Underground water condition detection method and system based on water flow numerical model simulation

Abstract

The invention discloses an underground water condition detection method and system based on water flow numerical model simulation, and relates to the data processing technology, and the method comprises the steps: building a terrain model containing a target region; determining a proportional relation in the terrain model, and cutting in multiple target directions of the terrain model based on the proportional relation; determining a plurality of representative sites according to the sections, collecting formation data of the representative sites, and mapping the collected formation data to the corresponding sections; fitting is carried out on corresponding sections based on stratums of the mapped representative sites, so that stratigraphic distribution is constructed; constructing a data sequence according to the acquired associated monitoring data of the underground water of the plurality of acquisition points and the rainfall data of the target area; and complementing the groundwater condition data of each section of the island terrain model by using a preset complementing network model according to the data sequence. Groundwater condition detection is realized in combination with a stratum structure model and a network completion mode, and a new method for underground water condition detection is provided.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a groundwater condition detection method and system based on water flow numerical model simulation. Background Art

[0002] Groundwater, with its stable flow and high quality, is a crucial source of water for agricultural irrigation, industrial production, and urban life. This is particularly true in arid and semi-arid regions where surface water is scarce, where groundwater is often the primary source of water. Groundwater resources play a crucial role in regional water supply. Without proper management, overuse can easily lead to subsidence and collapse, resulting in a decline in local water resources and even groundwater pollution, leading to serious environmental problems.

[0003] Traditional groundwater regime estimation is based on various hydraulic tests and numerical models to infer the available water volume of the groundwater system. Modeling is time-consuming and requires a large number of hydrogeological parameters. Due to factors such as complex geological conditions, uncertainty in water regime prediction parameters and the burden of mathematical modeling, a more accurate regional groundwater regime quantitative prediction tool has not yet been formed. It is impossible to accurately predict the groundwater regime in various parts of the region and cannot serve regional water supply management well. Summary of the Invention

[0004] The embodiment of the present application provides a groundwater detection method and system based on water flow numerical model simulation, which is used to realize groundwater detection by combining the stratum structure model and the complete network architecture design, and proposes a new method for groundwater detection.

[0005] The present application provides a method for detecting groundwater conditions based on a water flow numerical model simulation, including: Obtain terrain elevation data and build a terrain model containing the target area; Determining a proportional relationship in the terrain model, and performing sectioning in multiple target directions of the terrain model based on the proportional relationship to obtain an island terrain model of the target area, wherein the section depth at least completely covers the groundwater area; Determining a plurality of representative locations according to the section, collecting stratigraphic data of the representative locations, and mapping the collected stratigraphic data to corresponding sections; Fitting the strata based on the mapped representative locations at corresponding slices to construct a stratum distribution at each slice of the isolated island terrain model, and determining an initial groundwater distribution for each slice of the isolated island terrain model; Constructing a data series based on the associated monitoring data of groundwater at several collection points and the rainfall data of the target area; Complete the groundwater condition data of each section of the isolated island terrain model according to the data sequence by using a preset completion network model, so as to complete the groundwater condition detection.

[0006] Optionally, after obtaining the isolated island terrain model of the target area, it further includes: Determine the formation permeability coefficient distribution according to the formation data of the representative locations collected; Adjust the section depth of each section of the isolated island terrain model based on the formation permeability coefficient distribution, and Extract the geological feature orientation based on the formation data, and section along the fault zone direction preferentially according to the geological feature orientation.

[0007] Optionally, sectioning in multiple target directions of the terrain model based on the proportional relationship includes: setting the target directions of multiple sections so that the sections in multiple target directions are perpendicular to the horizontal plane, and the sections are parallel or intersecting; Fitting based on the mapped formation of the representative locations in the corresponding sections to construct the formation distribution in each section of the isolated island terrain model includes: Perform fitting on the current section according to the adjusted section depth of the section and the extracted geological feature orientation; and Correct the fitting result according to the geological features of the associated sections parallel and perpendicular to the current section.

[0008] Optionally, correcting the fitting result according to the geological features of the associated sections parallel and perpendicular to the current section includes: Extract the formation mutation features of the associated sections parallel and perpendicular to the current section; Based on the formation mutation features, search for similar formation patterns along the direction of the parallel section; Construct a Bayesian probability model and spatial constraints in the direction of the vertical section to calculate the formation extension, where the spatial constraints are used to make the adjacent sections conform to continuity at the junction; Correct the fitting result according to the distribution of the formation extension calculation results based on each associated section.

