A soil resistivity three-dimensional reconstruction method, system, computer device and medium

Abstract

The application provides a soil resistivity three-dimensional reconstruction method, system, computer device and medium, belongs to the field of infrastructure exploration, and comprises the following steps: obtaining two-dimensional resistivity survey line data and shallow resistivity measurement data, and fusing to obtain standardized resistivity data; performing geological adaptive discrete state interval division on the standardized resistivity data, and determining a plurality of resistivity state intervals; calculating the transition probability between the plurality of resistivity state intervals to obtain a transition probability matrix; performing three-dimensional grid subdivision on the underground space of the surrounding area of a transformer substation to obtain blank grid cells; interpolating and filling the standardized resistivity data to obtain a soil resistivity initial model; performing spatial constraint on the soil resistivity initial model based on the transition probability matrix, and performing iterative inversion calculation to obtain a soil three-dimensional resistivity model. The method overcomes the precision limitation problem caused by neglecting spatial continuity in the prior art, and guarantees the accuracy of the soil three-dimensional resistivity model constructed by inversion.

Description

Technical Field

[0001] This invention belongs to the field of infrastructure exploration, specifically relating to a method, system, computer equipment, and medium for three-dimensional reconstruction of soil resistivity. Background Technology

[0002] In the field of substation engineering, the three-dimensional spatial distribution data of soil resistivity is a core technical basis for substation site selection, grounding grid design optimization, and soil corrosion risk assessment. Based on relevant industry standards, substation grounding grid design requires accurate acquisition of soil resistivity distribution data within the 0-60m depth range. Furthermore, when measuring soil resistivity at the site, the maximum distance between adjacent electrodes should not be less than the maximum diagonal of the grounding grid to ensure that the grounding resistance meets safety standards and controls project costs. However, existing direct measurement techniques for soil resistivity detection are constrained by site conditions. Urban substations are surrounded by dense buildings and complex terrain, limiting the range of measurement lines and making it impossible to achieve full depth coverage from 0-60m underground. Data continuity is also poor, making it difficult to meet the requirements of high-resolution resistivity tomography inversion.

[0003] To address the aforementioned shortcomings, such as Figure 1 As shown, existing technologies often employ a process of direct measurement, two-dimensional inversion, and empirical interpolation to analyze soil resistivity in substation working areas. For example, traditional resistivity measuring equipment is used to deploy two-dimensional survey lines within the substation site for data collection. Traditional inversion tools such as RES2DINV are then used to invert the data from a single survey line based on the least squares algorithm, generating a resistivity profile along the survey line direction. Multiple two-dimensional profiles are then stitched together at horizontal distances, and resistivity values ​​for blank areas are inferred using engineers' experience, forming a rough three-dimensional distribution diagram. However, this method of substation soil resistivity analysis does not consider the spatial continuity of soil resistivity, cannot reflect the true stratigraphic distribution, and has limited accuracy in the inversion results. Summary of the Invention

[0004] To address the problem that existing soil resistivity analysis methods neglect the spatial continuity of soil resistivity and have limited accuracy in inversion results, this invention provides a method, system, computer equipment, and medium for three-dimensional reconstruction of soil resistivity.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for three-dimensional reconstruction of soil resistivity includes:

[0007] Two-dimensional resistivity survey data and shallow resistivity measurement data of the soil in the area surrounding the substation were obtained; the two-dimensional resistivity survey data and shallow resistivity measurement data were fused to obtain standardized resistivity data;

[0008] The standardized resistivity data is divided into geologically adapted discrete state intervals to determine multiple resistivity state intervals; the transition probability between multiple resistivity state intervals is calculated based on spatial neighborhood statistics to obtain the transition probability matrix;

[0009] The underground space surrounding the substation is divided into three-dimensional grids to obtain blank grid cells. Standardized resistivity data is interpolated and filled into the blank grid cells to obtain an initial soil resistivity model. The initial soil resistivity model is spatially constrained based on the transition probability matrix to obtain a Markov-constrained soil resistivity model. The Markov-constrained resistivity model is iteratively inverted to establish a spatial probability model of three-dimensional soil resistivity.

[0010] Optionally, the three-dimensional soil resistivity reconstruction method provided by the present invention further includes:

[0011] The two-dimensional resistivity measurement data and shallow resistivity measurement data are detected based on the Z-score and interquartile range method to obtain outlier detection results. Based on the outlier detection results, the normally distributed data with Z-scores greater than the preset value and the non-normal data with lower or upper bounds exceeding the preset range are filtered out to obtain the verified two-dimensional resistivity measurement data and shallow resistivity measurement data.

[0012] The calibrated two-dimensional resistivity measurement data is transformed into the same coordinate system as the shallow resistivity measurement data to obtain coordinate-transformed two-dimensional resistivity measurement data. The shallow resistivity measurement data is then corrected for depth deviation based on the terrain elevation data of the area surrounding the substation to obtain depth-deviation-corrected shallow resistivity measurement data. The coordinate-transformed two-dimensional resistivity measurement data and the depth-deviation-corrected shallow resistivity measurement data are then fused to obtain standardized resistivity data.

[0013] Optionally, the three-dimensional soil resistivity reconstruction method provided by the present invention further includes:

[0014] Based on the resistivity distribution characteristics of the standardized resistivity data and the geological conditions of the area surrounding the substation, the boundary of the state interval of the standardized resistivity data is determined; multiple resistivity state intervals are then determined based on the boundary of the state interval.

[0015] Optionally, the three-dimensional soil resistivity reconstruction method provided by the present invention further includes:

[0016] Based on depth, underground space is divided into shallow and deep layers, with the deep layer being deeper than the shallow layer.

