized spatiotemporal data to obtain a multi-level spatial feature map corresponding to each time step; and based on the multi-level spatial feature map, perform feature extraction and optimization through a bidirectional long short-term memory network and a graph convolutional network to obtain spatiotemporal correlation field features.

[0020] In an embodiment of the present invention, after obtaining standardized spatiotemporal data, further preprocessing is performed to obtain processed standardized spatiotemporal data. The processed standardized spatiotemporal data is cleaned, interpolated, and normalized, and has a uniform time step and spatial resolution. For the spatial data corresponding to each time step, a deep learning model (such as a three-dimensional convolutional neural network) can be used to extract geological features layer by layer from low to high levels, generating a feature map containing information at multiple scales, such as micro, meso, and macro scales. Finally, a multi-level spatial feature map corresponding to each time step is obtained. Spatiotemporal correlation field features refer to high-level features that integrate dual temporal and spatial dependencies. Spatiotemporal correlation field features are further derived based on the multi-level spatial feature map, a bidirectional long short-term memory network, and a graph convolutional network.

[0021] S103. Construct a three-dimensional geological knowledge map based on unstructured data and structured data; and input the physical parameters and spatiotemporal correlation field characteristics extracted from the three-dimensional geological knowledge map into a prediction model that integrates physical mechanisms to obtain a scoring result of geothermal potential and a real-time potential index within an uncertainty interval; physical parameters include permeability, fault location, and boundary conditions.

[0022] In an embodiment of the present invention, the real-time potential index provides a real-time quantitative score of geothermal potential. By integrating unstructured and structured data, graph database technologies (such as Neo4j) can be used to establish nodes (representing geological entities such as ore bodies, faults, and groundwater systems) and edges (representing relationships between entities, such as adjacency and material migration paths), forming a three-dimensional geological knowledge graph. Geophysical parameters such as thermal conductivity and permeability are then extracted from the 3D geological knowledge graph. These extracted spatiotemporal correlation field features, reflecting temporal evolution and spatial correlation, are then input into a prediction model that incorporates the physical mechanisms of the geothermal system. Ultimately, a comprehensive score of the region's geothermal potential is output. Taking into account various uncertainties, the geothermal potential index and its possible fluctuation range are calculated and displayed in real time. Physical parameters include permeability, fault location, and boundary conditions, as well as formation lithology and thermal conductivity.

[0023] S104. Based on the scoring results, the real-time potential index, the structured data, the unstructured data, and the historical disaster data, a target geological hazard risk map is obtained through multiple risk assessment models, and a development decision of the geothermal potential is deduced based on the target geological hazard risk map.

[0024] In an embodiment of the present invention, based on the obtained geothermal potential scoring results, the real-time potential index within the uncertainty interval, structured data, unstructured data and historical disaster data, and combined with multiple geological hazard risk assessment models (such as fault activity, induced earthquake risk, ground subsidence possibility and other risk assessment models), a comprehensive target geological hazard risk map of the region is generated; on this basis, the developability and priority of geothermal resources in different regions are further deduced to form scientific, dynamic and risk-controllable geothermal development decision-making recommendations.

[0025] The target geological hazard risk map includes high-risk, medium-risk, and low-risk areas. Multiple risk assessment models can be used, such as logistic regression models and random forest models. Each model outputs a single geological hazard risk map. Finally, the single geological hazard risk models are fused to produce the target geological hazard risk map.

