Underground cavern three-dimensional modeling method and system based on dynamic parametric node regulation and control

Abstract

The invention provides an underground cavern three-dimensional modeling method and system based on dynamic parameterization node regulation, and the method comprises the steps: firstly obtaining a multi-source sensing data set which covers section contour point cloud, geological environment and real-time monitoring data, generating an initial section contour curve based on the section contour point cloud data, obtaining a dynamically adjusted curve and node regulation and control parameters through dynamic parameterization node regulation and control, and generating a stretching path reference line according to the cavern axis coordinate parameters; the adjusted section profile curve is axially stretched to generate a three-dimensional cavern curved surface containing a local three-dimensional curved surface set, finally, section stitching is conducted on the local three-dimensional curved surface set, a connection area topological structure is generated according to joint direction parameters and optimized, and therefore an optimized three-dimensional cavern model is generated. Therefore, more accurate and flexible underground cavern three-dimensional modeling is realized.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular, to a three-dimensional modeling method and system for underground caverns based on dynamic parameterized node regulation. Background Art

[0002] In the field of geotechnical engineering, the three-dimensional modeling technology of underground caverns is of crucial significance for aspects such as engineering design, construction planning, and safety assessment. By constructing an accurate three-dimensional model, engineers can more intuitively understand the geological structure, spatial form of underground caverns, and their interaction with the surrounding environment, thus providing strong support for the smooth progress of the project. However, many problems that need to be solved urgently have emerged in the actual application of existing three-dimensional modeling technologies for underground caverns.

[0003] During the model construction process, most of the existing methods adopt static parameterized modeling methods. Although they can achieve the rapid generation of models to a certain extent, they lack the ability of dynamic adjustment. Once the geological conditions change or new situations occur during the construction process, the model cannot make corresponding adjustments in a timely manner, resulting in the disconnection between the model and the actual situation and reducing the practicability and reliability of the model.

[0004] At the same time, the existing three-dimensional modeling technologies for underground caverns also need to improve in terms of computational efficiency and automation. The modeling process often requires a large amount of manual intervention and cumbersome calculation steps, which not only increases the cost and time of modeling but also easily introduces human errors and affects the quality of modeling. Summary of the Invention

[0005] In view of the problems mentioned above, in combination with the first aspect of the present invention, embodiments of the present invention provide a three-dimensional modeling method for underground caverns based on dynamic parameterized node regulation, and the method includes: Obtain a multi-source perception data set of the underground cavern, where the multi-source perception data set includes a cross-section contour point cloud data set collected by a laser scanning device, a geological environment data set collected by a geological exploration device, and a real-time monitoring data set collected by a monitoring sensor; wherein, the geological environment data set includes surrounding rock grade parameters, surrounding rock stress distribution parameters, and cavern axis coordinate parameters; Generate an initial cross-section contour curve based on the cross-section contour point cloud data set, and regulate the initial cross-section contour curve through dynamic parameterized nodes to generate a dynamically adjusted cross-section contour curve and a node regulation parameter set; Generate a stretching path reference line based on the cavern axis coordinate parameters, and perform an axial stretching operation on the dynamically adjusted cross-section contour curve along the stretching path reference line to generate a three-dimensional cavern surface including a set of local three-dimensional surfaces; Perform a cross-section stitching operation on the set of local three-dimensional surfaces, generate a topological structure of the connection region based on the joint direction parameters in the set of geological environment data, and perform topological optimization processing on the connection regions of adjacent surfaces to generate an optimized three-dimensional chamber model.

[0006] In a possible implementation manner of the first aspect, generating an initial cross-section profile curve based on the set of cross-section profile point cloud data and regulating the initial cross-section profile curve through dynamically parameterized nodes to generate a dynamically adjusted cross-section profile curve and a set of node regulation parameters includes: Perform noise reduction processing on the set of cross-section profile point cloud data to obtain a noise-reduced set of cross-section profile point cloud data; wherein, the noise reduction processing includes operations of removing abnormal points and filling missing points by unifying the coordinate system based on the coordinate system conversion parameters in the set of geological environment data; Generate an initial cross-section profile curve based on the noise-reduced set of cross-section profile point cloud data, and define a set of key control nodes including node regulation parameters on the initial cross-section profile curve, wherein the node regulation parameters include coordinate parameters, curvature parameters, and tangent direction parameters; Apply constraint conditions to the curvature parameters of the set of key control nodes according to the surrounding rock grade parameters in the set of geological environment data to generate a constrained initial cross-section profile curve; wherein, the curvature parameters in the soft rock area are restricted to be lower than a first threshold, and the curvature parameters in the hard rock area are restricted to be higher than a second threshold; Update the coordinate parameters of the set of key control nodes according to the surrounding rock deformation parameters in the set of real-time monitoring data to generate a dynamically adjusted cross-section profile curve and a set of node regulation parameters including the updated parameters.

[0007] In a possible implementation manner of the first aspect, the performing noise reduction processing on the set of cross-section profile point cloud data to obtain a noise-reduced set of cross-section profile point cloud data includes: Perform a coordinate transformation alignment operation on the coordinate system of the set of cross-section profile point cloud data and the coordinate system of the set of geological environment data; Identify abnormal points in the set of cross-section profile point cloud data and remove the abnormal points, and fill the missing area after removing the abnormal points through a neighboring point interpolation algorithm that adjusts the interpolation weight based on the surrounding rock grade parameters to obtain a noise-reduced set of cross-section profile point cloud data.

[0008] In a possible implementation manner of the first aspect, the generating an initial cross-section profile curve based on the noise-reduced set of cross-section profile point cloud data and defining a set of key control nodes including node regulation parameters on the initial cross-section profile curve includes: Interpolate and fit the denoised cross-section contour point cloud data set using a B-spline curve to generate an initial cross-section contour curve; Calculate the local curvature parameter of each data point on the initial cross-section contour curve, and identify geometric feature points according to the mutation amplitude of the curvature parameter; Screen the geometric feature points based on the surrounding rock grade parameter in the geological environment data set, and retain the feature points with curvature parameters lower than the first threshold in the soft rock area and the feature points with curvature parameters higher than the second threshold in the hard rock area; Calculate the tangent direction parameter of the initial cross-section contour curve at the retained geometric feature points, and bind the coordinate parameter, corresponding curvature parameter and tangent direction parameter of the geometric feature points into a key control node set.

[0009] In a possible implementation manner of the first aspect, the method for updating the coordinate parameters of the key control node set according to the surrounding rock deformation parameter in the real-time monitoring data set to generate a dynamically adjusted cross-section contour curve and a node control parameter set including updated parameters includes: Convert the surrounding rock deformation parameter into a coordinate offset of the key control node set; Perform filtering processing on the coordinate offset to generate a smoothed node coordinate update amount; Adjust the coordinate parameters of the key control node set according to the smoothed node coordinate update amount, and refit the initial cross-section contour curve based on the curvature parameter and tangent direction parameter of the key control node set to generate a dynamically adjusted cross-section contour curve; Store the updated coordinate parameter, curvature parameter and tangent direction parameter as the node control parameter set.

[0010] In a possible implementation manner of the first aspect, the method for generating a stretching path reference line based on the axis coordinate parameter of the cavern, and performing an axial stretching operation on the dynamically adjusted cross-section contour curve along the stretching path reference line to generate a three-dimensional cavern surface including a local three-dimensional surface set includes: Generate an initial stretching path reference line based on the axis coordinate parameter of the cavern in the geological environment data set; Divide the initial stretching path reference line into a homogeneous rock mass stretching section and a heterogeneous rock mass stretching section according to the surrounding rock stress distribution parameter, and set an initial stretching step parameter for each stretching section; Dynamically reduce the initial stretching step parameter in the heterogeneous rock mass stretching section, and perform curvature continuity correction on the stretching path reference line in the arc area according to the curvature parameter in the node control parameter set to generate a corrected stretching path; Perform a segmented stretching operation on the dynamically adjusted cross-sectional profile curve along the corrected stretching path to generate a set of local three-dimensional surfaces including a homogeneous rock mass stretching segment surface and a heterogeneous rock mass stretching segment surface.

[0011] In a possible implementation manner of the first aspect, the dynamically reducing the initial stretching step size parameter within the heterogeneous rock mass stretching segment includes: Generate a stress distribution map based on the surrounding rock stress distribution parameters, and identify the stress concentration area and the fault zone area; Mark the stress concentration area and the fault zone area as the heterogeneous rock mass stretching segment, and reduce the initial stretching step size parameter to a preset ratio dynamically adjusted based on the curvature parameter in the node control parameter set.