[0017] The shallow layer is subdivided according to the shallow depth step size and the shallow horizontal step size to obtain shallow mesh cells; the deep layer is subdivided according to the deep depth step size and the deep horizontal step size to obtain deep mesh cells, wherein the deep depth step size is greater than the shallow depth step size, and the deep horizontal step size is greater than the shallow horizontal step size.

[0018] Based on three-dimensional spatial coordinates, shallow and deep mesh cells are assigned unique numbers to obtain blank mesh cells with spatial location allocation.

[0019] Optionally, the three-dimensional soil resistivity reconstruction method provided by the present invention further includes:

[0020] Standardized resistivity data with a density greater than the preset value are filled into the corresponding blank grid cells using ordinary kriging interpolation to obtain the initial model of soil resistivity in the dense area.

[0021] Standardized resistivity data with a density less than a preset value are filled into the corresponding blank grid cells through co-kriging interpolation, and the values ​​in the blank grid cells are corrected based on the prior resistivity of the area surrounding the substation, thus obtaining the initial model of soil resistivity in the sparse area.

[0022] Optionally, the three-dimensional soil resistivity reconstruction method provided by the present invention further includes:

[0023] An objective function is constructed with the goal of minimizing the residuals of the model response data of standardized resistivity data and Markov-constrained resistivity model.

[0024] Based on the improved conjugate gradient iterative inversion objective function, the parameters in the Markov constrained resistivity model are optimized according to the residual of the objective function. When the optimized Markov constrained resistivity model satisfies the preset convergence condition, the spatial probability model of soil three-dimensional resistivity is determined according to the optimized Markov constrained resistivity model.

[0025] Optionally, the three-dimensional soil resistivity reconstruction method provided by the present invention further includes:

[0026] Volume rendering is performed based on the resistivity of the spatial probability model of soil three-dimensional resistivity to obtain a volume rendering image.

[0027] Surface rendering is performed based on the isosurface corresponding to the pre-defined resistivity in the spatial probability model of soil three-dimensional resistivity to obtain a surface rendering image;

[0028] Based on the resistivity corresponding to the preset depth and preset horizontal plane in the spatial probability model of soil three-dimensional resistivity, slice rendering is performed to obtain vertical rendering images and horizontal rendering images.

[0029] The present invention also provides a three-dimensional soil resistivity reconstruction system, comprising:

[0030] The data preprocessing module is used to acquire two-dimensional resistivity survey line data and shallow resistivity measurement data of the soil in the area surrounding the substation; and to fuse the two-dimensional resistivity survey line data and shallow resistivity measurement data to obtain standardized resistivity data.

[0031] The parameter configuration module is used to divide the standardized resistivity data into geologically adapted discrete state intervals and determine multiple resistivity state intervals; based on spatial neighborhood statistics, the transition probability between multiple resistivity state intervals is calculated to obtain the transition probability matrix;

[0032] The 3D inversion module is used to perform 3D meshing of the underground space around the substation to obtain blank mesh cells; standardized resistivity data is interpolated and filled into the blank mesh cells to obtain an initial soil resistivity model; spatial constraints are applied to the initial soil resistivity model based on the transition probability matrix to obtain a Markov-constrained soil resistivity model; iterative inversion calculations are performed on the Markov-constrained resistivity model to establish a spatial probability model of the 3D soil resistivity.

[0033] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement any of the steps in a method for three-dimensional reconstruction of soil resistivity.

[0034] The present invention also provides a computer-readable storage medium storing a computer program that, when loaded by a processor, can execute any step of a method for three-dimensional reconstruction of soil resistivity.

[0035] The three-dimensional reconstruction method for soil resistivity provided by this invention has the following beneficial effects:

[0036] Because the three-dimensional soil resistivity reconstruction method provided by this invention introduces geologically adapted discrete state partitioning and a transition probability matrix based on spatial neighborhood statistics, it effectively characterizes the spatial distribution law and state transition characteristics of soil resistivity. It can explicitly introduce the spatial continuity constraint of soil resistivity in the inversion process, and load the transition probability matrix as a spatial constraint condition into the initial model and perform iterative inversion. This makes the final three-dimensional resistivity model not only fit the observed data, but also conform to the continuous spatial variation law of soil resistivity. It significantly improves the consistency between the inversion results and the actual stratigraphic distribution, overcomes the accuracy limitation problem caused by ignoring spatial continuity in the prior art, and ensures the accuracy of the soil three-dimensional resistivity model constructed by inversion. Attached Figure Description

[0037] To more clearly illustrate the embodiments and design schemes of the present invention, the accompanying drawings required for this embodiment will be briefly described below. The drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This is an example of existing soil resistivity analysis provided in the embodiments of the present invention;

[0039] Figure 2 This is a schematic diagram of a three-dimensional soil resistivity reconstruction method provided in an embodiment of the present invention;

[0040] Figure 3 This is an example of an old process for three-dimensional reconstruction of soil resistivity provided in an embodiment of the present invention;

[0041] Figure 4 This is an example of a three-dimensional soil resistivity reconstruction system provided in an embodiment of the present invention. Detailed Implementation

[0042] To enable those skilled in the art to better understand and implement the technical solutions of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and should not be construed as limiting the scope of protection of the present invention.