[0026] It can be understood that in the embodiment of the present invention, multi-source heterogeneous data related to geothermal potential is obtained, and the multi-source heterogeneous data is preprocessed to obtain standardized spatiotemporal data with unified spatiotemporal dimensions; the multi-source heterogeneous data includes structured data and unstructured data, the structured data includes geological basic data, geophysical data, drilling data and layer data, and the unstructured data includes geological survey report data and geological description data; the spatial data corresponding to each time step in the processed standardized spatiotemporal data is subjected to layer-by-layer feature extraction to obtain a multi-level spatial feature map corresponding to each time step; and based on the multi-level spatial feature map, a bidirectional long short-term memory network and a graph are used to extract the spatial data corresponding to each time step. A convolutional network extracts and optimizes features to obtain spatiotemporal correlation field characteristics. A three-dimensional geological knowledge graph is constructed based on unstructured and structured data. The physical parameters and spatiotemporal correlation field characteristics extracted from the 3D geological knowledge graph are then input into a prediction model that integrates physical mechanisms to obtain a geothermal potential score and a real-time potential index within the uncertainty interval. Physical parameters include permeability, fault location, and boundary conditions. A target geohazard risk map is generated based on the score, real-time potential index, structured data, unstructured data, historical disaster data, and multiple risk assessment models. Development decisions based on the target geohazard risk map are then inferred based on the target geohazard risk map. This process integrates structured (geological, geophysical, borehole, and map) and unstructured data (survey reports and geological descriptions) to enhance information coverage and reduce bias introduced by single data sources. Layer-by-layer convolution operations are used to extract spatial feature maps, capturing geothermal geological structural characteristics from local to global perspectives. Combining a bidirectional long short-term memory network with a graph convolutional network, the spatiotemporal correlation field features are constructed, effectively simulating the temporal evolution of geothermal systems and their spatial interactions. Key physical parameters such as permeability, fault location, and boundary conditions are extracted from the constructed three-dimensional geological knowledge map and used as important inputs to the prediction model, making the prediction results more physically meaningful. Introducing physical equations or constraints based on deep learning gives the model both powerful fitting capabilities and conformity to geophysical laws. The output of the scoring results is accompanied by uncertainty intervals, which facilitates risk assessment and resource exploration prioritization. Combining historical disaster data with multiple risk assessment models, a target geological hazard risk map is generated to identify potential high-risk areas.

[0027] In some embodiments of the present invention, S101 may be implemented through S1011 to S1012, which is described in the following steps.

[0028] S1011. Clean, detect and repair the multi-source heterogeneous data set to obtain first data, and uniformly project the first data to the same spatial coordinates to obtain second data.

[0029] S1012. Resample and interpolate the time series in the second data to obtain third data; and calculate derivative features and standardize the third data to obtain standardized spatiotemporal data.

[0030] In some embodiments of the present invention, multi-source heterogeneous data is cleaned to remove duplicate data, followed by further outlier detection and repair to obtain first data. This first data is then uniformly converted to the same spatial coordinates to obtain second data. The second data is resampled to a daily time step, and missing values ​​are repaired using cubic spline interpolation to achieve data uniformity in the temporal dimension, resulting in third data. Derived features are calculated from the third data, such as heat flux density calculated from thermal conductivity and temperature gradient. The third data is then Z-score normalized to eliminate the dimensionality of numerical features, ultimately generating standardized spatiotemporal data encompassing spatial, temporal, and feature dimensions.

[0031] In some embodiments of the present invention, S20 is included before S102, which is explained through the following steps.

[0032] S20, dividing the standardized spatiotemporal data into three-dimensional grids, and performing interpolation and time step alignment on the divided three-dimensional grids using an inverse distance weighted interpolation method to obtain processed standardized spatiotemporal data.

[0033] In some embodiments of the present invention, the standardized spatiotemporal data is divided into a three-dimensional grid, and the grid nodes in the three-dimensional grid are interpolated and time-step aligned in spatial and temporal dimensions using the inverse distance weighted interpolation method, thereby obtaining structured and time-synchronized processed standardized spatiotemporal data.

[0034] In some embodiments of the present invention, S103 may be implemented through S1031 to S1033, which is described in the following steps.

[0035] S1031. Identify geological entities from unstructured data, classify the geological entities, and obtain classification results.

