Underground water reserve dynamic early warning method and platform based on digital twinning

Abstract

The invention provides an underground water reserve dynamic early warning method and platform based on digital twinning, and relates to the technical field of digital twinning, and the method comprises the steps: collecting and obtaining the hydrogeological structure information and multi-mode underground water monitoring data of an underground water reserve region; constructing an underground water reserve twinborn model, wherein the underground water reserve twinborn model comprises a plurality of underground water reserve twinborn correlation models of a plurality of hydrogeological unit areas; historical groundwater mining data is called, and dynamic prediction and risk assessment of the groundwater reserve state are carried out in combination with the current groundwater reserve state and the current meteorological data; and presetting a groundwater reserve risk judgment boundary, and if the risk assessment result falls into the groundwater reserve risk judgment boundary, generating a groundwater reserve early warning signal. The technical problem that in the prior art, groundwater management mostly depends on a static model or a simplified mathematical model, complexity of groundwater and interaction between the groundwater and various external factors cannot be fully reflected, and then accurate groundwater early warning is difficult to achieve is solved.

Description

Technical Field

[0001] This invention relates to the field of digital twin technology, specifically to a method and platform for dynamic early warning of groundwater reserves based on digital twins. Background Technology

[0002] Groundwater, as an important water resource, is widely used in various fields such as agricultural irrigation, industrial production, and drinking water supply. However, the management of groundwater resources faces enormous challenges, especially against the backdrop of over-exploitation of groundwater, climate change, pollution, and unscientific use of water resources. Changes in groundwater reserves have a significant impact on the ecological environment.

[0003] Traditional groundwater management systems and early warning methods rely heavily on static or simplified mathematical models. These models cannot fully reflect the complexity of groundwater systems and their interaction with external factors such as climate change and mining activities. In particular, they lack comprehensive simulation and prediction of the dynamic changes in groundwater, which makes it impossible for early warning systems to predict changes in groundwater reserves in a timely manner and to cope with sudden groundwater resource crises, such as water depletion caused by drought or over-exploitation. Summary of the Invention

[0004] This application provides a method and platform for dynamic early warning of groundwater reserves based on digital twins, aiming to solve the technical problem that existing groundwater management technologies rely heavily on static models or simplified mathematical models, which cannot fully reflect the complexity of groundwater and its interaction with various external factors, thus making it difficult to achieve accurate groundwater early warning.

[0005] The first aspect disclosed in this application provides a method for dynamic early warning of groundwater reserves based on digital twins. The method includes: collecting hydrogeological structure information and multimodal groundwater monitoring data of a groundwater reserve area; constructing a groundwater reserve twin model of the groundwater reserve area based on the hydrogeological structure information and multimodal groundwater monitoring data, wherein the groundwater reserve twin model includes multiple groundwater reserve twin association models of multiple hydrogeological unit areas; retrieving historical groundwater extraction data of the groundwater reserve area, combining it with the current groundwater reserve status and current meteorological data, and inputting it into the multiple groundwater reserve twin association models to perform dynamic prediction and risk assessment of the groundwater reserve status; and presetting a groundwater reserve risk judgment boundary, and generating a groundwater reserve early warning signal if the risk assessment result falls within the groundwater reserve risk judgment boundary.

[0006] The second aspect of this application discloses a digital twin-based dynamic early warning platform for groundwater reserves. This platform is used in the aforementioned digital twin-based dynamic early warning method for groundwater reserves. The platform includes: a data acquisition module for acquiring hydrogeological structure information and multimodal groundwater monitoring data of the groundwater reserve area; a model construction module for constructing a groundwater reserve twin model of the groundwater reserve area based on the hydrogeological structure information and multimodal groundwater monitoring data, wherein the groundwater reserve twin model includes multiple groundwater reserve twin association models of multiple hydrogeological unit areas; a risk assessment module for retrieving historical groundwater extraction data of the groundwater reserve area, combining it with the current groundwater reserve status and current meteorological data, and inputting it into the multiple groundwater reserve twin association models to perform dynamic prediction and risk assessment of the groundwater reserve status; and an early warning signal generation module for presetting groundwater reserve risk judgment boundaries, and generating a groundwater reserve early warning signal if the risk assessment result falls within the groundwater reserve risk judgment boundaries.

[0007] One or more technical solutions provided in this application have at least the following beneficial effects: By collecting hydrogeological structure information and multimodal groundwater monitoring data from groundwater reserve areas, comprehensive data support is provided for subsequent model construction and prediction. Integrating data from different sources ensures multi-dimensional data coverage, enabling the model to more accurately reflect the actual situation of groundwater reserves. The groundwater reserve twin model constructed based on hydrogeological structure information and multimodal monitoring data can establish specific groundwater reserve twin correlation models in multiple hydrogeological unit areas. This twin model can dynamically simulate the distribution, changes, and mutual influences of groundwater reserves, comprehensively depicting the dynamic process of groundwater resources. By retrieving historical groundwater extraction data and combining it with current groundwater reserve status and meteorological data, and inputting it into multiple groundwater reserve twin correlation models for dynamic prediction and risk assessment, changes in groundwater reserves can be monitored in real time, and the dynamic characteristics of future groundwater reserves can be predicted, providing a quantitative assessment of potential risks. The preset groundwater reserve risk judgment boundary can automatically generate a groundwater reserve early warning signal when the assessment result falls within a specific risk range. This signal can prompt relevant departments to take timely emergency measures. This early warning mechanism can respond promptly according to changes in the groundwater reserve status, avoiding irreversible resource losses due to delayed response.