[0012] In a possible implementation manner of the first aspect, the performing a cross-section stitching operation on the set of local three-dimensional surfaces, generating a connection area topological structure based on the joint direction parameter in the geological environment data set, and performing topological optimization processing on the connection area of adjacent surfaces to generate an optimized three-dimensional cavern model includes: Perform boundary matching processing on the surfaces of adjacent stretching segments in the set of local three-dimensional surfaces to generate a set of local three-dimensional surfaces after boundary matching; wherein, the boundary matching processing eliminates the curvature mutation in the splicing area through a smoothing algorithm; Based on the cross-cavern position parameter and the joint direction parameter in the geological environment data set, generate a connection area topological structure forming a preset angle with the joint direction in the cross-cavern area; wherein, the connection area topological structure includes a main extension direction parameter and an auxiliary connection line parameter; Perform mesh division processing on the set of local three-dimensional surfaces after boundary matching according to the surrounding rock grade parameter to generate a three-dimensional cavern model including a dense mesh area; wherein, the density parameter of the dense mesh area is dynamically adjusted based on the curvature parameter in the node control parameter set; Convert the three-dimensional cavern model into construction adaptation format data including support structure parameters, and dynamically correct the coordinate parameters of the three-dimensional cavern model according to the real-time displacement parameters collected by the construction monitoring equipment to generate an optimized three-dimensional cavern model.

[0013] In a possible implementation manner of the first aspect, the generating a connection area topological structure forming a preset angle with the joint direction in the cross-cavern area based on the cross-cavern position parameter and the joint direction parameter in the geological environment data set includes: Extract the boundary coordinates of the cross area according to the cross-cavern position parameter to generate the cross area range; Determine the main extension direction of the cross - region range based on the joint direction parameter, and generate an auxiliary extension direction that forms a fixed angular difference with the main extension direction according to the preset included angle. Adjust the length parameter of the auxiliary connection line based on the tangent direction parameter in the node regulation parameter set, and construct a connection - region topological structure including the Voronoi diagram structure.

[0014] On the other hand, an embodiment of the present invention further provides a three - dimensional modeling system for underground caverns based on dynamic parametric node regulation, including a processor and a machine - readable storage medium. The machine - readable storage medium is connected to the processor. The machine - readable storage medium is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the machine - readable storage medium to implement the above - mentioned method.

[0015] Based on the above aspects, through the deep fusion and dynamic regulation mechanism of multi - source perception data, the embodiment of the present invention has achieved a significant improvement in the accuracy, adaptability and intelligent level of three - dimensional modeling of underground caverns. Specifically, by integrating the cross - section contour point cloud data obtained by laser scanning, the geological environment parameters provided by geological exploration, and real - time monitoring data, the introduction of dynamic parametric nodes enables the generation of cross - section contour curves not to be limited to static templates, but to be adaptively adjusted according to key parameters such as surrounding rock grades and stress distributions in the geological environment data set. This not only improves the accuracy of curve fitting, but also endows the model with the ability to dynamically respond to complex geological conditions, effectively solving the problem that it is difficult to balance geometric accuracy and geological adaptability in traditional modeling methods. Further, based on the stretching path reference line generated from the axis coordinate parameters of the cavern, axial stretching is performed in combination with the dynamically adjusted cross - section contour curve to form a three - dimensional cavern surface containing rich local three - dimensional surface features, which not only retains the fine details of the geometric shape of the cavern, but also realizes the global consistency of the model through parametric regulation. Particularly importantly, in the cross - section stitching stage, the joint direction parameter in the geological environment data is used to generate a connection - region topological structure, and topological optimization processing is performed on the adjacent surface connection region, which not only ensures the continuity of the physical connection of the model, but also optimizes the model structure from the perspective of geomechanics, making the finally generated three - dimensional cavern model highly unified in geometric shape, geological adaptability and mechanical rationality. Thus, the accuracy, efficiency and intelligent level of three - dimensional modeling of underground caverns have been significantly improved. Description of the Drawings

[0016] Figure 1 It is a schematic execution flow diagram of the three - dimensional modeling method for underground caverns based on dynamic parametric node regulation provided by the embodiment of the present invention.

[0017] Figure 2It is a schematic diagram of exemplary hardware and software components of the three-dimensional modeling system for underground caverns based on dynamic parameterized node regulation provided by an embodiment of the present invention. Detailed implementation manners

[0018] The present invention will be specifically described below with reference to the accompanying drawings of the specification. Figure 1 It is a schematic flowchart of a three-dimensional modeling method for underground caverns based on dynamic parameterized node regulation provided by an embodiment of the present invention. The three-dimensional modeling method for underground caverns based on dynamic parameterized node regulation will be introduced in detail below.

[0019] Step S110: Obtain a multi-source perception data set of the underground cavern. The multi-source perception data set includes a cross-section contour point cloud data set collected by a laser scanning device, a geological environment data set collected by a geological exploration device, and a real-time monitoring data set collected by a monitoring sensor. Among them, the geological environment data set includes surrounding rock grade parameters, surrounding rock stress distribution parameters, and cavern axis coordinate parameters.

[0020] In this step, to comprehensively understand the actual situation of the underground cavern, it is necessary to obtain a multi-source perception data set. Specifically, a high-precision laser scanning device can be used to densely scan each cross-section of the underground cavern to obtain a cross-section contour point cloud data set. For example, the laser scanning device scans the cavern cross-section every 0.1 meter, and each scan can obtain point cloud data of thousands of points. This point cloud data accurately records the contour information of the cavern cross-section, forming a set containing a large amount of three-dimensional coordinate data.

[0021] In addition, through geological exploration devices such as drilling devices and geophysical exploration devices, the geological environment around the underground cavern is deeply explored to collect a geological environment data set. For the surrounding rock grade parameters, rock samples can be collected and the rock mechanics test methods in related technologies can be used to analyze indicators such as the strength and integrity of the rock, so as to determine the surrounding rock grade. For example, according to relevant standards, the surrounding rock is divided into grades I-V. Grade I surrounding rock has the best stability, and grade V surrounding rock has the worst stability. The surrounding rock stress distribution parameters are measured by arranging stress sensors at different positions around the cavern. These stress sensors can monitor the stress changes of the rock mass in real time, and finally form a parameter set reflecting the stress state of the rock mass around the cavern. The cavern axis coordinate parameters are accurately determined by high-precision measuring devices such as total stations and GPS locators to obtain a set containing multiple coordinate points.

[0022] At the same time, various monitoring sensors can be used to collect real-time monitoring data sets of the cavern. For example, a displacement sensor is installed on the surface of the cavern surrounding rock to record the displacement change of the surrounding rock every hour, so as to obtain the deformation parameters of the surrounding rock; a seepage pressure sensor is installed inside the cavern to monitor the pressure change of groundwater in real time and obtain the seepage pressure parameters.

[0023] Step S120: generating an initial cross-sectional contour curve based on the cross-sectional contour point cloud data set, and regulating the initial cross-sectional contour curve through a dynamic parameterized node to generate a dynamically adjusted cross-sectional contour curve and a node regulation parameter set.

[0024] In this embodiment, after the cross-section contour point cloud data set is acquired, it is necessary to generate an initial cross-section contour curve based on the cross-section contour point cloud data set and perform dynamic regulation.

[0025] Step S121: performing noise reduction processing on the cross-section contour point cloud data set to obtain a noise-reduced cross-section contour point cloud data set; wherein the noise reduction processing includes the operations of eliminating abnormal points and completing missing points in a unified coordinate system based on the coordinate system conversion parameters of the geological environment data set.

[0026] Step S1211: performing coordinate conversion and alignment operations on the coordinate system of the cross-section contour point cloud data set and the coordinate system of the geological environment data set.

[0027] In the underground cavern data collection scenario, laser scanning equipment and geological exploration equipment use different coordinate systems due to their different working principles and application purposes. When collecting cross-section contour point cloud data, the laser scanning device constructs a coordinate system based on the position and orientation of its own equipment. For example, the equipment is placed at a certain position in the cavern, with the point where the equipment is located as the origin, the horizontal direction as the X-axis, the vertical upward direction as the Y-axis, and the direction perpendicular to the XY plane as the Z-axis. When collecting geological environment data, geological exploration equipment usually uses a geographic coordinate system or a coordinate system uniformly specified by the project, such as a geodetic coordinate system, in order to match geographic information and the overall project plan.