[0036] S1032. Identify semantic relationships between geological entities based on the classification results; and establish spatial relationships based on the semantic relationships and structured data.

[0037] S1033. Construct a three-dimensional geological knowledge map based on geological entities and their relationships.

[0038] In some embodiments of the present invention, geological entities (such as rock formations, faults, ore bodies, etc.) can be identified from unstructured data based on the conditional random field model of the bidirectional Transformer encoder, the identified geological entities can be classified using the geological science markup language ontology, and the semantic relationships between geological entities can be identified based on the classification results, such as the causal relationship of "high permeability rock formations promoting heat flow", and spatial relationships can be established based on semantic relationships and structured data, such as "faults cutting strata", and finally, a three-dimensional geological knowledge map can be constructed based on the geological entities and spatial relationships.

[0039] In some embodiments of the present invention, obtaining spatiotemporal correlation field features based on the multi-level spatial feature map, bidirectional long short-term memory network and graph convolutional network in S102 can be implemented through S1021 to S1023, which is explained by the following steps.

[0040] S1021. Fusing multi-level spatial feature maps to obtain a fused multi-scale spatial feature map, and processing the multi-scale spatial feature map through the spatiotemporal attention layer in the bidirectional long short-term memory network to obtain dynamic evolution features.

[0041] In some embodiments of the present invention, a feature pyramid network can be used to fuse multi-level spatial feature maps to generate a fused multi-scale spatial feature map. This multi-scale spatial feature map is then processed using a bidirectional long short-term memory network with an integrated spatiotemporal attention mechanism. This network focuses on key time steps and spatial regions, identifies the dynamic changes in parameters such as temperature and flow rate, and generates dynamic evolution features.

[0042] S1022. Calculate the correlation coefficients between the grids in the three-dimensional grid using the dynamic evolution characteristics to obtain an initial correlation matrix.

[0043] In some embodiments of the present invention, the Pearson correlation coefficient (Pearson correlation coefficient) between each grid in the three-dimensional grid is calculated by dynamically evolving characteristics, an initial correlation matrix is ​​constructed by the calculated Pearson correlation coefficient, and the spatiotemporal correlation field characteristics are further obtained by the initial correlation matrix and the graph convolutional network.

[0044] S1023. Obtain an optimized global correlation matrix through the graph convolutional network and the initial correlation matrix, and identify regions with similar spatiotemporal behaviors in the optimized global correlation matrix to obtain spatiotemporal correlation field features.

[0045] In some embodiments of the present invention, the initial matrix is ​​optimized using a graph convolutional network, taking into account the spatial adjacency and feature similarity between regions to obtain an optimized global correlation matrix. This optimized global correlation matrix identifies regions that exhibit similar behavior patterns in time and space. These regions with similar spatiotemporal behavior are then considered to be strongly correlated and serve as spatiotemporal correlation field features.

[0046] In some embodiments of the present invention, S104 may be implemented through S1041 to S1042, which is described in the following steps.

[0047] S1041. Input the scoring results, real-time potential index, structured data, unstructured data and historical disaster data into multiple risk assessment models respectively to obtain multiple single disaster assessment results.

[0048] S1042. A plurality of single disaster assessment results are integrated according to preset rules to obtain a comprehensive risk index; and a target geological disaster risk map is obtained based on the comprehensive risk index.

[0049] In some embodiments of the present invention, the scoring results, real-time potential index, structured data, unstructured data and historical disaster data are respectively input into multiple risk assessment models to obtain multiple single disaster assessment results (such as fault activity, induced earthquake risk, ground subsidence possibility, etc.). The single disaster assessment results can be further fused through the Dempster combination rule to generate a comprehensive risk index. According to the calculated comprehensive risk index, the calculated comprehensive risk index is applied to the corresponding geographical location using geographic information system software to generate a geological disaster risk distribution map, set thresholds to divide the study area into different risk levels (such as low, medium and high risks), select appropriate color schemes and symbol systems to represent risk areas of different levels, and add necessary map elements to finally obtain the target geological disaster risk map.