[0008] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0009] Figure 1 A schematic diagram of the dynamic early warning method for groundwater reserves based on digital twins provided in this application embodiment.

[0010] Figure 2 A schematic diagram of the structure of the groundwater reserve dynamic early warning platform based on digital twin provided in this application embodiment.

[0011] Figure labeling: Data acquisition module 10, model building module 20, risk assessment module 30, early warning signal generation module 40. Detailed Implementation

[0012] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structure, features and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0013] Example 1, as Figure 1 As shown in the embodiment of this application, a method for dynamic early warning of groundwater reserves based on digital twins is provided, the method comprising: Collect hydrogeological structure information and multimodal groundwater monitoring data of the groundwater reserve area.

[0014] Geological surveys and explorations are conducted to obtain information on the geological characteristics of groundwater reservoir areas, the distribution of groundwater layers, porosity, permeability, and recharge sources. This hydrogeological structural information helps to construct the framework of hydrogeological unit areas. Various types of sensors are deployed for groundwater monitoring to obtain multimodal groundwater monitoring data such as water level, temperature, conductivity, and permeability. Multimodal data can be collected through different monitoring devices, such as remote sensing, hydrological sensors, and groundwater monitoring wells. The obtained data helps to comprehensively assess the current status of groundwater reserves.

[0015] Based on the hydrogeological structure information and multimodal groundwater monitoring data, a groundwater reserve twin model of the groundwater reserve area is constructed, wherein the groundwater reserve twin model includes multiple groundwater reserve twin association models of multiple hydrogeological unit areas.

[0016] Under the concept of digital twins, a virtual simulation model of groundwater reserves is established by combining real-world groundwater reserve areas with digitized 3D models. This model reflects the dynamic behavior of groundwater and enables analysis and prediction. Based on hydrogeological structural information, the groundwater reserve area is divided into multiple hydrogeological unit areas. Each unit area can be simulated by establishing a twin-related model. The constructed groundwater reserve twin model not only reflects the water reserve status of each hydrogeological unit but also integrates the interactions and influences between different units. The models of each hydrogeological unit area are interconnected, forming a comprehensive groundwater reserve twin model.

[0017] Historical groundwater extraction data of the groundwater reserve area is retrieved, and combined with the current groundwater reserve status and current meteorological data, the data is input into the multiple groundwater reserve twin correlation models to perform dynamic prediction and risk assessment of the groundwater reserve status.

[0018] Historical groundwater extraction data includes information such as extraction volume, extraction time, and extraction location. This data helps to understand groundwater extraction behavior and changes in groundwater reserves under different conditions. Current groundwater reserve status, such as water level and volume, is used to assess the current state of groundwater reserves. For example, is the current groundwater reserve sufficient? What is the trend of water level changes? Are there any signs of over-extraction? Current meteorological data, such as precipitation and temperature, has a significant impact on groundwater reserve status. By obtaining real-time meteorological data, the prediction of groundwater reserve status can be supplemented.

[0019] Historical groundwater extraction data, current groundwater reserve status, and current meteorological data are input into a groundwater reserve twin model for dynamic simulation. By simulating groundwater change trends, future reserve conditions are predicted, and these predictions provide a basis for subsequent risk assessments. Based on the model's predicted dynamic characteristics of groundwater, the potential risks to groundwater reserves are assessed, such as the possibility of excessively low water levels or extraction exceeding sustainable limits. Potential risks need to be identified in advance.

[0020] A groundwater reserve risk assessment boundary is preset. If the risk assessment result falls within the groundwater reserve risk assessment boundary, a groundwater reserve early warning signal is generated.

[0021] Based on the different states and risk levels of groundwater reserves, early warning thresholds, i.e., risk assessment boundaries, are set. For example, when the groundwater reserve level drops below a certain critical value, a risk is considered to exist. The groundwater reserve risk assessment boundary can be set based on historical data, geological characteristics, and meteorological conditions. If the risk assessment result of the groundwater reserve exceeds the set risk assessment boundary, an early warning signal is triggered. The early warning signal includes notification, alarm, or activation of automatic control systems to take emergency measures, such as limiting groundwater extraction, increasing recharge sources, or initiating artificial water replenishment, to ensure the sustainable use of groundwater resources.