[0009] Optionally, correcting the fitting result according to the geological features of the associated sections parallel and perpendicular to the current section further includes correcting the karst cave structure through the following steps: For the current section, judge the permeability coefficient mutation feature according to the formation data of the representative locations collected; and At the associated positions in at least two adjacent parallel or intersecting associated sections, identify the similar features of the permeability coefficient mutation feature; Determine the karst cave structure according to the identified similar features.

[0010] Optionally, the preset completion network model includes a spatio-temporal graph convolutional network (ST-GCN) and a temporal convolutional network (TCN). Completing the groundwater condition data of each section of the isolated island terrain model using the preset completion network model according to the data sequence includes: Determine the geographical locations of each collection point and calculate the three-dimensional Euclidean distance between the collection points; Set a dynamic adjacency matrix based on the calculated three-dimensional Euclidean distance; Construct a spatial topology graph according to the dynamic adjacency matrix, where the associated monitoring data of the groundwater at the collection points is used as the node features of the spatial topology graph; Extract the spatio-temporal graph features of the spatial topology graph using spatio-temporal graph convolution (ST-GCN) to perform data completion based on the spatio-temporal graph features.

[0011] Optionally, completing the groundwater condition data of each section of the isolated island terrain model using the completion network model according to the data sequence further includes: Extract rainfall features from the rainfall data of the target area using the temporal convolutional network (TCN); Project the extracted rainfall features and spatio-temporal graph features to a unified dimension; Fuse the rainfall features and spatio-temporal graph features in the unified dimension to complete the groundwater condition data based on the fused features.

[0012] Optionally, completing the groundwater condition data based on the fused features is implemented using a decoder.

[0013] An embodiment of the present application also provides a groundwater condition detection system based on the simulation of a water flow numerical model, including a processor and a memory. A computer program is stored on the memory, and when the computer program is executed by the processor, the steps of the aforementioned groundwater condition detection method based on the simulation of a water flow numerical model are implemented.

[0014] An embodiment of the present application provides a new method for groundwater condition detection. By establishing an isolated island terrain model and combining a completion network model to complete the groundwater condition data of each section of the isolated island terrain model, groundwater condition detection can be performed on multiple sections under the isolated island terrain model, which better serves the groundwater management in urban, town, and rural areas.

[0015] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features, and advantages of the present application more obvious and understandable, the following specifically describes the embodiments of the present invention. Description of the Drawings

[0016] By reading the following detailed description of the preferred embodiments, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present invention. Also, throughout the drawings, the same reference numerals are used to represent the same components. In the drawings: Figure 1 Schematic diagram of the basic process of the groundwater condition detection method for this embodiment; Figure 2 Schematic diagram of the data completion process of the groundwater condition detection method for this embodiment based on the completion network model. Specific embodiments

[0017] Hereinafter, exemplary embodiments of the present disclosure will be described in more detail with reference to the drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be completely conveyed to those skilled in the art.

[0018] An embodiment of the present application proposes a groundwater condition detection method based on the simulation of a water flow numerical model, as Figure 1 shown, including the following steps: In step S101, topographic elevation data is obtained, and a topographic model including the target area is established. For example, a topographic model including the target area can be established according to the topographic elevation data DEM, and the target area is the area where groundwater condition detection is required.

[0019] In step S102, the proportional relationship in the topographic model is determined, and based on the proportional relationship, cuts are made in multiple target directions of the topographic model to obtain the isolated island topographic model of the target area, where the cut depth completely covers at least the groundwater area. That is, cuts are made in multiple directions based on the constructed topographic model, and the specific formation data after cutting is filled according to the data collected subsequently. In this way, a formation structure from the ground surface inward can be established, providing a basis for effectively analyzing and displaying the groundwater situation.

[0020] In step S103, multiple representative locations are determined according to the cuts, and the formation data of the representative locations is collected, and the collected formation data is mapped to the corresponding cuts. In a specific example, existing monitoring wells and other facilities can be used as representative locations, or collection points can be added to collect the formation data of the representative locations. The representative locations can be set distributively, and one representative location can exist in multiple cuts.

[0021] In step S104, the strata representing the locations based on the mapping are fitted on the corresponding cross-sections to construct the strata distribution on each cross-section of the isolated island terrain model, and, determine the initial groundwater distribution for each cross-section of the isolated island terrain model. Through fitting, the strata distribution of the target area can be obtained, and the initial groundwater distribution can be determined for the isolated island terrain model according to the historically explored underground conditions.