[0043] like Figure 1 As shown, existing substation soil resistivity analysis sets the electrode spacing of the two-dimensional survey lines to 20-40m during the data acquisition stage, with a maximum single electrode length of 120m and a collection depth limited to 0-15m. The data volume per acquisition is approximately 500-1000 sets. Furthermore, due to obstacles such as buildings and pipelines, the survey lines are prone to breaks, resulting in insufficient data integrity. Differentiated acquisition schemes are not designed for different geological zones, limiting data representativeness. In the two-dimensional inversion stage, relying solely on the survey line data without integrating prior geological information such as borehole stratigraphic data, it fails to reflect the spatial variation of soil resistivity in the horizontal direction. Moreover, the inversion algorithm does not address artifacts such as image distortion that are prone to occur in the filtered back-projection method. For data deeper than 15m, linear interpolation is used for estimation, resulting in errors exceeding 30% in some scenarios, leading to excessively low accuracy. In addition, the three-dimensional construction stage lacks quantitative spatial correlation analysis and fails to consider the differences in physical properties of soil at different depths, resulting in a 3D model that does not match the actual strata with less than 75% accuracy.

[0044] Considering that 3D Markov chains, as a probabilistic and statistical spatial modeling method, can describe the spatial distribution of variables through transition probability matrices, and have demonstrated application advantages in fields such as geological resource exploration and environmental monitoring, existing applications of 3D Markov chains in soil resistivity inversion lack data fusion mechanisms designed for off-site measurement scenarios in indoor substations, and do not consider the soil characteristics of different geological zones. Therefore, a complete technical solution suitable for substation engineering scenarios has not yet been formed, failing to address the aforementioned industry pain points. Based on this, this invention proposes an indirect measurement solution based on 3D Markov chains. Through four core innovations—multi-source data fusion, geologically adapted Markov modeling, deep hierarchical inversion, and engineering-grade output—it achieves accurate reconstruction of soil resistivity at depths of 0-60m underground.

[0045] Example 1

[0046] This application provides a method for three-dimensional reconstruction of soil resistivity, specifically as follows: Figure 2 As shown, it includes the following steps:

[0047] Step 11: Obtain two-dimensional resistivity survey data and shallow resistivity measurement data of the soil in the area surrounding the substation; fuse the two-dimensional resistivity survey data and shallow resistivity measurement data to obtain standardized resistivity data.

[0048] The fusion of two-dimensional resistivity measurement data and shallow resistivity measurement data can be achieved through the following steps:

[0049] Step 111: Detect outlier detection results by using the Z-score and interquartile range method on the two-dimensional resistivity measurement line data and shallow resistivity measurement data. Based on the outlier detection results, filter out the normally distributed data with Z-scores greater than the preset value and the non-normal data with lower or upper bounds exceeding the preset range in the two-dimensional resistivity measurement line data and shallow resistivity measurement data, to obtain the verified two-dimensional resistivity measurement line data and shallow resistivity measurement data.

[0050] Step 112: Convert the calibrated two-dimensional resistivity measurement data to the same coordinate system as the shallow resistivity measurement data to obtain coordinate-transformed two-dimensional resistivity measurement data; perform depth deviation correction on the shallow resistivity measurement data based on the terrain elevation data of the area surrounding the substation to obtain depth deviation-corrected shallow resistivity measurement data; fuse the coordinate-transformed two-dimensional resistivity measurement data and the depth deviation-corrected shallow resistivity measurement data to obtain standardized resistivity data.

[0051] Specifically, we first acquire two-dimensional resistivity measurement data and shallow resistivity measurement data, and then generate standardized data volumes through format verification, outlier filtering, and coordinate unification, providing a high-quality data foundation for subsequent modeling and inversion.

[0052] For example, the system collects 2D resistivity measurement line data in CSV format, containing information such as the measurement line number, X / Y / Z coordinates, resistivity value, and measurement time; shallow resistivity measurement data in CSV format, containing information such as the coordinate system, resistivity pixel value, and measurement range; or raw data in SEG-Y format output from a high-density resistivity transducer such as the DUK-2A. The information is determined by parsing the trace head information and data volume length. Then, the collected data in various formats is validated. The first level of validation checks the integrity of the fields, ensuring that the 2D resistivity measurement line data includes the core fields of X / Y / Z coordinates and resistivity value. The second level checks the physical rationality, limiting the resistivity value range to 0.1-10000 Ω·m, marking data outside this range as pending review. The third level checks spatial continuity, marking data segments with a resistivity gradient > 50% / m between adjacent measurement points as suspicious. When validation fails, a data error report containing the error location and reason is generated, providing a basis for data correction.

[0053] For data that passes validation, a combination of Z-score and Interquartile Range (IQR) testing method is used. The Z-score is used to process normally distributed data. The data was identified as outlier and processed by IQR for non-normally distributed data, with the lower bound set to "". The upper bound is set to "". Data exceeding the upper and lower bounds will be classified as abnormal data. The ranges or thresholds of the Z-score and IQR can be adjusted by those skilled in the art based on actual geological conditions, and this application does not impose any restrictions.

[0054] In addition, for detected outliers, this invention provides two processing modes: automatic filtering and manual confirmation. For example, in the automatic filtering mode, the system automatically removes outliers and backs them up to an independent data file, generating a backup log. In the manual confirmation mode, a person manually judges suspected outliers based on measurement records, retains reasonable data and marks them with a "manual confirmation" label, thereby ensuring the authenticity of the data.

[0055] After the data detection is completed, coordinate transformation is performed on the two-dimensional resistivity measurement data and shallow resistivity measurement data. For example, it is converted to mainstream coordinate systems such as "CGCS2000, WGS84, Beijing 54, Xi'an 80". The coordinate transformation is achieved through four parameters: "X translation, Y translation, rotation angle, and scale factor" or seven parameters: "X translation, Y translation, Z translation, X rotation, Y rotation, Z rotation, and scale factor". The transformation error is controlled within 0.01m, and a coordinate transformation report containing the transformation parameters and error values ​​is generated.