[0028] To align the two coordinate systems, the coordinate system transformation parameters in the geological environment data set, namely the rotation matrix and translation vector, are required. The rotation matrix is ​​used to describe the rotation relationship between the coordinate systems, which can rotate the points in the cross-section contour point cloud data set from the original coordinate system to a position consistent with the direction of the geological environment data set coordinate system. The translation vector is used to adjust the position of the point in the new coordinate system so that it corresponds to the origin of the geological environment data set coordinate system.

[0029] Taking a specific point P as an example, assume its coordinates in the coordinate system of the cross-section contour point cloud data set (denoted as coordinate system A) are (x1, y1, z1). First, perform a rotation operation. The rotation matrix R is a 3×3 matrix that contains the rotation angle and direction information. Write the coordinates of point P in column vector form and perform matrix multiplication with the rotation matrix R. The process of matrix multiplication is that the elements of the first row of the rotation matrix are multiplied by the elements of the column vector of point P's coordinates and then added together to obtain the X coordinate of the rotated point P'; the elements of the second row are multiplied by the elements of the column vector of point P's coordinates and then added together to obtain the Y coordinate; the elements of the third row are multiplied by the elements of the column vector of point P's coordinates and then added together to obtain the Z coordinate. Assume the coordinates of the rotated point P' are (x2, y2, z2).

[0030] Next, perform a translation operation. The translation vector T is a three-dimensional vector (tx, ty, tz), which represent the translation distances in the X, Y, and Z directions respectively. Add the corresponding components in the translation vector T to the coordinates of the rotated point P', that is, add tx to x2, add ty to y2, and add tz to z2, to obtain the point P'' (x2 + tx, y2 + ty, z2 + tz) in the coordinate system of the geological environment data set (denoted as coordinate system B).

[0031] For each point in the cross-section contour point cloud data set, the above rotation and translation operations need to be performed in sequence. In actual operations, computer programming can be used to efficiently complete this process. For example, using the NumPy library in Python, store the point cloud data as a two-dimensional array, where each row represents the coordinates of a point. By looping through each row in the array, perform rotation and translation calculations for each point, and finally achieve the coordinate system conversion of the entire point cloud data set.

[0032] Step S1212: Identify abnormal points in the cross-section contour point cloud data set and remove the abnormal points, and complete the missing area after removing the abnormal points through a neighboring point interpolation algorithm that adjusts the interpolation weight based on the surrounding rock grade parameters, to obtain a denoised cross-section contour point cloud data set.

[0033] In this embodiment, after completing the coordinate system alignment, there may be abnormal points in the cross-section contour point cloud data set. The reasons for the generation of these abnormal points are diverse. For example, during the laser scanning process, dust and water vapor in the surrounding environment will scatter the laser, resulting in deviations in the measured point positions; problems with the accuracy of the equipment itself may produce inaccurate measurement results under complex geological conditions; operator errors, such as inaccurate installation positions of the equipment, etc.

[0034] The distance threshold method is used to identify abnormal points. For each point in the cross-section contour point cloud data set, its adjacent points need to be determined. The adjacent point range can be determined by setting a neighborhood radius. For example, all points within a spherical space with a radius of 0.5 meters centered on this point are regarded as adjacent points.

[0035] After determining adjacent points, calculate the distances from this point to all adjacent points, and then find the average value of these distances. Assume that the distance threshold is set to 0.2 meters. If the average distance from a certain point to its adjacent points is greater than 0.2 meters, this point is considered an abnormal point and is removed from the point cloud data set.

[0036] After removing abnormal points, missing areas will appear in the point cloud data set, which need to be filled. The nearest neighbor interpolation algorithm based on adjusting the interpolation weight according to the surrounding rock grade parameters is adopted.

[0037] For different surrounding rock grades, the physical properties and stability of the rock are different, which affect the reliability of the laser scanning data. In soft rock areas, the rock structure is loose and the stability is poor. The laser scanning data may have large errors and low data reliability. Therefore, increase the interpolation weight, such as setting it to 0.8. In hard rock areas, the rock structure is compact and the stability is good. The laser scanning data is accurate and the data reliability is high. Set the interpolation weight to 0.2.

[0038] For each missing point in the missing area, find its surrounding nearest neighbor points. Assume that there are three nearest neighbor points P1(x1, y1, z1), P2(x2, y2, z2) and P3(x3, y3, z3) around a missing point, and assign interpolation weights w1, w2 and w3 respectively, and w1 + w2 + w3 = 1.

[0039] Calculate the coordinates of the missing point through weighted average. The X coordinate of the missing point is equal to w1 times the X coordinate of P1 plus w2 times the X coordinate of P2 plus w3 times the X coordinate of P3; the Y coordinate and Z coordinate are calculated in the same way. Calculate all the missing points in the missing area in this way to complete the filling of the missing area and obtain the denoised cross-section contour point cloud data set.

[0040] Step S122: Generate an initial cross-section contour curve based on the denoised cross-section contour point cloud data set, and define a key control node set including node control parameters on the initial cross-section contour curve, where the node control parameters include coordinate parameters, curvature parameters and tangent direction parameters.

[0041] Step S1221: Use a B-spline curve to interpolate and fit the denoised cross-section contour point cloud data set to generate an initial cross-section contour curve.

[0042] In this embodiment, the B-spline curve has good locality and smoothness and is suitable for fitting point cloud data. Use the denoised cross-section contour point cloud data as the control points of the B-spline curve, and these control points determine the general shape of the curve.

[0043] The order of the B-spline curve is selected as 3 because the B-spline curve of order 3 can better fit the point cloud data while ensuring the smoothness of the curve. The knot vector determines the transition mode of the curve between different control points and needs to be determined according to the distribution of the point cloud data.

[0044] Interpolation fitting is performed through the calculation formula of the B-spline curve. During the calculation process, the positions of the control points and the values of the knot vector are continuously adjusted to make the curve approximate the point cloud data as much as possible. The method for evaluating the fitting effect is to calculate the error between the curve and the point cloud data, and the error is the distance from the points on the curve to the corresponding points in the point cloud data.

[0045] To minimize the error, an iterative method is adopted. In each iteration, the positions of the control points and the values of the knot vector are fine-tuned according to the current error situation, and then the curve and the error are recalculated. After multiple iterations, when the error reaches an acceptable range, the initial cross-section contour curve that accurately fits the denoised cross-section contour point cloud data set is obtained.

[0046] Step S1222: Calculate the local curvature parameters of each data point on the initial cross-section contour curve, and identify geometric feature points according to the mutation amplitude of the curvature parameters.

[0047] In this embodiment, after generating the initial cross-section contour curve, the local curvature parameters of each data point on the curve need to be calculated. The curvature parameter reflects the degree of curvature of the curve at a certain point and is very important for analyzing the geometric features of the curve.

[0048] The numerical differentiation method is used to calculate the first derivative and the second derivative. For a point P on the curve, two adjacent points P1 and P2 before and after it are selected, and the first derivative and the second derivative of this point are approximately calculated by the finite difference method. The first derivative represents the tangent slope of the curve at this point, and the second derivative is related to the degree of curvature of the curve.

[0049] The curvature parameter of this point is obtained by calculating the ratio of the first derivative and the second derivative. To accurately identify the positions where the curve shape changes greatly, the curvature mutation threshold is set to 0.1.

[0050] Traverse all the points on the initial cross-section contour curve, and compare the difference between the curvature parameter of each point and the curvature parameter of the adjacent point. If the difference between the curvature parameter of a certain point and the curvature parameter of the adjacent point is greater than 0.1, the curve shape changes greatly at this point, and it is regarded as a geometric feature point.

[0051] Step S1223: Screen the geometric feature points based on the surrounding rock grade parameters in the geological environment data set, and retain the feature points with curvature parameters lower than the first threshold in the soft rock area and the feature points with curvature parameters higher than the second threshold in the hard rock area.

[0052] In this embodiment, the mechanical properties of rock masses with different surrounding rock grades are different, which affects the shape of the cross-section contour curve. The geometric feature points are screened according to the surrounding rock grade parameters in the geological environment data set.

[0053] The surrounding rock grade parameters divide the rock masses around the cavern into soft rock areas and hard rock areas. Soft rocks have low strength and strong deformation ability. The curvature of the cross-section contour curve in the soft rock area should not be too large, otherwise it may cause excessive deformation or even failure of the rock mass. Set the first threshold to 0.05. For the geometric feature points located in the soft rock area, if their curvature parameters are lower than 0.05, it indicates that the curve has a small degree of bending, which conforms to the mechanical properties of soft rocks, and these feature points are retained.