[0050] In an embodiment of the present invention, on a three-dimensional visualization platform, based on web 3D graphics technology, geological structures such as strata, faults, and heat reservoirs are constructed as three-dimensional models, supporting users to interactively cut and view the internal structure; the distribution of geothermal potential is displayed in the form of a heat map, and the target geological hazard risk map is superimposed on it to intuitively reflect the potential and risk levels of different areas; through dynamic simulation functions, the changes in parameters such as temperature field and groundwater level after the implementation of different mining plans are displayed.

[0051] Reference Figure 2 , shows a schematic structural diagram of an electronic device according to an embodiment of the present invention. The specific embodiment of the present invention does not limit the specific implementation of the electronic device.

[0052] like Figure 2As shown, the electronic device may include: a processor (processor) 502, a communications interface (Communications Interface 504), a memory (memory) 506, and a communication bus 508.

[0053] in: The processor 502 , the communication interface 504 , and the memory 506 communicate with each other via a communication bus 508 .

[0054] The communication interface 504 is used to communicate with other electronic devices or servers.

[0055] The processor 502 is configured to execute the program 510 , and specifically may execute the relevant steps in the above method embodiment.

[0056] Specifically, the program 510 may include program codes, which include computer operation instructions.

[0057] Processor 502 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention. The one or more processors included in a smart device may be of the same type, such as one or more CPUs, or different types, such as one or more CPUs and one or more ASICs.

[0058] The memory 506 is used to store the program 510. The memory 506 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage.

[0059] The program 510 may be specifically configured to enable the processor 502 to execute operations corresponding to the methods described in the above method embodiments.

[0060] The specific implementation of each step in program 510 can be found in the corresponding descriptions of the corresponding steps and units in the above-mentioned method embodiments, and will not be repeated here. Those skilled in the art will clearly understand that for the convenience and brevity of description, the specific working processes of the above-mentioned devices and modules can refer to the corresponding process descriptions in the above-mentioned method embodiments, and will not be repeated here.

[0061] It should be pointed out that, according to the needs of implementation, the various components / steps described in the embodiments of the present invention can be split into more components / steps, or two or more components / steps or partial operations of components / steps can be combined into new components / steps to achieve the purpose of the embodiments of the present invention.

[0062] The methods according to the embodiments of the present invention described above can be implemented in hardware, firmware, or as software or computer code that can be stored on a recording medium (such as a CD ROM, RAM, floppy disk, hard disk, or magneto-optical disk), or as computer code originally stored on a remote recording medium or non-transitory machine-readable medium downloaded over a network and then stored on a local recording medium. Thus, the methods described herein can be processed by such software stored on a recording medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware (such as an ASIC or FPGA). It will be understood that a computer, processor, microprocessor controller, or programmable hardware includes a storage component (e.g., RAM, ROM, flash memory, etc.) that can store or receive software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the methods described herein are implemented. Furthermore, when a general-purpose computer accesses the code for implementing the methods described herein, the execution of the code transforms the general-purpose computer into a dedicated computer for performing the methods described herein.

[0063] Those skilled in the art will appreciate that the units and method steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the embodiments of the present invention.

[0064] The above implementation methods are only used to illustrate the embodiments of the present invention, and are not intended to limit the embodiments of the present invention. Ordinary technicians in the relevant technical field may make various changes and modifications without departing from the spirit and scope of the embodiments of the present invention. Therefore, all equivalent technical solutions also fall within the scope of the embodiments of the present invention, and the scope of patent protection of the embodiments of the present invention should be defined by the claims.