[0022] Furthermore, this includes: Based on the hydrogeological structure information, the groundwater reserve area is divided into multi-scale hydrogeological grids to generate multiple hydrogeological unit areas with different spatial resolutions; three-dimensional modeling is performed on the multiple hydrogeological unit areas to generate multiple hydrogeological three-dimensional twin models; multimodal sensor arrays are deployed in the multiple hydrogeological unit areas, and time-synchronized groundwater data monitoring is performed to obtain multiple multimodal groundwater monitoring data with location identifiers; based on the location identifiers, the multiple multimodal groundwater monitoring data are synchronized to the multiple hydrogeological three-dimensional twin models to construct a scene and generate multiple groundwater reserve twin association models for the multiple hydrogeological unit areas.

[0023] Based on the hydrogeological structure information of the groundwater reserve area, a decision is made on how to divide the area into multiple hydrogeological unit regions. These unit regions can be divided at multiple scales according to the characteristics of the groundwater reserve. This means that some areas are subdivided using higher-resolution grids for detailed subdivision, while some simpler or geologically homogeneous areas can use lower-resolution grids. For example, higher-resolution grids are used in areas with thin aquifers or complex groundwater flows; lower-resolution grids are used in areas with simple geological structures. Grids for multiple hydrogeological unit regions are generated, with each grid representing a specific hydrogeological unit. These gridded areas will serve as the basic units for subsequent hydrogeological 3D modeling.

[0024] Within each hydrogeological unit area, a corresponding three-dimensional hydrogeological model is constructed based on geological structure information. Modeling software, such as geological modeling software, is used to transform this information into a three-dimensional mesh. Each mesh represents a physical unit of groundwater flow and storage, reflecting groundwater conditions at different depths and locations. Within each hydrogeological unit, groundwater distribution, porosity, permeability, and other properties are considered. These characteristics are used as parameters in the model to simulate the dynamic processes of groundwater flow, storage, and change. These three-dimensional hydrogeological models correspond to actual groundwater storage areas, enabling precise simulation of groundwater behavior in each area within a virtual environment through digital means.

[0025] Based on the aforementioned division of hydrogeological unit areas, various types of sensors are deployed within the groundwater reserve area, such as water level sensors, temperature and humidity sensors, conductivity sensors, and permeability sensors. Each sensor can measure different physical or chemical properties of groundwater, such as water level changes, conductivity, and temperature. These sensors collect data periodically, and the data acquisition from all sensors is synchronized to ensure that data from different sensors are synchronized and can reflect various characteristics of groundwater at the same point in time. Each sensor's data includes a location identifier to correlate the measured values ​​with specific geographical locations and hydrogeological unit areas.

[0026] Multimodal groundwater monitoring data is synchronized with corresponding hydrogeological unit areas in a 3D hydrogeological twin model based on location identifiers. Through spatial mapping and temporal synchronization, it is ensured that the monitoring data at every moment is accurately reflected in the 3D model. By synchronizing these multimodal data into the model, a virtual scene of groundwater reserves is constructed. This virtual scene can display the dynamic changes of different hydrogeological unit areas in time and space, including the changing trends of groundwater reserves and the interactions between units. Finally, by combining all the data and the 3D model, multiple groundwater reserve twin correlation models are formed. Each twin model can accurately reflect the groundwater reserve status of different hydrogeological unit areas and can perform dynamic simulation and prediction.

[0027] Furthermore, based on the aforementioned hydrogeological structure information, the groundwater reserve area is divided into multi-scale hydrogeological grids to generate multiple hydrogeological unit areas with different spatial resolutions, including: The hydrogeological structure information is decomposed at multiple scales to obtain multi-level hydrogeological feature information; based on the multi-level hydrogeological feature information, a high-resolution grid scale and a low-resolution grid scale are generated; based on a preset grid division rule, and in combination with the high-resolution grid scale and the low-resolution grid scale, the groundwater reserve area is divided into hydrogeological grids to generate the multiple hydrogeological unit areas.

[0028] Hydrogeological structural information includes groundwater distribution, permeability, porosity, aquifer thickness, geological faults, and groundwater flow paths. This information reflects the spatial characteristics of groundwater. Multi-scale decomposition is a technique for analyzing regional information at different spatial scales. In this context, multi-scale refers to analyzing hydrogeological structures at different levels, from coarse to fine. For example, low-resolution scales are suitable for relatively simple geological areas, while high-resolution scales are used for areas with complex groundwater flow and storage characteristics. Through multi-scale decomposition, the hydrogeological characteristics of the entire groundwater reservoir area are divided into multiple levels of feature information. For example, the first level of hydrogeological feature information may be the overall distribution of groundwater, while finer levels include details such as the permeability distribution and recharge flow of groundwater layers.

[0029] In groundwater reserve areas, some regions exhibit complex hydrogeological features, such as variations in aquifer depth, faults, and complex groundwater flow directions. High-resolution grids are required to more accurately describe these details. For example, areas with rapid water level changes, uneven groundwater flow, or unique geological structures demand higher spatial resolution to reflect their hydrological dynamics. Conversely, areas with simpler geological structures and less variation can be modeled using low-resolution grids. This not only reduces computational burden but also improves model efficiency. Low-resolution grids are suitable for areas with relatively stable groundwater flow or homogeneous geological features. By selecting appropriate grid scales based on the hydrogeological characteristics of different regions, high-resolution and low-resolution grids can be rationally combined, avoiding over-gridization in simple areas while ensuring sufficient resolution in complex areas.