[0054] Hard rocks have high strength and weak deformation ability. The cross-section contour curve in the hard rock area can have a large curvature change. Set the second threshold to 0.15. For the geometric feature points located in the hard rock area, if their curvature parameters are higher than 0.15, it indicates that the curve has a large degree of bending, which conforms to the mechanical properties of hard rocks, and these feature points are retained.

[0055] Through screening, the geometric feature points that do not meet the geological conditions are removed, and only the feature points that match the surrounding rock grade are retained.

[0056] Step S1224: Calculate the tangent direction parameters of the initial cross-section contour curve at the retained geometric feature points, and bind the coordinate parameters, corresponding curvature parameters, and tangent direction parameters of the geometric feature points into a set of key control nodes.

[0057] In this embodiment, after screening out the qualified geometric feature points, calculate the tangent direction parameters of the initial cross-section contour curve at these feature points. The tangent direction parameters reflect the tangent direction of the curve at this point and are very important for subsequent curve regulation and 3D modeling.

[0058] Use the method of numerical differentiation to calculate the first derivative of the curve at each retained geometric feature point, and the direction of this first derivative is the tangent direction. The specific approach is similar to that when calculating the curvature parameters. Select two adjacent points before and after this point, and approximately calculate the first derivative by the finite difference method.

[0059] Bind the coordinate parameters, corresponding curvature parameters, and tangent direction parameters of the geometric feature points together to form a key control node. For example, for the geometric feature point P(x, y, z), its curvature parameter is k, and the tangent direction is T(tx, ty, tz), then (x, y, z, k, tx, ty, tz) is a key control node.

[0060] Combine all the key control nodes corresponding to the retained geometric feature points to form a set of key control nodes.

[0061] Step S123: Apply constraint conditions to the curvature parameters of the key control node set according to the surrounding rock grade parameters in the geological environment data set to generate an initial section profile curve after constraint; wherein, the curvature parameters in the soft rock area are restricted to be lower than the first threshold, and the curvature parameters in the hard rock area are restricted to be higher than the second threshold.

[0062] In this embodiment, the curvature requirements of the section profile curve for rock masses of different surrounding rock grades are different. Soft rock has low strength and strong deformation ability. Excessive curvature of the section profile curve in the soft rock area may lead to excessive deformation or even failure of the rock mass. Therefore, the curvature parameters in the soft rock area are restricted to be lower than the first threshold, and the first threshold is set to 0.05.

[0063] For the key control nodes located in the soft rock area, if their curvature parameters are greater than 0.05, their curvature parameters are adjusted to 0.05 to ensure that the curve shape conforms to the mechanical properties of the soft rock.

[0064] Hard rock has high strength and weak deformation ability. The section profile curve in the hard rock area can have a large curvature change. The curvature parameters in the hard rock area are restricted to be higher than the second threshold, and the second threshold is set to 0.15.

[0065] For the key control nodes located in the hard rock area, if their curvature parameters are less than 0.15, their curvature parameters are adjusted to 0.15 to make full use of the mechanical properties of the hard rock.

[0066] After adjusting the curvature parameters of the key control node set, based on the adjusted key control node set, interpolation fitting is performed again using the B-spline curve. This process is similar to step S1221. By continuously adjusting the positions of the control points and the values of the knot vectors, the curve approximates the adjusted key control nodes. After multiple iterative calculations, an initial section profile curve after constraint is generated to make it more conform to the actual geological situation.

[0067] Step S124: Update the coordinate parameters of the key control node set according to the surrounding rock deformation parameters in the real-time monitoring data set to generate a dynamically adjusted section profile curve and a node control parameter set containing the updated parameters.

[0068] Step S1241: Convert the surrounding rock deformation parameters into coordinate offsets of the key control node set.

[0069] In this embodiment, the surrounding rock deformation parameters in the real-time monitoring data set reflect the deformation of the surrounding rock of the underground cavern in the actual engineering process. To make the section profile curve reflect the real-time deformation of the surrounding rock, it is necessary to convert the surrounding rock deformation parameters into coordinate offsets of the key control node set.

[0070] Establish the mapping relationship between the surrounding rock deformation parameters and the coordinate offset through pre-experimental data or theoretical analysis. For example, in the laboratory, simulate the loading experiments on different types of rock masses, and record the coordinate change rules of the key control nodes under different deformation conditions; or establish a rock mass mechanical model through theoretical analysis to deduce the relationship between the surrounding rock deformation and the coordinate change of the key control nodes.

[0071] For each key control node, calculate the coordinate offset according to the mapping relationship based on the surrounding rock deformation parameters at its location. For example, for the key control node P(x, y, z), the deformation amount of the surrounding rock at its location in the x direction is Δdx, the deformation amount in the y direction is Δdy, and the deformation amount in the z direction is Δdz. According to the mapping relationship, the coordinate offset of this key control node in the x direction is Δx, the coordinate offset in the y direction is Δy, and the coordinate offset in the z direction is Δz.

[0072] Step S1242: Perform filtering processing on the coordinate offset to generate the updated amount of the smoothed node coordinates.

[0073] In this embodiment, since the real-time monitoring data may contain noise and mutation values, which will cause abnormal fluctuations in the coordinate offset and affect the smoothness and accuracy of the cross-section contour curve, it is necessary to perform filtering processing on the coordinate offset.

[0074] Adopt the moving average filtering method, and select the filtering window size to be 5. For the coordinate offset sequence of each key control node, take the average value of its adjacent 5 coordinate offsets as the smoothed coordinate offset of this point.

[0075] For example, the coordinate offset sequence of a key control node in the x direction is [Δx1, Δx2, Δx3, Δx4, Δx5, Δx6, …]. For the 3rd coordinate offset Δx3, its smoothed coordinate offset is (Δx1 + Δx2 + Δx3 + Δx4 + Δx5) divided by 5.

[0076] Perform such filtering processing on the coordinate offsets of all key control nodes in the three directions of X, Y, and Z to obtain the updated amount of the smoothed node coordinates, which can effectively remove the influence of noise and mutation values.

[0077] Step S1243: Adjust the coordinate parameters of the key control node set according to the updated amount of the smoothed node coordinates, and re-fit the initial cross-section contour curve based on the curvature parameter and tangent direction parameter of the key control node set to generate the dynamically adjusted cross-section contour curve.

[0078] In this embodiment, after obtaining the smoothed node coordinate update amounts, the coordinate parameters of the key control node set are adjusted according to these update amounts. For each key control node, its original coordinate parameters are added with the smoothed node coordinate update amounts to obtain the updated coordinate parameters.

[0079] For example, for the key control node P(x, y, z), its smoothed node coordinate update amount is (Δx, Δy, Δz), and the updated coordinate parameters are (x + Δx, y + Δy, z + Δz).

[0080] After updating the coordinate parameters of the key control nodes, based on the curvature parameters and tangent direction parameters of the key control node set, the updated key control nodes are interpolated and fitted again using a B-spline curve. This process is similar to the interpolation and fitting process for generating the initial cross-section profile curve. By continuously adjusting the positions of the control points and the values of the node vectors, the curve approximates the updated key control nodes. After multiple iterative calculations, a dynamically adjusted cross-section profile curve is generated, which can accurately reflect the real-time deformation of the surrounding rock.

[0081] Step S1244: Store the updated coordinate parameters, curvature parameters, and tangent direction parameters as the node regulation parameter set.

[0082] In this embodiment, after generating the dynamically adjusted cross-section profile curve, the coordinate parameters, curvature parameters, and tangent direction parameters of the updated key control nodes are stored as the node regulation parameter set. This node regulation parameter set contains the latest parameter information of all key control nodes. These node regulation parameters are stored in a database or a file for subsequent calling and analysis.

[0083] Step S130: Generate a stretching path baseline based on the chamber axis coordinate parameters, and perform an axial stretching operation on the dynamically adjusted cross-section profile curve along the stretching path baseline to generate a three-dimensional chamber surface containing a set of local three-dimensional surfaces.

[0084] Step S131: Generate an initial stretching path baseline based on the chamber axis coordinate parameters in the geological environment data set.