[0022] In step S105, a data sequence is constructed based on the associated monitoring data of the groundwater at several acquisition points and the rainfall data of the target area. In a specific example, the associated monitoring data of the groundwater may include the permeability coefficient, the formation water storage rate, the lithology, etc., and also include the monitored water levels at the acquisition points. In a specific example, the groundwater situation is also strongly associated with the regional rainfall. Therefore, in the embodiments of the present application, a data sequence is further constructed based on the rainfall data.

[0023] In step S106, the groundwater situation data on each cross-section of the isolated island terrain model is complemented by using a preset complementation network model according to the data sequence to complete the detection of the groundwater situation. After the complementation, the groundwater situation distribution on multiple cross-sections of the isolated island terrain model can be obtained, thus completing the detection.

[0024] The embodiments of the present application propose a new method for detecting the groundwater situation. By establishing an isolated island terrain model and combining a complementation network model to complement the groundwater situation data on each cross-section of the isolated island terrain model, the groundwater situation of multiple cross-sections under the isolated island terrain model can be detected, which better serves the groundwater management in urban, town, and rural areas.

[0025] In some embodiments, after obtaining the isolated island terrain model of the target area, it further includes: Determine the formation permeability coefficient distribution according to the formation data of the representative locations collected; Adjust the cross-section depth of each cross-section of the isolated island terrain model based on the formation permeability coefficient distribution. By adjusting the cross-section depth, the trend of the groundwater situation can be represented in the isolated island terrain model.

[0026] Extract the geological feature orientation based on the formation data, and perform cross-sections along the fault zone direction preferentially according to the geological feature orientation. In a specific example, the geological feature orientation represents the distribution and trend of different formations underground, and performing cross-sections along the fault zone direction preferentially according to the geological feature orientation can improve the accuracy of capturing the karst cave structure in the karst landform area.

[0027] In some embodiments, performing cross-sections in multiple target directions on the terrain model based on the proportional relationship includes: setting the target directions of multiple cross-sections such that the cross-sections in multiple target directions are perpendicular to the horizontal plane, and the cross-sections are parallel or intersecting with each other. That is, each cross-section is perpendicular to the horizontal plane, and the cross-sections can be in a parallel or intersecting relationship with each other.

[0028] The strata representing locations based on the mapping are fitted at corresponding cross-sections to construct the strata distribution at each cross-section of the isolated island terrain model, including: According to the cross-section depth of the adjusted cross-section and the extracted geological feature orientation, fitting is performed at the current cross-section. In a specific example, three-dimensional fitting can be performed according to the geological feature orientation distribution of discrete acquisition points on multiple cross-sections and the adjusted cross-section depth to obtain the strata situation under the isolated island terrain model.

[0029] The fitting result is corrected according to the geological features of the associated cross-sections parallel and perpendicular to the current cross-section.

[0030] In some embodiments, correcting the fitting result according to the geological features of the associated cross-sections parallel and perpendicular to the current cross-section includes: Extracting the strata mutation features of the associated cross-sections parallel and perpendicular to the current cross-section, such as obvious changes in rock layers and soil layers.

[0031] Based on the strata mutation features, similar strata patterns are searched in the direction parallel to the cross-section. For example, strata with similar parameters such as permeability coefficient and lithology can be searched in the direction parallel to the cross-section according to the data of each acquisition point.

[0032] A Bayesian probability model and spatial constraints are constructed in the direction perpendicular to the cross-section to infer the strata extension, where the spatial constraints are used to make the adjacent cross-sections conform to continuity at the junction. In a specific example, a Bayesian probability model is constructed based on each acquisition point in the direction perpendicular to the cross-section, and the specific spatial constraints can be determined according to the geological feature orientation and the feature continuity at the cross-section intersection line.

[0033] The fitting result is corrected according to the distribution of the strata extension inference results based on each associated cross-section. After correction, different strata can be presented using different colors or structures in the isolated island terrain model and each cross-section.

[0034] In some embodiments, correcting the fitting result according to the geological features of the associated cross-sections parallel and perpendicular to the current cross-section further includes correcting the karst cave structure through the following steps: For the current cross-section, according to the strata data of the representative location collected, the permeability coefficient mutation feature is judged. In a specific example, if there is a continuous strata distribution, there will be no mutation in the permeability coefficient. The possible karst cave structure is detected by judging the mutation of the permeability coefficient.