[0030] During the gridding process, standardized rules are set according to actual needs, such as grid size, shape, and grid density. These rules help ensure that the grid is consistent with the geological characteristics of the groundwater reserve area and meets accuracy requirements. Based on the multi-level feature information and grid scales of different resolutions obtained earlier, rules are used to combine high-resolution and low-resolution grids to perform hydrogeological grid division of the entire area. After grid division, the entire groundwater reserve area is divided into multiple hydrogeological unit areas. Each unit area corresponds to a certain geographic spatial range and hydrogeological characteristics. These unit areas will serve as the basic units for 3D modeling of the groundwater reserve and provide data support for subsequent dynamic simulation and risk assessment.

[0031] Furthermore, it also includes: Based on the multi-level hydrogeological feature information, the hydrogeological attributes of the groundwater reserve area are evaluated to obtain a groundwater reserve spatial distribution map; according to the high-resolution grid scale and low-resolution grid scale, unsupervised clustering and adaptive grid division of the groundwater reserve spatial distribution map are performed to generate the multiple hydrogeological unit regions.

[0032] By analyzing the hydrogeological characteristics of different hydrogeological units, the groundwater properties of each region are assessed. The assessment includes: permeability assessment, which evaluates the permeability of the groundwater layer, affecting groundwater flow velocity and distribution; water storage capacity assessment, which evaluates the water storage capacity of the groundwater reservoir to help determine groundwater reserves; water level change assessment, which evaluates groundwater level conditions in different regions and identifies regional differences in water level changes; and recharge and extraction, which assesses groundwater recharge flow and extraction volume based on geological structure and climatic conditions. Based on the above assessment results, different hydrogeological characteristics are mapped to specific spatial locations, forming a spatial distribution map of groundwater reserve areas. This map shows the spatial distribution of groundwater reserves, such as areas with abundant groundwater and areas with scarce groundwater, areas with high water levels and areas with low water levels, etc., which forms the basis for subsequent model analysis.

[0033] Unsupervised clustering is a data analysis method that uses cluster analysis on the spatial distribution map of groundwater reserves to group areas with similar hydrogeological characteristics into the same category. These clusters are based on attributes such as groundwater reserves, permeability, and recharge. The advantage of unsupervised clustering is that it does not require pre-defined groups or categories; it can automatically group data according to the natural distribution patterns of the data.

[0034] Adaptive mesh generation refers to the process of adjusting the mesh density within a groundwater reserve area based on changes in hydrogeological characteristics. Specifically, for areas with drastic changes in hydrogeological characteristics, such as areas with complex groundwater flow and significant differences in permeability, high-resolution meshes are used. These areas require more precise meshing to better capture the dynamic characteristics of hydrogeological changes. Conversely, for areas with relatively stable or simple hydrogeological characteristics, such as areas with abundant and evenly distributed groundwater reserves, low-resolution meshes are used. This not only saves computational resources but also improves model efficiency. Based on the spatial distribution map of the groundwater reserve and clustering results, the mesh scale is automatically adjusted. For different regions, the mesh generation adaptively changes according to their hydrogeological characteristics to ensure that each hydrogeological unit region is reasonably described.

[0035] By combining unsupervised clustering with adaptive grid partitioning, multiple hydrogeological unit regions are ultimately generated. Each unit region represents a spatial region with specific hydrogeological characteristics, such as a region with high permeability and abundant groundwater, or a region with poor permeability and low water level. These unit regions will serve as the basis for subsequent modeling, monitoring, and prediction.

[0036] Furthermore, it also includes: Based on the multi-level hydrogeological characteristic information, data records of groundwater reserves of the same type are retrieved. These records include groundwater extraction data, aquifer permeability parameters, recharge flow data, and groundwater reserve status information. Using the groundwater extraction data, aquifer permeability parameters, and recharge flow data as input data, and the groundwater reserve status information as output data, a groundwater reserve status simulation model is constructed based on machine learning. A scenario is then built based on the groundwater reserve status simulation model and the multiple hydrogeological three-dimensional twin models.

[0037] Based on multi-level hydrogeological information, regions with similar hydrogeological characteristics to the target area are identified. These similar regions can serve as reference data sources to aid in the establishment of simulation models. Groundwater reserve data records from these similar regions are queried and obtained, specifically including: groundwater extraction data of the same type, reflecting the extraction situation in various groundwater units and helping to understand the rate of groundwater resource consumption and its changes; aquifer permeability, a crucial factor in groundwater flow, affecting its flow velocity and distribution; obtaining permeability parameters of similar aquifers allows for the inference of permeability characteristics in the target area; similar recharge flow data, reflecting the rate and manner of groundwater recharge, helping to understand the rate of groundwater replenishment; the magnitude and stability of the recharge flow directly affect groundwater reserves; and groundwater reserve status information of the same type, including groundwater level, reserves, and current hydrological status in the target area or similar regions, to help establish a dynamic reserve model.