Claims

1. A method for intelligent assessment and development decision-making of geothermal potential in urban underground space, characterized in that: include: Acquire multi-source heterogeneous data related to geothermal potential and pre-process the multi-source heterogeneous data to obtain standardized spatiotemporal data with unified spatiotemporal dimensions; the multi-source heterogeneous data includes structured data and unstructured data, the structured data includes geological basic data, geophysical data, borehole data, and layer data, and the unstructured data includes geological survey report data and geological description data; Performing layer-by-layer feature extraction on the spatial data corresponding to each time step in the processed standardized spatiotemporal data to obtain a multi-level spatial feature map corresponding to each time step; and based on the multi-level spatial feature map, performing feature extraction and optimization through a bidirectional long short-term memory network and a graph convolutional network to obtain spatiotemporal correlation field features; constructing a three-dimensional geological knowledge map based on the unstructured data and the structured data; and inputting physical parameters extracted from the three-dimensional geological knowledge map and the spatiotemporal correlation field characteristics into a prediction model integrating physical mechanisms to obtain a geothermal potential scoring result and a real-time potential index within an uncertainty interval; the physical parameters including permeability, fault location, and boundary conditions; Based on the scoring results, the real-time potential index, the structured data, the unstructured data and the historical disaster data, a target geological hazard risk map is obtained through multiple risk assessment models, and a geothermal potential development decision is deduced based on the target geological hazard risk map.

2. The method according to claim 1, characterized in that The preprocessing of the multi-source heterogeneous data to obtain standardized spatiotemporal data with unified spatiotemporal dimensions includes: Cleaning, outlier detection, and repairing the multi-source heterogeneous data to obtain first data, and uniformly projecting the first data to the same spatial coordinates to obtain second data; Resampling and interpolation processing are performed on the time series in the second data to obtain third data; and derivative feature calculation and standardization processing are performed on the third data to obtain the standardized spatiotemporal data.

3. The method according to claim 1, characterized in that Before extracting layer-by-layer features from the spatial data corresponding to each time step in the processed standardized spatiotemporal data to obtain a multi-level spatial feature map corresponding to each time step, the method further includes: The standardized spatiotemporal data is divided into three-dimensional grids, and the divided three-dimensional grids are interpolated and time-step aligned using an inverse distance weighted interpolation method to obtain the processed standardized spatiotemporal data.

4. The method according to claim 1, wherein The constructing of a three-dimensional geological knowledge map based on the unstructured data and the structured data includes: Identifying geological entities from the unstructured data and classifying the geological entities to obtain classification results; Identifying semantic relationships between the geological entities based on the classification results; and establishing spatial relationships based on the semantic relationships and the structured data; The three-dimensional geological knowledge map is constructed based on the geological entities and the spatial relationships.

5. The method according to claim 1, wherein Based on the multi-level spatial feature map, feature extraction and optimization are performed through a bidirectional long short-term memory network and a graph convolutional network to obtain spatiotemporal correlation field features, including: Fusing the multi-level spatial feature maps to obtain a fused multi-scale spatial feature map, and processing the multi-scale spatial feature map through the spatiotemporal attention layer in the bidirectional long short-term memory network to obtain a dynamic evolution feature; Calculating the correlation coefficients between the grids in the three-dimensional grid using the dynamic evolution characteristics to obtain an initial correlation matrix; An optimized global correlation matrix is ​​obtained through the graph convolutional network and the initial correlation matrix, and regions with similar spatiotemporal behaviors in the optimized global correlation matrix are identified to obtain the spatiotemporal correlation field features.

6. The method according to any one of claims 1 to 5, characterized in that The step of obtaining a target geological hazard risk map through multiple risk assessment models based on the scoring result, the real-time potential index, the structured data, the unstructured data, and the historical disaster data includes: Inputting the scoring results, the real-time potential index, the structured data, the unstructured data, and the historical disaster data into the multiple risk assessment models to obtain multiple single disaster assessment results; The multiple single disaster assessment results are integrated according to preset rules to obtain a comprehensive risk index; and the target geological disaster risk map is obtained based on the comprehensive risk index.

Images