[0056] Furthermore, considering the depth deviation caused by topographic undulations in the shallow resistivity measurement data, topographic elevation data of the survey area is obtained based on a global topographic database. The depth benchmark is unified to the geoid, and the shallow resistivity measurement data is corrected. Finally, standardized resistivity data in binary .bin format is output for easy reading by subsequent units, along with a data preprocessing report including verification results, outlier handling records, and coordinate and depth correction details. In addition, a depth comparison map can be used to visually display the differences between the data before and after correction, ensuring the accuracy of the depth data.

[0057] Step 12: Divide the standardized resistivity data into geologically adapted discrete state intervals to determine multiple resistivity state intervals; calculate the transition probability between multiple resistivity state intervals based on spatial neighborhood statistics to obtain the transition probability matrix.

[0058] The resistivity state range can be divided using the following steps:

[0059] Step 121: Based on the resistivity distribution characteristics of the standardized resistivity data and the geological conditions of the area surrounding the substation, determine the state interval boundaries of the standardized resistivity data; determine multiple resistivity state intervals based on the state interval boundaries.

[0060] Specifically, after the two-dimensional resistivity survey data and shallow resistivity measurement data are preprocessed to obtain standardized resistivity data, a spatial probability model of soil resistivity can be established through discrete state division and transition probability matrix calculation, providing core parameters for three-dimensional inversion calculation.

[0061] For example, the resistivity range of substation soil can be divided into 5 discrete states: S1 corresponds to 0-20Ω·m, S2 corresponds to 20-50Ω·m, S3 corresponds to 50-100Ω·m, S4 corresponds to 100-300Ω·m, and S5 corresponds to 300-1000Ω·m. Custom settings for 2-10 states are supported to meet the technical requirements of different geological regions such as loess areas, alluvial plains, and mountainous areas.

[0062] Subsequently, based on the resistivity distribution characteristics such as the mean, standard deviation, and quartiles of the standardized resistivity data, each state interval is allocated through equal-frequency partitioning to ensure a balanced proportion of data in each state, thereby preventing any single state from exceeding 40% or falling below 5%. Furthermore, the interval boundaries can be manually fine-tuned, with the data proportion of each state displayed in real time during the fine-tuning process to assist in optimizing the partitioning scheme.

[0063] The system also includes a built-in geological zoning parameter library, containing preset zoning templates such as loess areas, alluvial plains, and mountainous areas. When the area where the substation is located is determined, the typical resistivity range of that area is automatically loaded, and the weight coefficients of the discrete states are adjusted. For example, for mountainous areas, the weight of the S4 high-resistivity state is increased, thereby improving the adaptability of the model to the actual geological conditions.

[0064] Next, valid samples are extracted from the standardized data volume, for example, using a spatially uniform sampling method of extracting one sample per 10m×10m×5m grid unit, with increased sampling density for key areas, such as increasing the sampling density in the core area of ​​the grounding grid by 2-3 times. The sample size is set to 2000-500 groups. In addition, to ensure the statistical significance of matrix calculations, when the sample is insufficient, a prompt is displayed indicating the need for supplementary data collection and suggesting the collection area.

[0065] After selecting valid samples, neighborhood windows of patterns such as "3×3×3", "2×2×2", or "4×4×4" are selected based on spatial neighborhood statistics. The frequency of each state transitioning to adjacent states in space is counted, generating a state transition frequency table. The transition probability is determined based on the ratio of the transition frequency to the total frequency, thus constructing an N×N transition probability matrix, where N is the number of discrete states, and element P(i,j) in the matrix represents the probability of transitioning from state i to state j. Finally, a discrete state partitioning table is determined, including state number, resistivity range, data proportion, and geological zoning adaptation description. A transition probability matrix with elements retained to four decimal places is generated, along with a matrix detection report containing sample information, test results, and optimization records.

[0066] Furthermore, once the transition probability matrix is ​​constructed, it can also be determined through X. 2 The chi-square test was used to verify the statistical reasonableness of the transition probability, and the test results P and X were then compared. 2 The system outputs statistics in real time, identifying anomalous transition terms when P < 0.05. It also supports manual correction of anomalous probability values, automatically recalculating the matrix normalization coefficients after correction to ensure the sum of probabilities in each row is 1, thus guaranteeing matrix validity.

[0067] Step 13: Perform three-dimensional meshing on the underground space around the substation to obtain blank mesh cells; interpolate standardized resistivity data to fill the blank mesh cells to obtain an initial soil resistivity model; apply spatial constraints to the initial soil resistivity model based on the transition probability matrix to obtain a Markov-constrained soil resistivity model; perform iterative inversion calculations on the Markov-constrained resistivity model to establish a spatial probability model of three-dimensional soil resistivity.

[0068] The blank grid cells can be determined by the following steps:

[0069] Step 131: Divide the underground space into shallow and deep layers based on depth, where the depth of the deep layer is greater than that of the shallow layer.

[0070] Step 132: Divide the shallow layer according to the shallow depth step size and the shallow horizontal step size to obtain shallow mesh cells; divide the deep layer according to the deep depth step size and the deep horizontal step size to obtain deep mesh cells, wherein the deep depth step size is greater than the shallow depth step size, and the deep horizontal step size is greater than the shallow horizontal step size.