[0038] To construct a groundwater reserve state simulation model, commonly used machine learning algorithms include regression models, time series models, and ensemble learning methods. Before model training, data cleaning and preprocessing, such as normalization and missing value imputation, are performed. Then, existing data is used for training, and the model's performance and generalization ability are evaluated through methods such as cross-validation. By continuously optimizing the model parameters, the prediction results are made as accurate as possible. The main function of the final groundwater reserve state simulation model is to predict the future trend of groundwater reserve changes, thereby providing data support for decision-making.

[0039] Multiple hydrogeological 3D twin models are virtual reconstructions of the spatial and hydrological characteristics of groundwater storage areas. These models can intuitively present the distribution, flow, and storage of groundwater. Groundwater storage state simulation models provide predictions of changes in groundwater storage state. By combining the output of the groundwater storage state simulation models with multiple hydrogeological 3D twin models, comprehensive groundwater storage scenarios are generated. These scenarios can demonstrate the evolution of groundwater storage under different extraction strategies, recharge rates, or climatic conditions.

[0040] Furthermore, dynamic prediction of groundwater reserve status includes: Multiple groundwater reserve states are defined, and a Markov model is constructed. The Markov model is initialized based on the state transition probabilities between the multiple groundwater reserve states. Based on historical groundwater extraction data, current groundwater reserve states, and current meteorological data, continuous extraction simulation is performed on the multiple groundwater reserve twin correlation models. A simulated groundwater reserve dynamic feature sequence is generated based on the extraction simulation results. The simulated groundwater reserve dynamic feature sequence is used as an observation sequence and input into the Markov model. The distribution of groundwater reserve states at several future times is predicted based on a forward algorithm. The multiple groundwater reserve states include high reserve state, low reserve state, and crisis reserve state.

[0041] Several groundwater reserve states are defined, including: High Reserve State: Ample groundwater reserves, high water level, relatively low extraction volume, and large recharge flow; Low Reserve State: Low groundwater reserves, declining water level, requiring increased recharge or reduced extraction volume; Critical Reserve State: Groundwater reserves are at an extremely low level, water level is close to depletion, and there is a potential risk of resource depletion. These definitions are based on factors such as groundwater level, reserves, and extraction volume, and are adjusted according to the specific circumstances of the region in practical applications.

[0042] A Markov model is a mathematical model that predicts future states based on current states. It defines multiple states and determines the probability of state transitions between different states based on historical data. For example, if the current state is a high-reserve state, the probability of transitioning to a low-reserve state is relatively small, and the probability of transitioning to a crisis-reserve state is also relatively low, but the probability of transitioning to a low-reserve state is relatively high. The probability of state transitions is obtained by statistical analysis of historical mining data and hydrological monitoring data.

[0043] The state transition probabilities contain the transition probabilities between all states. These probability values ​​are used to initialize the Markov model so that it can predict the state at the next moment based on the current state.

[0044] Historical groundwater extraction data reflects the extraction patterns and influencing factors of groundwater resources. Current groundwater reserve status reflects the current state of groundwater. Current meteorological data has a significant impact on groundwater recharge; for example, increased precipitation enhances groundwater recharge capacity, while drought reduces recharge. By combining this information, continuous groundwater extraction simulations are conducted to simulate dynamic characteristic sequences of groundwater reserves. These sequences include water level changes, reserve changes, and extraction volume, serving as inputs for subsequent prediction models.

[0045] The generated simulated groundwater reserve dynamic characteristic sequence is used as the observation sequence for the Markov model. These observation sequences contain dynamic change information of groundwater, such as water level changes and reserve changes, and form the basis for model prediction. The forward algorithm, a type of algorithm in the Markov model, is used to calculate the state distribution from the initial state to future times. The specific process is as follows: Based on the current groundwater reserve state, the Markov model is initialized, initial state probabilities are set, and the state transition probabilities and observation sequences are used to recursively calculate the state distribution for future times. Each prediction depends on the state of the previous step and the observation value at the current time. The forward algorithm provides a probability distribution for the groundwater reserve state at each time point, representing the probability of the groundwater being in each state at that time. The final output is the probability distribution of the groundwater reserve state at several future times, used to predict whether the groundwater reserve will enter a low-reserve or critical-reserve state, thereby conducting risk assessment and early warning.

[0046] Furthermore, a risk assessment should be conducted, including: Based on the various groundwater reserve states, define corresponding groundwater reserve risk levels; based on the distribution of groundwater reserve states, conduct risk exposure analysis under the various groundwater reserve risk levels, and generate risk assessment results.