[0085] In this embodiment, the chamber axis coordinate parameters detail the three-dimensional coordinate information of the chamber axis at different positions. For example, the coordinate of the chamber axis at position 1 is (x1, y1, z1), the coordinate at position 2 is (x2, y2, z2), and so on. There are multiple such coordinate points.

[0086] Thus, these discrete coordinate points can be connected in sequence. Taking the first point P1(x1, y1, z1) and the second point P2(x2, y2, z2) as an example, in three-dimensional space, a line segment is formed from P1 to P2. The connection method here is based on the principle that two points in space determine a straight line. By finding the corresponding positions of these two points in a three-dimensional rectangular coordinate system and then connecting them with a straight line. Then, another line segment is formed from P2 to P3(x3, y3, z3), and so on, finally forming a broken line composed of multiple line segments, which initially presents the general trend of the chamber axis.

[0087] In actual operation, with the help of three-dimensional drawing software, these coordinate points can be input into the three-dimensional drawing software, and the three-dimensional drawing software can automatically connect these points in sequence to form a broken line, which is convenient for visually observing the initial shape of the chamber axis.

[0088] Among them, in order to make this broken line more conform to the actual chamber trend, it needs to be smoothed. Here, the spline interpolation method is adopted. Spline interpolation can generate a smooth curve according to the nodes of the broken line. Specifically, spline interpolation is a method of constructing a function through known nodes, which can make the generated curve transition naturally at the nodes and avoid sharp corners.

[0089] First, select a suitable spline function, such as a cubic spline function. The cubic spline function has good smoothness and local control characteristics, and is suitable for fitting complex curves such as chamber axes. For three adjacent points Pi, Pi+1, Pi+2, the cubic spline function can calculate a curve segment that can smoothly transition between these three points according to the coordinates of these three points and their relative position relationships.

[0090] The specific process of determining the curve segment is as follows: For each curve segment, four parameters need to be determined to describe the cubic spline function. The determination of these four parameters is based on the following conditions: The curve segment should pass through the corresponding coordinate points at the starting point and the ending point, and the first derivative (tangent slope) of the curve segment at the starting point and the ending point should be equal to the first derivative of the adjacent curve segment at this point, and the second derivative should also be continuous. These parameters are determined by solving a system of linear equations.

[0091] Assume that the cubic spline function between two adjacent points Pi and Pi+1 is S(t)=a + bt + ct² + dt³, where t is a parameter that varies between 0 and 1. When t = 0, S(0) is equal to the coordinates of Pi, and when t = 1, S(1) is equal to the coordinates of Pi+1. At the same time, according to the conditions of continuous first derivative and second derivative, a series of equations can be listed, and the values of a, b, c, and d are obtained by solving these equations.

[0092] Such calculations are performed on the curve segments between all adjacent points, and finally all the curve segments are joined together to obtain a smooth curve, which is the initial stretching path baseline, describing the orientation of the cavern in three-dimensional space.

[0093] Step S132: Divide the initial stretching path baseline into a homogeneous rock mass stretching segment and a heterogeneous rock mass stretching segment according to the surrounding rock stress distribution parameters, and set an initial stretching step parameter for each stretching segment.

[0094] In this embodiment, after generating the initial stretching path baseline, it is necessary to divide it according to the surrounding rock stress distribution parameters, which reflect the stress distribution of the rock mass around the cavern.

[0095] Specifically, by analyzing the surrounding rock stress distribution parameters in detail, the regions with uniform stress distribution and the regions with non-uniform stress distribution are identified. The specific judgment method is to divide the initial stretching path baseline into several small segments at equal intervals. Each small segment corresponds to a spatial region. Let the coordinates of the starting point and the ending point of each small segment on the initial stretching path baseline be Qi and Qi+1 respectively, and a series of data of the surrounding rock stress values are measured in the corresponding spatial region.

[0096] For the spatial region corresponding to each small segment, calculate the change rate of the surrounding rock stress in this region. The calculation method is as follows: First, find the maximum stress value Sm and the minimum stress value Sn in this region, and then use ((Sm - Sn) / Sn)*100% to obtain the stress change rate. This is based on the principle that the stress change rate can reflect the degree of uniform stress distribution. The smaller the change rate, the more uniform the stress distribution.

[0097] If the stress change rate of a certain region is less than 10%, it is considered that the stress distribution in this region is uniform, and the stretching path corresponding to this region is divided into a homogeneous rock mass stretching segment; if the stress change rate of a certain region is greater than 30%, it is considered that the stress distribution in this region is non-uniform, and the stretching path corresponding to this region is divided into a heterogeneous rock mass stretching segment. In actual operation, stress sensors can be used to monitor the stress at different positions around the cavern in real time, record the data, and then perform regional division according to the above method.

[0098] After the area division is completed, set the initial stretching step length parameter for each stretching segment. For the stretching segment of homogeneous rock mass, since its stress distribution is uniform and the deformation of the rock mass is relatively stable, a relatively large initial stretching step length parameter can be set, such as 1 meter. In this way, in the subsequent stretching operation, stretching can be carried out at a larger interval, reducing the amount of calculation. For the stretching segment of heterogeneous rock mass, since its stress changes greatly and the deformation of the rock mass is complex, more refined modeling is required. Therefore, a smaller initial stretching step length parameter is set, such as 0.5 meter, in order to more accurately simulate the rock mass situation in this area. Setting different stretching step length parameters is to improve the calculation efficiency on the premise of ensuring the accuracy of the model according to the characteristics of different rock mass areas.

[0099] Step S133: Dynamically reduce the initial stretching step length parameter within the heterogeneous rock mass stretching segment, and perform curvature continuity correction on the stretching path reference line of the arc area according to the curvature parameter in the node control parameter set to generate a corrected stretching path.

[0100] Step S1331: Generate a stress distribution map based on the surrounding rock stress distribution parameters, and identify the stress concentration area and the fault zone area.

[0101] In this embodiment, first, divide the three-dimensional space around the cavern into several small grid units, and each grid unit can be set as a cube with a side length of 0.1 meter. Taking a corner point of the cavern as the origin, divide the space according to the set rules to form regularly arranged grid units. The purpose of this division is to discretize the continuous three-dimensional space, facilitating the accurate measurement and analysis of the stress in each small area.

[0102] Then, use stress sensors to measure the surrounding rock stress values at the center positions of each grid unit, and use them as the stress representative values of the grid units. Assume that the center coordinates of the i-th grid unit are (xi, yi, zi), and the measured stress value is Si. The stress sensor can be a resistance strain gauge sensor, etc., which can convert stress into an electrical signal and obtain the stress value by measuring the magnitude of the electrical signal. Record the measured stress values for subsequent drawing of stress contour maps.

[0103] Next, draw a stress contour map based on these stress representative values. First, determine a stress value range, which can be determined according to the maximum and minimum values of all measured stress values. Then select a series of equally spaced stress values within this range as the values of the contour lines. For each contour line value, find all the points in the three-dimensional space where the stress representative values are equal to this value, and connect these points to form a contour line.

[0104] By continuously adjusting the values of the contour lines, a series of contour lines are drawn. The density of the contour lines reflects the rate of stress change. The area where the stress contour lines are dense indicates a large stress change, while the area where the stress contour lines are sparse indicates a small stress change. During the actual drawing process, professional drawing software such as Surfer software can be used. It can automatically draw the stress contour map based on the input stress data and can visually display the stress distribution.

[0105] By conducting a detailed analysis of the stress distribution atlas, stress concentration areas and fault zone areas are identified. Among them, the stress concentration area usually appears as the area where the stress contour lines are dense. When the spacing of the stress contour lines in a certain area is less than 0.5 meters, this area can be determined as the stress concentration area. This is because a small spacing of the stress contour lines indicates a drastic stress change in this area, and there is likely to be stress concentration. The fault zone area may appear as the area where the stress value changes abruptly. When the stress difference between adjacent grid cells is greater than 5 MPa, this area can be regarded as a possible fault zone. Because the fault will cause a sudden change in the mechanical properties of the rock mass, thus causing a sudden change in stress.

[0106] Step S1332: Mark the stress concentration area and the fault zone area as the heterogeneous rock mass tensile section, and reduce the initial tensile step size parameter to a preset ratio dynamically adjusted based on the curvature parameter in the node control parameter set.

[0107] In this embodiment, after identifying the stress concentration area and the fault zone area, the stress concentration area and the fault zone area are marked as the heterogeneous rock mass tensile section. Because the mechanical properties of the rock mass in the stress concentration area and the fault zone area change greatly, more refined modeling is required to accurately describe the situation. In actual operation, the parts corresponding to these areas on the initial tensile path baseline can be marked, for example, using different colors or symbols to distinguish the homogeneous rock mass tensile section and the heterogeneous rock mass tensile section.