[0035] At the associated positions in at least two adjacent parallel or intersecting associated cross-sections, the similar features of the permeability coefficient mutation feature are identified, and further the similar features of the mutation are identified at the possible karst cave covering positions in the associated cross-sections.

[0036] Determine the size and scope of the karst cave structure based on the identified similar features.

[0037] In some embodiments, the preset completion network model includes a spatio-temporal graph convolutional network ST-GCN and a temporal convolutional network TCN. The embodiment of the present application designs a completion network model architecture, which is implemented based on the spatio-temporal graph convolutional network ST-GCN and the temporal convolutional network TCN, as Figure 2 shown.

[0038] Completing the groundwater condition data of each section of the isolated island terrain model using the preset completion network model according to the data sequence includes: Determine the geographical locations of each collection point and calculate the three-dimensional Euclidean distance between the collection points; Set a dynamic adjacency matrix based on the calculated three-dimensional Euclidean distance. The adjacency matrix A ij can be defined as when d ij ≤, A ij = exp (-d² ij / σ²) , otherwise it is 0. The parameters σ and R here are determined through seepage tests. For example, σ = 500 meters and R = 2 kilometers.

[0039] Construct a spatial topology graph according to the dynamic adjacency matrix, where the associated monitoring data of the groundwater at the collection points is used as the node features of the spatial topology graph. For example, the associated monitoring data such as the permeability coefficient, formation water storage rate, and water level change is used as the node features.

[0040] Use spatio-temporal graph convolution ST-GCN to extract the spatio-temporal graph features of the spatial topology graph for data completion based on the spatio-temporal graph features. Specifically, graph convolution operations can be performed based on the spatial topology graph.

[0041] In some embodiments, as Figure 2 shown, completing the groundwater condition data of each section of the isolated island terrain model using the completion network model according to the data sequence further includes: Use the temporal convolutional network TCN to extract rainfall features from the rainfall data in the target area. For example, dilated causal convolution can be used to stack dilated convolutional layers (dilation factors = 1, 2, 4, 8), the convolution kernel size is set to 5, and each layer contains a residual connection.

[0042] Project the extracted rainfall features and spatio-temporal graph features into a unified dimension. The output Hg of ST-GCN and the output Ht of TCN can be projected into a unified dimension.

[0043] Fuse the rainfall characteristics of the unified dimension with the spatio-temporal map characteristics to complete the groundwater regime data based on the fused characteristics. In a specific example, the fused characteristics can be generated by calculating the spatio-temporal feature correlation. Then, use the decoder to complete the groundwater regime data based on the fused characteristics. Through the solution of this application, the efficiency and accuracy of groundwater detection can be improved based on ST-GCN and TCN, and the results of water regime detection are presented through the isolated island terrain model, providing a new and efficient auxiliary solution for groundwater management.

[0044] An embodiment of this application also proposes a groundwater regime detection system based on the simulation of a water flow numerical model, including a processor and a memory. A computer program is stored on the memory, and when the computer program is executed by the processor, the steps of the groundwater regime detection method based on the simulation of the water flow numerical model as described above are implemented.

[0045] In addition, although exemplary embodiments have been described herein, their scope includes any and all embodiments based on the present disclosure with equivalent elements, modifications, omissions, combinations (e.g., solutions that cross various embodiments), adaptations, or changes. It is not limited to the examples described in this specification or during the implementation of this application, and the examples will be interpreted as non-exclusive.

[0046] The above description is intended to be illustrative rather than restrictive. For example, the above examples (or one or more of their solutions) can be used in combination with each other. For example, those of ordinary skill in the art can use other embodiments when reading the above description.

[0047] The above embodiments are only exemplary embodiments of the present disclosure. Those skilled in the art can make various modifications or equivalent replacements to the present invention within the essence and protection scope of the present disclosure, and such modifications or equivalent replacements should also be regarded as falling within the protection scope of the present invention.

Claims

1. A groundwater condition detection method based on the simulation of a water flow numerical model, characterized in that, Including: Obtain terrain elevation data and establish a terrain model including the target area; Determine the scale relationship in the terrain model, and perform cross-sections in multiple target directions of the terrain model based on the scale relationship to obtain the isolated island terrain model of the target area, where the cross-section depth completely covers at least the groundwater area; Determine multiple representative locations according to the cross-sections, collect the formation data of the representative locations, and map the collected formation data to the corresponding cross-sections; Fit based on the formation of the mapped representative locations in the corresponding cross-sections to construct the formation distribution in each cross-section of the isolated island terrain model, and determine the initial groundwater distribution for each cross-section of the isolated island terrain model; Construct a data sequence according to the associated monitoring data of groundwater at several acquisition points and the rainfall data of the target area; Complete the groundwater condition detection by using a preset completion network model to complete the groundwater condition data of each cross-section of the isolated island terrain model according to the data sequence.