[0047] Groundwater reserve risk level is a quantitative result of groundwater reserve status, used to represent the degree of risk that may exist under different reserve statuses. For example: under high reserve status, the water level is sufficient, the reserves are abundant, and the water resources are sufficient to meet demand in the short term. The risk level under this status is low, defined as low risk level; under low reserve status, the water level is declining, the reserves are decreasing, the extraction is increasing, and the groundwater resources are close to depletion. If no intervention is taken, it may enter a crisis state. The risk level under this status is high, defined as medium risk level; under crisis reserve status, the groundwater resources are extremely scarce, the water level is close to depletion, the extraction is too large, and normal water supply cannot be maintained. At this time, emergency response measures are required, such as water restriction and increased recharge. The risk level under this status is the highest, defined as high risk level.

[0048] Risk exposure analysis mainly assesses the degree of water resource stress and its potential impact under different groundwater reserve states. Based on Markov models and historical data, the distribution of groundwater reserve states at different time points is first determined. For each time point, there will be a groundwater reserve state, such as high reserve, low reserve, and crisis reserve. This distribution reflects the probability of different states occurring.

[0049] Risk exposure analysis relies on several factors: groundwater reserves (the amount of reserves determines the degree of risk exposure; lower reserves mean higher risk); extraction and recharge rates (excessive extraction or insufficient recharge exacerbates groundwater scarcity, increasing risk exposure); regional dependence (some regions are highly dependent on groundwater, resulting in greater risk exposure; for example, some cities heavily reliant on groundwater are more vulnerable to scarcity); and meteorological factors (variations in precipitation and temperature fluctuations affect groundwater reserves, thus influencing risk exposure). Risk assessment results are derived through a comprehensive analysis of risk exposure under different conditions. This helps managers determine the risk level of groundwater reserves and provides a basis for decision-making.

[0050] Furthermore, it also includes: Based on the twin correlation models of multiple groundwater reserves in the multiple hydrogeological unit areas, the hydrogeological correlation between the regions is analyzed; based on the hydrogeological correlation, a multi-regional coordinated emergency response is carried out for the groundwater reserve early warning signal.

[0051] Groundwater reserves are not confined to a single site but consist of multiple hydrogeological units interconnected by factors such as flow, infiltration, recharge, and extraction. By analyzing the relationships between these groundwater reserve twin models, the interactive effects between units can be identified. For example, over-extraction in one area may lead to a drop in water levels in neighboring areas, or water recharge in some areas may affect the reserve status of other areas. For instance, spatiotemporal analysis can be used to analyze the dynamic changes in groundwater across different regions. For example, water level changes in one area under specific meteorological conditions may affect water levels in other areas. Through geological structure and flow mechanisms, methods such as correlation analysis, regression analysis, and graph network analysis can be employed to identify and quantify the degree of correlation between different hydrogeological units.

[0052] Based on hydrogeological correlations, groundwater status information from different regions is integrated to construct a coordinated response mechanism. This mechanism can be centralized or distributed, flexibly responding to the needs and available resources of different regions. For example, when groundwater reserves in a certain area drop to a low or critical state, emergency measures are immediately taken to allocate groundwater resources from surrounding areas for replenishment. For areas where groundwater reserves have already entered a critical state, artificial water replenishment measures can be taken according to the actual situation, such as increasing reservoir water transfers and deep groundwater extraction.

[0053] Furthermore, if the risk assessment result does not fall within the risk determination boundary of the groundwater reserve, a continuous monitoring instruction is generated, and continuous monitoring of the groundwater reserve area is carried out according to the continuous monitoring instruction.

[0054] If a crisis situation is not reached, for example, if the reserves are still at a high or low level, a continuous monitoring instruction will be automatically generated to instruct continued monitoring of the groundwater reserve status. This ensures that the monitoring content reflects the actual situation of the current groundwater reserves in a timely manner. Continuous monitoring not only helps with short-term risk warnings but also provides data support for long-term groundwater reserve management, ensuring that groundwater reserves can be effectively managed and allocated under different risk levels.

[0055] Example 2, based on the same inventive concept as the digital twin-based groundwater reserve dynamic early warning method in the previous examples, such as... Figure 2 As shown in the embodiment of this application, a dynamic early warning platform for groundwater reserves based on digital twins is provided, the platform comprising: The data acquisition module 10 is used to collect hydrogeological structure information and multimodal groundwater monitoring data of the groundwater reserve area; the model construction module 20 is used to construct a groundwater reserve twin model of the groundwater reserve area based on the hydrogeological structure information and multimodal groundwater monitoring data, wherein the groundwater reserve twin model includes multiple groundwater reserve twin association models of multiple hydrogeological unit areas; the risk assessment module 30 is used to retrieve historical groundwater extraction data of the groundwater reserve area, combine it with the current groundwater reserve status and current meteorological data, and input it into the multiple groundwater reserve twin association models to perform dynamic prediction and risk assessment of the groundwater reserve status; the early warning signal generation module 40 is used to preset the groundwater reserve risk judgment boundary, and generate a groundwater reserve early warning signal if the risk assessment result falls within the groundwater reserve risk judgment boundary.