[0108] The determination of the preset ratio is related to the curvature parameter in the node control parameter set. First, a curvature parameter threshold is set, for example, set to 0.2. The preset ratio is set in segments according to the size of the curvature parameter: when the curvature parameter is less than 0.2, the preset ratio is 0.8; when the curvature parameter is between 0.2 and 0.5, the preset ratio is 0.6; when the curvature parameter is greater than 0.5, the preset ratio is 0.4. This setting of the preset ratio is because the larger the curvature parameter, the greater the degree of curvature of the curve, and a more refined tensile step size is required to accurately describe the shape of the curve.

[0109] For each tensile point in the tensile section of the heterogeneous rock mass, according to the curvature parameter in the corresponding set of node control parameters, select the corresponding preset ratio. Then multiply the initial tensile step length parameter by this preset ratio to obtain the reduced tensile step length parameter. For example, if the initial tensile step length parameter is 0.5 meters, the curvature parameter of a certain tensile point is 0.3, and the corresponding preset ratio is 0.6, then the reduced tensile step length parameter is 0.5 × 0.6 = 0.3 meters. In actual calculations, the curvature parameter and the corresponding preset ratio of each tensile point can be stored in a table or an array, and the reduced tensile step length parameter can be automatically calculated through a program.

[0110] Step S1333: Perform curvature continuity correction on the tensile path reference line of the arc region according to the curvature parameter in the set of node control parameters to generate a corrected tensile path.

[0111] In this embodiment, in the tensile path reference line, the curvature change in the arc region may be discontinuous, which will affect the quality of the three-dimensional cavern surface. Therefore, it is necessary to perform curvature continuity correction on the tensile path reference line of the arc region according to the curvature parameter in the set of node control parameters.

[0112] Specifically, first, identify the arc region in the tensile path reference line. Determine the arc region by judging the change in the included angle between adjacent points. Let three adjacent points be Rj, Rj+1, and Rj+2, and calculate the included angle between vector RjRj+1 and vector Rj+1Rj+2. Here, the dot product formula of vectors can be used to calculate the included angle, that is , where θ is the included angle between the two vectors.

[0113] If the change in the included angle is greater than 10°, it is considered that the section from Rj to Rj+2 may be an arc region. Make such a judgment for all adjacent three points on the tensile path reference line to find all possible arc regions. In actual operation, a program can be written to traverse all points on the tensile path reference line, calculate the change in the included angle between adjacent points, and thus automatically identify the arc region.

[0114] For each point in the arc region, according to the curvature parameter in the corresponding set of node control parameters, adjust the position of this point so that the curvature change in the arc region is more continuous. Let this point be Pk, and its adjacent points be Pk-1 and Pk+1, and calculate the curvature parameter kP of Pk and the curvature parameters kP-1 and kP+1 of Pk-1 and Pk+1.

[0115] Calculate the curvature difference Δk = |kP - kP-1| and Δk' = |kP - kP+1|. If Δk or Δk' is greater than the preset curvature continuity threshold (such as 0.05), then fine-tune the position of Pk according to the magnitude and direction of the difference.

[0116] If \(kP>kP - 1\) and \(\Delta k>0.05\), move \(Pk\) a certain distance in the direction of decreasing curvature. The moving distance is determined according to the size of the difference. The larger the difference, the greater the moving distance. For example, the moving distance \(d = m\times\Delta k\) can be set, where \(m\) is a coefficient determined according to the actual situation, such as \(0.1\). In the actual adjustment process, the method of three-dimensional coordinate transformation can be used to move the position of the point to ensure that the adjusted point is still within a reasonable range.

[0117] After performing the above position adjustment on all points within the arc region, connect the adjusted points in sequence to obtain the corrected stretching path, and the curvature change of its arc region is smoother. In the process of connecting points, methods such as spline interpolation can also be used to ensure that the corrected stretching path is still a smooth curve.

[0118] Step S134: Perform a segmented stretching operation on the dynamically adjusted cross-sectional profile curve along the corrected stretching path to generate a set of local three-dimensional surfaces including a homogeneous rock mass stretching section surface and a heterogeneous rock mass stretching section surface.

[0119] In this embodiment, after obtaining the corrected stretching path, perform a segmented stretching operation on the dynamically adjusted cross-sectional profile curve along the corrected stretching path.

[0120] Specifically, for the homogeneous rock mass stretching section, stretch according to the set initial stretching step parameter (such as 1 meter). Starting from the starting point of the stretching path, let the current stretching point be \(Ts\), and the dynamically adjusted cross-sectional profile curve be \(C\).

[0121] Determine the tangent direction of the stretching path at \(Ts\). Here, the tangent direction can be obtained by calculating the first derivative of the stretching path at this point. Move the cross-sectional profile curve \(C\) 1 meter along this tangent direction. During the movement, keep the shape of the cross-sectional profile curve \(C\) unchanged, and only perform translation and rotation in space position to make it consistent with the direction of the stretching path. The specific translation and rotation operations can be realized using a three-dimensional transformation matrix.

[0122] Repeat the above operation, perform stretching every 1 meter. Each stretching generates a new position of the cross-sectional profile curve. Connect the cross-sectional profile curves at these different positions in sequence to form the homogeneous rock mass stretching section surface. In the connection process, the method of surface fitting can be used to ensure that the formed surface is smooth and continuous.

[0123] For the heterogeneous rock mass stretching section, stretch according to the dynamically reduced stretching step parameter (such as 0.3 meter). Starting from the position where the stretching path enters the heterogeneous rock mass stretching section, let the current stretching point be \(Tn\). Similarly, move the dynamically adjusted cross-sectional profile curve \(C\) 0.3 meter along the tangent direction of the stretching path at \(Tn\).

[0124] During the movement, keep the shape of the cross-section contour curve C unchanged, and translate and rotate it in space to adapt to the direction of the stretching path. Repeat this operation, perform stretching every 0.3 meters, and connect the cross-section contour curves at different positions in sequence to form the stretching section surface of the heterogeneous rock mass. In this process, due to the small stretching step, stretching operations need to be performed more frequently to ensure that the surface shape of the heterogeneous rock mass area can be accurately described.

[0125] Combine the generated stretching section surface of the homogeneous rock mass and the stretching section surface of the heterogeneous rock mass together to form a local three-dimensional surface set, which completely presents the three-dimensional surface morphology of the cavern in different rock mass areas. During the combination process, it is necessary to ensure the continuity and smoothness of the two surfaces at the splicing position, which can be achieved using a surface splicing algorithm, such as a splicing method based on spline curves, so that the combined local three-dimensional surface set conforms to the actual situation both visually and physically.

[0126] Step S140: Perform a cross-section stitching operation on the local three-dimensional surface set, generate a connection area topological structure based on the joint direction parameters in the geological environment data set, and perform topological optimization processing on the connection areas of adjacent surfaces to generate an optimized three-dimensional cavern model.

[0127] Step S141: Perform boundary matching processing on the surfaces of adjacent stretching sections in the local three-dimensional surface set to generate a local three-dimensional surface set after boundary matching; wherein, the boundary matching processing eliminates the curvature mutation in the splicing area through a smoothing algorithm.

[0128] In this embodiment, in the local three-dimensional surface set, problems such as curvature mutation may exist at the splicing positions of the surfaces of adjacent stretching sections, affecting the quality of the three-dimensional cavern model. Therefore, it is necessary to perform boundary matching processing on the surfaces of adjacent stretching sections, and a spline interpolation smoothing algorithm can be used to eliminate the curvature mutation in the splicing area.

[0129] First, extract the boundary point coordinates of the surfaces of adjacent stretching sections to form a boundary point set. For each boundary point in the splicing area, find its corresponding point on the boundary of the adjacent surface. Then, use the spline interpolation algorithm to generate a smooth transition curve between these corresponding points. For example, for the boundary points P1 and P2 of two adjacent surfaces, select a suitable spline function, such as a cubic spline function, and calculate a series of points on the transition curve according to the coordinates of P1, P2, and their adjacent points, so that the curvature change of the transition curve is smooth at P1 and P2.

[0130] By performing the above processing on all the boundary points in the splicing area, the curvature change at the splicing positions of the surfaces of adjacent stretching sections becomes more natural, and a local three-dimensional surface set after boundary matching is generated.