[0071] Step 133: Assign unique numbers to shallow and deep mesh cells based on three-dimensional spatial coordinates to obtain blank mesh cells with spatial location allocation.

[0072] Specifically, such as Figure 3 As shown, the underground space is divided into multiple layers, including shallow, intermediate, and deep layers, using a depth-layered meshing strategy. Different 3D meshing techniques are employed for each layer. For example, the shallow layer (0-15m depth) corresponds to the core area of ​​the grounding grid design and requires high resolution, so a horizontal step size of 2m and a depth step size of 0.5m are used. The intermediate layer (15-30m depth) is meshed using a horizontal step size of 4m and a depth step size of 1m, and the deep layer (30-60m depth) is meshed using a horizontal step size of 8m and a depth step size of 2m. The specific number of layers and the step size for each layer can be set according to actual needs; this application does not impose any restrictions. Furthermore, the step size can be adjusted in real-time by calculating the total number of mesh cells, for example, controlling it to 50,000-100,000 to avoid computational overload. When the cell limit is exceeded, a prompt to increase the step size is displayed.

[0073] Subsequently, within the area surrounding the substation with unified coordinates, a three-dimensional grid spatial coordinate system is generated, with the X / Y axes corresponding to the horizontal direction and the Z axis corresponding to the depth direction. Each grid cell is assigned a unique number to achieve a precise correspondence between the grid cell and its spatial location. For example, the grid cell with X=0m, Y=0m, and Z=0m is assigned the number "G001-001-001".

[0074] The initial model for soil resistivity can be constructed through the following steps:

[0075] Step 134: Fill the corresponding blank grid cells with the standardized resistivity data whose data density is greater than the preset value through ordinary kriging interpolation to obtain the initial model of soil resistivity in the dense area.

[0076] Step 135: Standardized resistivity data with a data density less than the preset value are filled into the corresponding blank grid cells by co-kriging interpolation, and the values ​​in the blank grid cells are corrected based on the prior resistivity of the area surrounding the substation to obtain the initial model of soil resistivity in the sparse area.

[0077] Specifically, after obtaining blank grids, kriging interpolation can be used to spatially interpolate the standardized data volume, filling the initial resistivity values ​​of the blank grid cells. For example, ordinary kriging interpolation is used in data-dense areas such as around the survey line, while co-kriging interpolation is used in data-sparse areas such as deep, unmeasured areas, combined with geological prior information such as formation resistivity values ​​from borehole data for correction. Furthermore, interpolation errors can be displayed using an error variance plot, marking cells with errors exceeding 15% as requiring inversion optimization.

[0078] Subsequently, the transition probability matrix and geological zoning parameters were loaded into the initial model as spatial constraints. For each grid cell, based on the resistivity state of its neighboring cells, the state probability distribution of that cell was calculated using the transition probability. The median of the resistivity interval corresponding to the state with the highest probability was selected as the correction value for the initial model, ensuring that the model conforms to the spatial correlation law of soil resistivity.

[0079] The three-dimensional resistivity model of soil can be constructed through the following steps:

[0080] Step 136: Construct an objective function with the goal of minimizing the residuals of the model response data of the standardized resistivity data and the Markov constrained resistivity model.

[0081] Step 137: Based on the improved conjugate gradient iterative inversion objective function, optimize the parameters in the Markov constrained resistivity model according to the residual of the objective function. When the optimized Markov constrained resistivity model satisfies the preset convergence condition, determine the spatial probability model of soil three-dimensional resistivity according to the optimized Markov constrained resistivity model.

[0082] Specifically, the objective function is constructed with the goal of minimizing the residual between the observed data and the model response data. As shown in formula (1):

[0083] (1)

[0084] in, For observational data in a standardized data volume, For model response data, For model parameters such as resistivity of grid cells, As the initial model, A data weight matrix that is automatically assigned based on data precision. This is the model weight matrix constructed based on Markov transition probabilities. These are the regularization parameters automatically selected using the L-curve method.

[0085] Then, the improved conjugate gradient method is used to solve the objective function. The default number of iterations is 20-50, but it can be set according to actual needs. Furthermore, the residual is calculated after each iteration, as shown in formula (2):

[0086] (2)

[0087] in, The iteration automatically stops when the residual is less than 5%. Then, a convergence determination mechanism based on both residual convergence and model parameter convergence is used. If the convergence condition is not met, the regularization parameter is automatically adjusted. Alternatively, the iteration can be repeated after each iteration step to ensure stable and reliable inversion results and avoid inversion bias caused by improper parameter settings. The final soil 3D resistivity model is obtained when the residual change rate is <0.1% for three consecutive iterations and the model parameter change rate is <0.5% for three consecutive iterations. The final output is a VTK format 3D resistivity model file that can be read by professional visualization tools for subsequent analysis. It includes an iterative inversion report containing grid parameters, initial model information, iteration process data, convergence verification results, etc., and displays the residual distribution layered by depth, intuitively presenting the residual distribution map of high residual regions.

[0088] Step 14: Perform volume rendering based on the resistivity of the spatial probability model of soil three-dimensional resistivity to obtain a volume rendering image.

[0089] Step 15: Render the surface based on the isosurface corresponding to the pre-set resistivity in the spatial probability model of soil three-dimensional resistivity to obtain the surface rendering image.

[0090] Step 16: Based on the resistivity corresponding to the preset depth and preset horizontal plane in the spatial probability model of soil three-dimensional resistivity, slice and render to obtain vertical and horizontal rendered images.

[0091] Specifically, once the three-dimensional soil resistivity model is constructed, it can be presented in an intuitive form through multi-mode rendering and interactive analysis functions, helping engineers quickly identify soil resistivity distribution patterns and abnormal areas, and providing visual support for grounding design decisions.