[0056] Furthermore, the model building module 20 is used to perform the following operation steps: Based on the hydrogeological structure information, the groundwater reserve area is divided into multi-scale hydrogeological grids to generate multiple hydrogeological unit areas with different spatial resolutions; three-dimensional modeling is performed on the multiple hydrogeological unit areas to generate multiple hydrogeological three-dimensional twin models; multimodal sensor arrays are deployed in the multiple hydrogeological unit areas, and time-synchronized groundwater data monitoring is performed to obtain multiple multimodal groundwater monitoring data with location identifiers; based on the location identifiers, the multiple multimodal groundwater monitoring data are synchronized to the multiple hydrogeological three-dimensional twin models to construct a scene and generate multiple groundwater reserve twin association models for the multiple hydrogeological unit areas.

[0057] Furthermore, the model building module 20 is used to perform the following operation steps: The hydrogeological structure information is decomposed at multiple scales to obtain multi-level hydrogeological feature information; based on the multi-level hydrogeological feature information, a high-resolution grid scale and a low-resolution grid scale are generated; based on a preset grid division rule, and in combination with the high-resolution grid scale and the low-resolution grid scale, the groundwater reserve area is divided into hydrogeological grids to generate the multiple hydrogeological unit areas.

[0058] Furthermore, the model building module 20 is used to perform the following operation steps: Based on the multi-level hydrogeological feature information, the hydrogeological attributes of the groundwater reserve area are evaluated to obtain a groundwater reserve spatial distribution map; according to the high-resolution grid scale and low-resolution grid scale, unsupervised clustering and adaptive grid division of the groundwater reserve spatial distribution map are performed to generate the multiple hydrogeological unit regions.

[0059] Furthermore, the model building module 20 is used to perform the following operation steps: Based on the multi-level hydrogeological characteristic information, data records of groundwater reserves of the same type are retrieved. These records include groundwater extraction data, aquifer permeability parameters, recharge flow data, and groundwater reserve status information. Using the groundwater extraction data, aquifer permeability parameters, and recharge flow data as input data, and the groundwater reserve status information as output data, a groundwater reserve status simulation model is constructed based on machine learning. A scenario is then built based on the groundwater reserve status simulation model and the multiple hydrogeological three-dimensional twin models.

[0060] Furthermore, the risk assessment module 30 is used to perform the following operational steps: Multiple groundwater reserve states are defined, and a Markov model is constructed. The Markov model is initialized based on the state transition probabilities between the multiple groundwater reserve states. Based on historical groundwater extraction data, current groundwater reserve states, and current meteorological data, continuous extraction simulation is performed on the multiple groundwater reserve twin correlation models. A simulated groundwater reserve dynamic feature sequence is generated based on the extraction simulation results. The simulated groundwater reserve dynamic feature sequence is used as an observation sequence and input into the Markov model. The distribution of groundwater reserve states at several future times is predicted based on a forward algorithm. The multiple groundwater reserve states include high reserve state, low reserve state, and crisis reserve state.

[0061] Furthermore, the risk assessment module 30 is used to perform the following operational steps: Based on the various groundwater reserve states, define corresponding groundwater reserve risk levels; based on the distribution of groundwater reserve states, conduct risk exposure analysis under the various groundwater reserve risk levels, and generate risk assessment results.

[0062] Furthermore, it also includes an emergency response module for performing the following steps: Based on the twin correlation models of multiple groundwater reserves in the multiple hydrogeological unit areas, the hydrogeological correlation between the regions is analyzed; based on the hydrogeological correlation, a multi-regional coordinated emergency response is carried out for the groundwater reserve early warning signal.

[0063] Furthermore, if the risk assessment result does not fall within the risk determination boundary of the groundwater reserve, a continuous monitoring instruction is generated, and continuous monitoring of the groundwater reserve area is carried out according to the continuous monitoring instruction.

[0064] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for dynamic early warning of groundwater reserves based on digital twinning, characterized in that, The method comprises: Collecting hydrogeological structure information and multi-modal groundwater monitoring data of the groundwater reserve area; Based on the hydrogeological structure information and multi-modal groundwater monitoring data, a groundwater reserve twin model of the groundwater reserve area is constructed, wherein the groundwater reserve twin model comprises a plurality of groundwater reserve twin correlation models of a plurality of hydrogeological unit areas; Retrieve the historical groundwater exploitation data of the groundwater reserve area, combine the current groundwater reserve state and the current meteorological data, input the plurality of groundwater reserve twin correlation models to dynamically predict and risk assess the groundwater reserve state; Pre-set groundwater reserve risk judgment boundary, if the risk assessment result falls into the groundwater reserve risk judgment boundary, generate groundwater reserve early warning signal.

2. The digital-twin-based groundwater reserve dynamic early warning method according to claim 1, characterized in that, It includes: According to the hydrogeological structure information, the groundwater reserve area is divided into multiple scales of hydrogeological grid, and the multiple hydrogeological unit areas with different spatial resolution are generated; Three-dimensional modeling is performed on the multiple hydrogeological unit areas to generate multiple hydrogeological three-dimensional twin models; Lay out multi-modal sensing array in the multiple hydrogeological unit areas, and perform time-synchronized groundwater data monitoring to obtain multiple multi-modal groundwater monitoring data with location identification; Based on the location identification, the multiple multi-modal groundwater monitoring data are synchronized to the multiple hydrogeological three-dimensional twin models for scene construction to generate the multiple groundwater reserve twin correlation models of the multiple hydrogeological unit areas.