[0131] Step S142: Based on the cross-chamber position parameters and joint direction parameters in the geological environment data set, generate a topological structure of a connection area in the cross-chamber area that forms a preset angle with the joint direction; wherein, the topological structure of the connection area includes a main extension direction parameter and an auxiliary connection line parameter.

[0132] Step S1421: Extract the boundary coordinates of the intersection area according to the cross-chamber position parameters, and generate the intersection area range.

[0133] In this embodiment, the cross-chamber position parameters are extracted from the geological environment data set. The cross-chamber position parameters record the specific position information of the cross-chamber in the three-dimensional space. According to these specific position information, the boundary points of the intersection area can be determined.

[0134] For example, by analyzing the axis coordinates and section dimensions of the cross-chamber, find the coordinates of the outermost edge points of the intersection area, and connect the coordinates of these outermost edge points to form a closed polygon, which is the intersection area range. During specific operation, the center line of the cross-chamber can be determined first, and then the boundary point positions can be determined by expanding a set distance from the center line to both sides according to the section dimensions. By processing the position parameters of all cross-chambers, the complete intersection area range can be obtained.

[0135] Step S1422: Determine the main extension direction of the intersection area range based on the joint direction parameters, and generate an auxiliary extension direction that forms a fixed angular difference with the main extension direction according to the preset angle.

[0136] In this embodiment, the joint direction parameters reflect the strike of the joints in the rock mass. In the intersection area range, the main extension direction is determined according to the joint direction parameters. The main extension direction can be obtained by statistically analyzing the main strike of the joints in the intersection area.

[0137] For example, use joint measurement equipment to obtain the joint direction data of multiple points in the intersection area, and statistically analyze the joint direction with the highest occurrence frequency as the main extension direction. The preset angle is determined according to the actual engineering requirements and geological conditions. Assume the preset angle is 30°. Based on the main extension direction, rotate 30° clockwise or counterclockwise to obtain the auxiliary extension direction. When determining the main extension direction and the auxiliary extension direction, the distribution characteristics of the joints and the mechanical properties of the rock mass should be considered to ensure the rationality of the topological structure of the connection area.

[0138] Step S1423: Adjust the length parameter of the auxiliary connection line based on the tangent direction parameter in the node control parameter set, and construct a topological structure of the connection area including the Voronoi diagram structure.

[0139] In this embodiment, after determining the main extension direction and the auxiliary extension direction, it is necessary to construct the topological structure of the connection area. The length parameter of the auxiliary connection line is adjusted according to the tangent direction parameter in the node control parameter set.

[0140] For each node in the intersection area, its tangent direction parameter reflects the tangent direction of the curve at that node. According to the tangent direction parameter, the connection mode and length of the auxiliary connection line with the node are determined. For example, if the included angle between the tangent direction of the node and the main extension direction is large, the length of the auxiliary connection line is appropriately increased to better connect adjacent nodes. When constructing the topological structure of the connection area including the Voronoi diagram structure, the nodes in the intersection area are used as the generating points of the Voronoi diagram. According to the generation rules of the Voronoi diagram, the intersection area is divided into multiple Voronoi polygons, and each polygon corresponds to a generating point. The auxiliary connection lines connect the generating points of adjacent Voronoi polygons to form the topological structure of the connection area. By adjusting the length parameter of the auxiliary connection line, the topological structure of the connection area can better meet the geological conditions and engineering requirements.

[0141] Step S143: Perform mesh division processing on the set of locally three-dimensional surfaces after boundary matching according to the surrounding rock grade parameter to generate a three-dimensional cavern model including a dense mesh area; wherein, the density parameter of the dense mesh area is dynamically adjusted based on the curvature parameter in the node control parameter set.

[0142] In this embodiment, after obtaining the set of locally three-dimensional surfaces after boundary matching, mesh division processing is performed on it according to the surrounding rock grade parameter. Different mesh division strategies are adopted for areas with different surrounding rock grades. In the soft rock area, due to the poor stability of the rock, a finer mesh division is required to accurately simulate the deformation and failure of the rock mass. Therefore, the mesh is appropriately densified in the soft rock area.

[0143] For example, the mesh density in the soft rock area is set to twice that in the hard rock area. The density parameter of the dense mesh area is dynamically adjusted based on the curvature parameter in the node control parameter set. When the curvature parameter in the node control parameter set is large, it indicates that the curve in this area changes violently, and a denser mesh is required to accurately describe the surface shape. For example, a curvature parameter threshold is set, and when the curvature parameter is greater than this threshold, the mesh density is increased by 50%. By performing mesh division processing on the entire set of locally three-dimensional surfaces after boundary matching, a three-dimensional cavern model including a dense mesh area is generated.

[0144] Step S144: Convert the three-dimensional cavern model into construction-adapted format data including support structure parameters, and dynamically correct the coordinate parameters of the three-dimensional cavern model according to the real-time displacement parameters collected by the construction monitoring equipment to generate an optimized three-dimensional cavern model.

[0145] In this embodiment, the generated three-dimensional cavern model is converted into construction-adapted format data including support structure parameters. First, support structure parameters are extracted from the geological environment data set and the design document, such as the positions, lengths, and diameters of anchor bolts, the layout of cable bolts, the thickness and materials of linings, etc. These support structure parameters are integrated with the three-dimensional cavern model and converted into CAD format data commonly used in construction or other adapted format data. For example, the geometric information and mechanical parameters of the support structure are added to the corresponding nodes and elements of the three-dimensional cavern model.

[0146] Then, the coordinate parameters of the three-dimensional cavern model are dynamically corrected according to the real-time displacement parameters collected by the construction monitoring equipment. Construction monitoring equipment, such as displacement sensors, inclinometers, etc., monitors the displacement changes of the cavern during the construction process in real time. These real-time displacement parameters are fed back into the three-dimensional cavern model to adjust the coordinate parameters of each node in the three-dimensional cavern model. For example, for a node, its coordinates are moved in the corresponding direction according to the real-time displacement parameters at its location. By continuously feeding back the real-time displacement parameters into the three-dimensional cavern model, the three-dimensional cavern model becomes more in line with the actual construction situation, and an optimized three-dimensional cavern model is generated.

[0147] Based on the above steps, the embodiment of the present invention realizes a significant improvement in the accuracy, adaptability, and intelligence level of the three-dimensional modeling of underground caverns through the deep fusion and dynamic regulation mechanism of multi-source perception data. Specifically, by integrating the cross-section contour point cloud data obtained by laser scanning, the geological environment parameters provided by geological exploration, and the real-time monitoring data, the introduction of dynamically parameterized nodes enables the generation of the cross-section contour curve not to be limited to a static template, but to be adaptively adjusted according to key parameters such as the surrounding rock grade and stress distribution in the geological environment data set. This not only improves the accuracy of curve fitting but also endows the model with the ability to dynamically respond to complex geological conditions, effectively solving the problem that it is difficult to balance geometric accuracy and geological adaptability in traditional modeling methods. Further, based on the stretching path reference line generated from the cavern axis coordinate parameters, axial stretching is performed in combination with the dynamically adjusted cross-section contour curve to form a three-dimensional cavern surface containing rich local three-dimensional surface features. This not only retains the fine details of the cavern geometric shape but also achieves global consistency of the model through parametric regulation. Particularly importantly, in the cross-section stitching stage, the joint direction parameters in the geological environment data are used to generate the topological structure of the connection area, and the topological optimization of the adjacent surface connection area is carried out, which not only ensures the continuity of the model in physical connection but also optimizes the model structure from the perspective of geomechanics, making the finally generated three-dimensional cavern model highly unified in geometric shape, geological adaptability, and mechanical rationality. Thus, the accuracy, efficiency, and intelligence level of the three-dimensional modeling of underground caverns are significantly improved.

[0148] Figure 2 FIG. 2 shows a schematic diagram of exemplary hardware and software components of an underground cavern three-dimensional modeling system 100 based on dynamic parametric node regulation that can implement the inventive concept provided by some embodiments of the present invention. For example, the processor 120 can be used on the underground cavern three-dimensional modeling system 100 based on dynamic parametric node regulation and is used to execute the functions in the present invention.

[0149] The underground cavern three-dimensional modeling system 100 based on dynamic parametric node regulation can be a general-purpose server or a special-purpose server, both of which can be used to implement the method for underground cavern three-dimensional modeling based on dynamic parametric node regulation of the present invention. Although only one server is shown in the present invention, for convenience, the functions described in the present invention can be implemented in a distributed manner on multiple similar platforms to balance the processing load.