[0092] For example, multi-mode grayscale rendering technology can provide volume rendering images, surface rendering images, and slice rendering images of a three-dimensional soil resistivity model. For instance, the volume rendering image can be based on a three-dimensional resistivity model and rendered using the "resistivity-grayscale value mapping" rule. The lower the resistivity, the lighter the grayscale, and the higher the resistivity, the darker the grayscale. The default grayscale gradient range is set to 0-1000Ω・m, corresponding to 0-255 grayscale values. It also supports setting the critical grounding design range of 50-100Ω・m to medium grayscale to highlight the distribution of this range and other custom mapping relationships. Furthermore, the volume rendering image supports spatial rotation, scaling, and translation operations and can be output as an image file for report preparation. Surface rendering images extract isosurfaces corresponding to specific resistivity levels in a 3D model, such as the 50Ω·m and 200Ω·m surfaces, and display them using a "wireframe + dot fill" method. The wireframe is a solid black line, and the density of the dot fill is positively correlated with how close the resistivity is to the target value; for example, a higher density indicates that the resistivity is closer to the target value. Surface rendering images support the simultaneous extraction and display of multiple isosurfaces, distinguished by different dot densities, clearly presenting the spatial distribution boundaries of different resistivity ranges. Slice rendering images offer two modes: "horizontal slicing" and "vertical slicing." Horizontal slicing allows selection of any depth, such as Z=5m or Z=10m, to display the resistivity distribution on a horizontal plane at that depth. Vertical slicing allows selection of any horizontal position, such as X=50m or Y=80m, to display the resistivity distribution within that vertical plane. The slice results are labeled with resistivity values, for example, one value every five grid cells, retaining one decimal place. Continuous slice animations are supported, visually demonstrating the variation of resistivity with depth or horizontal position.

[0093] Furthermore, based on preset anomaly criteria such as "high resistivity anomaly: resistivity > 300 Ω·m and continuous area > 100 m²" and "low resistivity anomaly: resistivity < 20 Ω·m and continuous volume > 50 m³", it can automatically scan the 3D model and mark anomaly areas, which are enclosed by black dashed boxes. An anomaly list is generated, recording the anomaly type, location, range, and resistivity statistics such as mean, maximum, and minimum values. It also allows engineers to manually draw profile lines in 3D space by determining the start and end coordinates, automatically extracting the corresponding vertical profile and annotating the depth scale, horizontal distance scale, and resistivity grayscale legend. It also supports extracting multiple profile lines simultaneously and displaying them in separate windows for comparison, facilitating the analysis of resistivity distribution differences at different locations. Moreover, through spatial positioning, it can obtain the core information of the unit in real time, including the unit number, spatial coordinates, resistivity value, discrete state, and inversion residual. It supports selecting any area for data statistics, automatically calculates the total number of grid cells, mean resistivity, standard deviation, maximum and minimum values ​​of the area, and generates an area statistical report to provide data support for the selection of grounding electrode placement areas.

[0094] In summary, it can output PNG / JPG / GIF format multi-mode rendered images with legends and scale annotations, anomaly area reports including anomaly location maps, statistical information and engineering suggestions, SVG format cross-sectional drawings with editable vector graphics, and Excel format regional statistical reports for further data analysis.

[0095] In addition, in order to convert the 3D inversion results into a format that meets the engineering design requirements, the following design can be used to achieve collaborative adaptation with grounding design tools and GIS software. Specifically, for grounding design software compatible formats, a CDEGS-specific ".res" format file is generated, containing grid cell coordinates and resistivity values. This file can be directly imported into the CDEGS "Soil Model" module without manual data format conversion, reducing data processing workload in the design process. An AutoCAD-compatible ".dwg" format file is also generated, containing a 3D resistivity model wireframe and anomaly area markers. This file can be overlaid with grounding grid design drawings to assist in determining the location of grounding electrodes. For GIS software compatible formats, a deeply layered GeoTIFF format image file is generated, containing spatial coordinate information. This file can be directly imported into GIS software such as ArcGIS and QGIS for overlay analysis with topographic and geological maps of the survey area, integrating multi-source geographic information to assist in substation site selection and grounding scheme optimization. For general data formats, a VTK format file is generated, preserving complete 3D model information and supporting further analysis with professional visualization software. This file includes cell numbers, coordinates, resistivity values, and anomaly markers. A CSV format grid cell information table is easily processed using tools such as Excel and Python. A JSON format structured data file contains full project process information and supports data interaction with other engineering software.

[0096] In summary, the soil resistivity 3D reconstruction method provided by this invention integrates 2D resistivity survey data and shallow resistivity measurement data through an indirect measurement mode, enabling the acquisition of large-scale soil resistivity data across a depth of 0-60m and a range of 400m×400m, breaking through the limitations of existing measurement ranges. By establishing a multi-source data verification and coordinate unification mechanism, it eliminates data format differences and coordinate deviations, achieving accurate fusion of multi-source data, fully leveraging the advantages of different data sources, and improving the quality of the inversion base data. By constructing a transition probability matrix based on Markov constraint models such as 3D Markov chain models, it quantifies the spatial correlation of soil resistivity, accurately describes the transfer law of resistivity between adjacent grid units, and improves the rationality of the inversion results. Through a differentiated grid partitioning scheme of shallow subdivision and deep coarsening, it reduces the accuracy of deep grids while maintaining high shallow resolution, balancing computational efficiency and accuracy. Furthermore, it develops multi-mode 3D visualization functions such as rendering and slice rendering, supporting format compatibility with GIS software and grounding grid design software, meeting the needs of engineering decision-making and design.