2. The groundwater condition detection method based on the simulation of the water flow numerical model according to claim 1, wherein After obtaining the isolated island terrain model of the target area, it further includes: Determine the formation permeability coefficient distribution according to the formation data of the collected representative locations; Adjust the cross-section depth of each cross-section of the isolated island terrain model based on the formation permeability coefficient distribution, and Extract the geological feature orientation based on the formation data, and preferentially perform cross-sections along the fault zone direction according to the geological feature orientation.

3. The groundwater condition detection method based on water flow numerical model simulation according to claim 2, characterized in that, Performing cross-sections in multiple target directions of the terrain model based on the scale relationship includes: setting the target directions of multiple cross-sections so that the cross-sections in multiple target directions are perpendicular to the horizontal plane, and the cross-sections are parallel or intersecting; Fitting based on the formation of the mapped representative locations in the corresponding cross-sections to construct the formation distribution in each cross-section of the isolated island terrain model includes: Performing fitting on the current cross-section according to the adjusted cross-section depth of the cross-section and the extracted geological feature orientation; and Correcting the fitting result according to the geological features of the associated cross-sections parallel and perpendicular to the current cross-section.

4. The groundwater condition detection method based on the simulation of the water flow numerical model according to claim 3, wherein Correcting the fitting result according to the geological features of the associated cross-sections parallel and perpendicular to the current cross-section includes: Extracting the formation mutation features of the associated cross-sections parallel and perpendicular to the current cross-section; Searching for similar formation patterns along the parallel cross-section direction based on the formation mutation features; Constructing a Bayesian probability model and spatial constraints in the vertical cross-section direction to deduce the formation extension, where the spatial constraints are used to make the adjacent cross-sections conform to continuity at the junction; Correcting the fitting result according to the distribution of the formation extension deduction results based on each associated cross-section.

5. The groundwater condition detection method based on the simulation of the water flow numerical model according to claim 3, characterized in that, Correcting the fitting result according to the geological features of the associated cross-sections parallel and perpendicular to the current cross-section also includes correcting the karst cave structure through the following steps: For the current cross-section, judge the permeability coefficient mutation features according to the formation data of the collected representative locations; and Identify the similar features of the permeability coefficient mutation features at the associated positions in at least two adjacent parallel or intersecting associated cross-sections; Determine the karst cave structure according to the identified similar features.

6. The groundwater condition detection method based on water flow numerical model simulation according to claim 1, wherein, The preset completion network model includes a spatio-temporal graph convolutional network ST-GCN and a temporal convolutional network TCN; Completing the groundwater condition data of each section of the isolated island terrain model according to the data sequence by using a preset completion network model includes: Determining the geographical locations of each collection point and calculating the three-dimensional Euclidean distance between the collection points; Setting a dynamic adjacency matrix based on the calculated three-dimensional Euclidean distance; Constructing a spatial topology graph according to the dynamic adjacency matrix, wherein the associated monitoring data of the groundwater at the collection points is used as the node features of the spatial topology graph; Using spatio-temporal graph convolution ST-GCN to extract the spatio-temporal graph features of the spatial topology graph for data completion based on the spatio-temporal graph features.

7. The groundwater condition detection method based on the simulation of the water flow numerical model according to claim 6, characterized in that, Completing the groundwater condition data of each section of the isolated island terrain model according to the data sequence by using the completion network model further includes: Using a temporal convolutional network TCN to extract rainfall features from the rainfall data of the target area; Projecting the extracted rainfall features and spatio-temporal graph features into a unified dimension; Fusing the rainfall features and spatio-temporal graph features in the unified dimension to complete the groundwater condition data based on the fused features.

8. The groundwater condition detection method based on the simulation of the water flow numerical model according to claim 6, wherein Completing the groundwater condition data based on the fused features is implemented by using a decoder.

9. A groundwater condition detection system based on water flow numerical model simulation, characterized in that, It includes a processor and a memory, and a computer program is stored on the memory. When the computer program is executed by the processor, the steps of the groundwater condition detection method based on the simulation of the water flow numerical model as described in any one of claims 1 to 8 are implemented.

Images