3. The digital-twin-based groundwater reserve dynamic early warning method according to claim 2, characterized in that, According to the hydrogeological structure information, the groundwater reserve area is divided into multiple scales of hydrogeological grid, and the multiple hydrogeological unit areas with different spatial resolution are generated, which comprises: Multi-scale decomposition is performed on the hydrogeological structure information to obtain multi-level hydrogeological feature information; Based on the multi-level hydrogeological feature information, high-resolution grid scale and low-resolution grid scale are generated; Based on the preset grid division rule, combined with the high-resolution grid scale and low-resolution grid scale, the hydrogeological grid division is performed on the groundwater reserve area to generate the multiple hydrogeological unit areas.

4. The digital-twin-based groundwater reserve dynamic early warning method according to claim 3, characterized in that, It also includes: Based on the multi-level hydrogeological feature information, the hydrogeological attribute of the groundwater reserve area is evaluated to obtain a groundwater reserve spatial distribution map; According to the high-resolution grid scale and low-resolution grid scale, unsupervised clustering and adaptive grid division of the groundwater reserve spatial distribution map are performed to generate the multiple hydrogeological unit areas.

5. The digital-twin-based groundwater reserve dynamic early warning method according to claim 3, characterized in that, It also includes: According to the multi-level hydrogeological feature information, query and obtain the same type of groundwater reserve data record, which includes the same type of groundwater exploitation amount data, the same type of aquifer permeability parameter, the same type of recharge flow data, and the same type of groundwater reserve state information; Taking the same type of groundwater exploitation amount data, the same type of aquifer permeability parameter, and the same type of recharge flow data as input data, and taking the same type of groundwater reserve state information as output data, a groundwater reserve state simulation model is constructed based on machine learning; According to the groundwater reserve state simulation model and the plurality of hydrogeological three-dimensional twin models, a scenario is constructed.

6. The digital-twin-based groundwater reserve dynamic early warning method according to claim 1, characterized in that, Dynamic prediction of the groundwater reserve state is performed, including: a plurality of groundwater reserve states are defined, a Markov model is constructed, and the Markov model is initialized based on state transition probabilities between the plurality of groundwater reserve states; based on the historical groundwater exploitation data, the current groundwater reserve state, and the current meteorological data, continuous exploitation simulation is performed on the plurality of groundwater reserve twin correlation models, and a simulated groundwater reserve dynamic feature sequence is generated according to the exploitation simulation result; the simulated groundwater reserve dynamic feature sequence is input into the Markov model as an observation sequence, and the groundwater reserve state distribution at future time points is predicted based on a forward algorithm; wherein the plurality of groundwater reserve states include a high reserve state, a low reserve state, and a crisis reserve state.

7. The digital-twin-based groundwater reserve dynamic early warning method according to claim 6, characterized in that, Risk assessment is performed, including: According to the plurality of groundwater reserve states, a plurality of groundwater reserve risk levels are defined; According to the groundwater reserve state distribution, risk exposure analysis under the plurality of groundwater reserve risk levels is performed, and a risk assessment result is generated. 8.The digital-twin-based groundwater reserve dynamic early-warning method of claim 1, wherein, Further comprising: According to the plurality of groundwater reserve twin correlation models of the plurality of hydrogeological unit regions, the hydrogeological correlation relationship between the regions is analyzed; Based on the hydrogeological correlation relationship, a multi-region linkage emergency response of the groundwater reserve early warning signal is performed. 9.The digital-twin-based groundwater reserve dynamic early-warning method of claim 1, wherein, If the risk assessment result does not fall within the groundwater reserve risk determination boundary, a continuous monitoring instruction is generated, and continuous monitoring of the groundwater reserve region is performed according to the continuous monitoring instruction.

10. A digital-twin-based groundwater reserve dynamic early-warning platform, characterized in that, The platform for implementing the digital twin-based groundwater reserve dynamic early warning method of any one of claims 1-9, comprising: a data acquisition module for acquiring hydrogeological structure information and multi-modal groundwater monitoring data of a groundwater reserve region; a model construction module for constructing a groundwater reserve twin model of the groundwater reserve region based on the hydrogeological structure information and multi-modal groundwater monitoring data, wherein the groundwater reserve twin model includes a plurality of groundwater reserve twin correlation models of a plurality of hydrogeological unit regions; a risk assessment module for retrieving historical groundwater exploitation data of the groundwater reserve region, combining current groundwater reserve state and current meteorological data, and inputting the plurality of groundwater reserve twin correlation models for dynamic prediction and risk assessment of the groundwater reserve state; an early warning signal generation module for presetting a groundwater reserve risk determination boundary, and generating a groundwater reserve early warning signal if the risk assessment result falls within the groundwater reserve risk determination boundary.

Images