[0097] Example 2

[0098] This application also provides a three-dimensional soil resistivity reconstruction system, comprising:

[0099] The data preprocessing module is used to acquire two-dimensional resistivity survey line data and shallow resistivity measurement data of the soil in the area surrounding the substation; and to fuse the two-dimensional resistivity survey line data and shallow resistivity measurement data to obtain standardized resistivity data.

[0100] The parameter configuration module is used to divide the standardized resistivity data into geologically adapted discrete state intervals and determine multiple resistivity state intervals; based on spatial neighborhood statistics, the transition probability between multiple resistivity state intervals is calculated to obtain the transition probability matrix;

[0101] The 3D inversion module is used to perform 3D meshing of the underground space around the substation to obtain blank mesh cells; standardized resistivity data is interpolated and filled into the blank mesh cells to obtain an initial soil resistivity model; spatial constraints are applied to the initial soil resistivity model based on the transition probability matrix to obtain a Markov-constrained soil resistivity model; iterative inversion calculations are performed on the Markov-constrained resistivity model to establish a spatial probability model of the 3D soil resistivity.

[0102] Specifically, such as Figure 4 As shown, the soil resistivity 3D reconstruction system provided by this invention revolves around the entire process of "data input - preprocessing - modeling and inversion - visualization - result output," constructing five core technical units: "multi-source data fusion unit - Markov parameter configuration unit - 3D inversion calculation unit - 3D visualization unit - engineering adaptation output unit," forming a closed-loop technical system for 3D inversion and reconstruction of soil resistivity at substation sites. Each technical unit works collaboratively, inputting 2D resistivity survey line data and shallow resistivity measurement data in CSV format into the multi-source data fusion unit. After fusion, the Markov parameter configuration unit processes and outputs a discrete state table and transition probability matrix. Subsequently, the 3D inversion calculation unit constructs a 3D resistivity model, and the 3D visualization unit outputs volume rendering maps, slice maps, and other rendering results. Finally, the engineering adaptation output unit outputs GeoTIFF format data, VTK format data, and an engineering report. This achieves the transformation from indirect measurement data to engineering-usable results, thus meeting the technical requirements of 3D soil resistivity reconstruction in substation engineering scenarios.

[0103] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory. The processor executes the computer program to implement the steps in an embodiment of a method for three-dimensional reconstruction of soil resistivity. Specific implementation methods can be found in the method embodiments, and will not be repeated here.

[0104] Furthermore, the present invention also provides a non-transitory computer-readable storage medium containing instructions, on which a computer program is stored. For example, a memory containing instructions that can be executed by a processor of a computer device to perform the above-described method. For example, the non-transitory computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc. When the computer program is executed by the processor, it can implement the steps in an embodiment of a three-dimensional soil resistivity reconstruction method. Specific implementation methods can be found in the method embodiments, which will not be repeated here.

[0105] Those skilled in the art will understand that embodiments of the present invention can provide methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0106] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0107] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0108] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0109] It should be noted that the specific embodiments described above enable those skilled in the art to more fully understand the present invention, but do not limit the present invention in any way. Therefore, although the present invention has been described in detail in this specification and embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the present invention; and all technical solutions and improvements that do not depart from the spirit and scope of the present invention are covered within the protection scope of the present invention patent. No reference numerals in the claims should be construed as limiting the scope of the claims. Any simple variations or equivalent substitutions of technical solutions that can be readily obtained by those skilled in the art within the scope of the technology disclosed in the present invention are within the protection scope of the present invention.

Claims

1. A method for three-dimensional reconstruction of soil resistivity, characterized in that, include: Two-dimensional resistivity survey data and shallow resistivity measurement data of the soil in the area surrounding the substation are obtained; the two-dimensional resistivity survey data and shallow resistivity measurement data are fused to obtain standardized resistivity data; The standardized resistivity data is divided into geologically adapted discrete state intervals to determine multiple resistivity state intervals. Based on spatial neighborhood statistical calculations, the transition probability between the multiple resistivity state intervals is obtained, resulting in a transition probability matrix. The underground space surrounding the substation is divided into three-dimensional meshes to obtain blank mesh units. Specifically, this involves dividing the underground space around the substation into shallow and deep layers based on depth, wherein the depth of the deep layer is greater than that of the shallow layer; subdividing the shallow layer according to the shallow layer depth step size and the shallow layer horizontal step size to obtain shallow mesh units; subdividing the deep layer according to the deep layer depth step size and the deep layer horizontal step size to obtain deep mesh units, wherein the deep layer depth step size is greater than the shallow layer depth step size, and the deep layer horizontal step size is greater than the shallow layer horizontal step size; and assigning unique numbers to the shallow and deep mesh units based on three-dimensional spatial coordinates to obtain blank mesh units with spatially assigned locations. The standardized resistivity data is interpolated and filled into the blank grid cells to obtain an initial soil resistivity model. Specifically, this includes filling the standardized resistivity data with a data density greater than a preset value into the corresponding blank grid cells using ordinary kriging interpolation to obtain an initial soil resistivity model for dense areas; and filling the standardized resistivity data with a data density less than a preset value into the corresponding blank grid cells using co-kriging interpolation, and correcting the values ​​in the blank grid cells based on the prior resistivity of the area surrounding the substation to obtain an initial soil resistivity model for sparse areas. Based on the transition probability matrix, the initial soil resistivity model is spatially constrained to obtain a Markov-constrained resistivity model. The Markov-constrained resistivity model is then iteratively inverted to establish a spatial probability model of the three-dimensional soil resistivity. This includes constructing an objective function with the goal of minimizing the residual between the standardized resistivity data and the model response data of the Markov-constrained resistivity model; iteratively inverting the objective function based on the improved conjugate gradient; optimizing the parameters in the Markov-constrained resistivity model based on the residual of the objective function; and determining the spatial probability model of the three-dimensional soil resistivity based on the optimized Markov-constrained resistivity model when the optimized Markov-constrained resistivity model satisfies the preset convergence condition.

2. The method for three-dimensional reconstruction of soil resistivity according to claim 1, characterized in that, The standardized resistivity data obtained by fusing the two-dimensional resistivity measurement data and the shallow resistivity measurement data includes: The two-dimensional resistivity measurement data and shallow resistivity measurement data are detected based on Z-score and interquartile range method to obtain outlier detection results; based on the outlier detection results, normal distribution data with Z-score greater than preset value and non-normal data with lower or upper bounds exceeding preset range are filtered out to obtain verified two-dimensional resistivity measurement data and shallow resistivity measurement data. The verified two-dimensional resistivity measurement data is converted to the same coordinate system as the shallow resistivity measurement data to obtain coordinate-transformed two-dimensional resistivity measurement data; the shallow resistivity measurement data is then corrected for depth deviation based on the terrain elevation data of the area surrounding the substation to obtain depth-deviation-corrected shallow resistivity measurement data; the coordinate-transformed two-dimensional resistivity measurement data and the depth-deviation-corrected shallow resistivity measurement data are then fused to obtain the standardized resistivity data.

3. The method for three-dimensional reconstruction of soil resistivity according to claim 1, characterized in that, The standardized resistivity data is divided into geologically adapted discrete state intervals to determine multiple resistivity state intervals, including: Based on the resistivity distribution characteristics of the standardized resistivity data and the geological conditions of the area surrounding the substation, the boundary of the state interval of the standardized resistivity data is determined; and the plurality of resistivity state intervals are determined based on the boundary of the state interval.

4. The method for three-dimensional reconstruction of soil resistivity according to claim 1, characterized in that, After determining the spatial probability model of the soil's three-dimensional resistivity, the method further includes: Volume rendering is performed based on the resistivity of the spatial probability model of the three-dimensional resistivity of the soil to obtain a volume rendering image; Surface rendering is performed based on the isosurface corresponding to the pre-set resistivity in the spatial probability model of the three-dimensional resistivity of the soil to obtain a surface rendering image; Based on the resistivity corresponding to the preset depth and preset horizontal plane in the spatial probability model of soil three-dimensional resistivity, slice rendering is performed to obtain vertical rendering images and horizontal rendering images.

5. A three-dimensional soil resistivity reconstruction system, characterized in that, include: The data preprocessing module is used to acquire two-dimensional resistivity survey line data and shallow resistivity measurement data of the soil in the area surrounding the substation; and to fuse the two-dimensional resistivity survey line data and shallow resistivity measurement data to obtain standardized resistivity data. The parameter configuration module is used to divide the standardized resistivity data into geologically adapted discrete state intervals and determine multiple resistivity state intervals. Based on spatial neighborhood statistical calculations, the transition probability between the multiple resistivity state intervals is obtained, resulting in a transition probability matrix. A 3D inversion module is used to perform 3D meshing of the underground space surrounding the substation to obtain blank mesh cells. Specifically, this includes dividing the underground space around the substation into shallow and deep layers based on depth, where the deep layer is deeper than the shallow layer; subdividing the shallow layer according to a shallow depth step size and a shallow horizontal step size to obtain shallow mesh cells; subdividing the deep layer according to a deep depth step size and a deep horizontal step size to obtain deep mesh cells, where the deep depth step size and the deep horizontal step size are greater than the shallow depth step size and the deep horizontal step size are greater than the shallow horizontal step size; assigning unique numbers to the shallow and deep mesh cells based on 3D spatial coordinates to obtain spatially assigned blank mesh cells; and interpolating and filling the blank mesh cells with standardized resistivity data to obtain an initial soil resistivity model. Specifically, this includes filling the corresponding blank mesh cells with standardized resistivity data of higher density than a preset value using ordinary kriging interpolation to obtain an initial soil resistivity model for dense areas; and filling the blank mesh cells with data of lower density... The standardized resistivity data at preset values ​​are filled into the corresponding blank grid cells through co-kriging interpolation, and the values ​​in the blank grid cells are corrected based on the prior resistivity of the area surrounding the substation to obtain an initial soil resistivity model for the sparse region. The initial soil resistivity model is then spatially constrained based on the transition probability matrix to obtain a Markov-constrained resistivity model. The Markov-constrained resistivity model is iteratively inverted to establish a spatial probability model for the three-dimensional soil resistivity. Specifically, this includes constructing an objective function with the goal of minimizing the residual between the standardized resistivity data and the model response data of the Markov-constrained resistivity model; iteratively inverting the objective function based on the improved conjugate gradient; optimizing the parameters in the Markov-constrained resistivity model based on the residual of the objective function; and determining the spatial probability model for the three-dimensional soil resistivity based on the optimized Markov-constrained resistivity model when the optimized model satisfies the preset convergence condition.

6. A computer device, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the three-dimensional soil resistivity reconstruction method according to any one of claims 1 to 4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is loaded by the processor, it is able to execute the steps of the three-dimensional soil resistivity reconstruction method according to any one of claims 1 to 4.

Images