[0150] For example, the underground cavern three-dimensional modeling system 100 based on dynamic parametric node regulation can include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and different forms of storage media 140, such as disks, ROM, or RAM, or any combination thereof. Exemplarily, the underground cavern three-dimensional modeling system 100 based on dynamic parametric node regulation can also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The method of the present invention can be implemented according to these program instructions. The underground cavern three-dimensional modeling system 100 based on dynamic parametric node regulation further includes an input / output (I / O) interface 150 between the computer and other input / output devices.

[0151] For ease of explanation, only one processor is described in the underground cavern three-dimensional modeling system 100 based on dynamic parametric node regulation. However, it should be noted that the underground cavern three-dimensional modeling system 100 in the present invention can also include multiple processors. Therefore, the steps executed by one processor described in the present invention can also be jointly executed or separately executed by multiple processors. For example, if the processor of the underground cavern three-dimensional modeling system 100 based on dynamic parametric node regulation executes steps A and B, it should be understood that steps A and B can also be jointly executed by two different processors or separately executed in one processor. For example, the first processor executes step A, the second processor executes step B, or the first processor and the second processor jointly execute steps A and B.

[0152] In addition, an embodiment of the present invention further provides a readable storage medium, in which computer-executable instructions are preset. When a processor executes the computer-executable instructions, the above three-dimensional underground cavern modeling method based on dynamic parameterized node regulation is implemented.

[0153] It should be noted that, in order to simplify the description of the present invention disclosure and thus help the understanding of one or more embodiments of the invention, in the foregoing description of the embodiments of the present invention, sometimes multiple features are merged into one embodiment, drawing or description thereof.

Claims

1. A three-dimensional modeling method for underground caverns based on dynamic parameterized node control, characterized in that: The method comprises: Acquire a multi-source sensing data set of an underground cavern, wherein the multi-source sensing data set includes a cross-sectional contour point cloud data set collected by a laser scanning device, a geological environment data set collected by a geological exploration device, and a real-time monitoring data set collected by a monitoring sensor; wherein the geological environment data set includes surrounding rock grade parameters, surrounding rock stress distribution parameters, and cavern axis coordinate parameters; Generating an initial cross-sectional contour curve based on the cross-sectional contour point cloud data set, and regulating the initial cross-sectional contour curve through a dynamic parameterized node to generate a dynamically adjusted cross-sectional contour curve and a set of node regulation parameters; Generating a stretching path reference line based on the cavity axis coordinate parameters, performing an axial stretching operation on the dynamically adjusted cross-section contour curve along the stretching path reference line, and generating a three-dimensional cavity surface including a local three-dimensional surface set; A section stitching operation is performed on the local three-dimensional surface set, a connection area topology structure is generated based on the joint direction parameters in the geological environment data set, and a topological optimization process is performed on the connection areas of adjacent surfaces to generate an optimized three-dimensional cave model.

2. The underground cavern three-dimensional modeling method based on dynamic parameterized node control according to claim 1 is characterized in that: The generating of the initial cross-sectional contour curve based on the cross-sectional contour point cloud data set, and regulating the initial cross-sectional contour curve through a dynamic parameterized node to generate a dynamically adjusted cross-sectional contour curve and a set of node regulation parameters, includes: Performing noise reduction processing on the cross-section contour point cloud data set to obtain a noise-reduced cross-section contour point cloud data set; wherein the noise reduction processing includes an operation of eliminating abnormal points and completing missing points in a unified coordinate system based on coordinate system conversion parameters of the geological environment data set; Generate an initial cross-sectional contour curve based on the denoised cross-sectional contour point cloud data set, and define a key control node set including node control parameters on the initial cross-sectional contour curve, wherein the node control parameters include coordinate parameters, curvature parameters and tangent direction parameters; Applying constraint conditions to the curvature parameters of the key control node set according to the surrounding rock grade parameters in the geological environment data set to generate a constrained initial cross-section profile curve; wherein the curvature parameters of the soft rock area are restricted to be lower than a first threshold, and the curvature parameters of the hard rock area are restricted to be higher than a second threshold; The coordinate parameters of the key control node set are updated according to the surrounding rock deformation parameters in the real-time monitoring data set, and a dynamically adjusted cross-section profile curve and a node control parameter set containing the updated parameters are generated.

3. The underground cavern three-dimensional modeling method based on dynamic parameterized node control according to claim 2 is characterized in that: The step of performing noise reduction processing on the cross-section contour point cloud data set to obtain a noise-reduced cross-section contour point cloud data set includes: Performing coordinate conversion and alignment operations on the coordinate system of the cross-section contour point cloud data set and the coordinate system of the geological environment data set; Abnormal points are identified and removed from the cross-section contour point cloud data set, and the missing areas after the abnormal points are removed are supplemented by a neighboring point interpolation algorithm that adjusts the interpolation weight based on the surrounding rock grade parameter to obtain a denoised cross-section contour point cloud data set.

4. The underground cavern three-dimensional modeling method based on dynamic parameterized node control according to claim 2 is characterized in that: The generating of an initial cross-sectional contour curve based on the cross-sectional contour point cloud data set after noise reduction, and defining a key control node set including node control parameters on the initial cross-sectional contour curve, comprises: Using a B-spline curve to interpolate and fit the noise-reduced cross-section contour point cloud data set to generate an initial cross-section contour curve; Calculating the local curvature parameter of each data point on the initial cross-sectional contour curve, and identifying the geometric feature point according to the mutation amplitude of the curvature parameter; The geometric feature points are screened based on the surrounding rock grade parameters in the geological environment data set, and feature points with curvature parameters lower than a first threshold in a soft rock region and feature points with curvature parameters higher than a second threshold in a hard rock region are retained; The tangent direction parameters of the initial cross-section contour curve are calculated at the retained geometric feature points, and the coordinate parameters of the geometric feature points, the corresponding curvature parameters and the tangent direction parameters are bound as a key control node set.

5. The underground cavern three-dimensional modeling method based on dynamic parameterized node control according to claim 2 is characterized in that: The updating of the coordinate parameters of the key control node set according to the surrounding rock deformation parameters in the real-time monitoring data set, generating a dynamically adjusted cross-section profile curve and a node control parameter set containing updated parameters, includes: Converting the surrounding rock deformation parameters into coordinate offsets of the key control node set; Filtering the coordinate offset to generate smoothed node coordinate update; Adjusting the coordinate parameters of the key control node set according to the smoothed node coordinate update amount, and refitting the initial cross-section contour curve based on the curvature parameter and tangent direction parameter of the key control node set to generate a dynamically adjusted cross-section contour curve; The updated coordinate parameters, curvature parameters and tangent direction parameters are stored as the node control parameter set.

6. The underground cavern three-dimensional modeling method based on dynamic parameterized node control according to claim 1 is characterized in that: The step of generating a stretching path reference line based on the cavity axis coordinate parameters, performing an axial stretching operation on the dynamically adjusted cross-section contour curve along the stretching path reference line, and generating a three-dimensional cavity surface including a local three-dimensional surface set, comprises: Generate an initial stretching path baseline based on the cavern axis coordinate parameters in the geological environment data set; Dividing the initial stretching path baseline into a homogeneous rock mass stretching section and a heterogeneous rock mass stretching section according to the surrounding rock stress distribution parameters, and setting an initial stretching step length parameter for each stretching section; Dynamically reducing the initial stretching step length parameter in the inhomogeneous rock mass stretching section, and performing curvature continuity correction on the stretching path baseline of the arc region according to the curvature parameter in the node control parameter set to generate a corrected stretching path; A segmented stretching operation is performed on the dynamically adjusted cross-section contour curve along the corrected stretching path to generate a local three-dimensional surface set including a homogeneous rock mass stretching segment surface and a heterogeneous rock mass stretching segment surface.

7. The underground cavern three-dimensional modeling method based on dynamic parameterized node control according to claim 6 is characterized in that: The dynamically reducing the initial stretching step length parameter in the heterogeneous rock mass stretching section includes: Generate a stress distribution map based on the surrounding rock stress distribution parameters to identify stress concentration areas and fault zone areas; The stress concentration area and the fault zone area are marked as inhomogeneous rock mass stretching sections, and the initial stretching step parameter is reduced to a preset ratio dynamically adjusted based on the curvature parameter in the node control parameter set.

8. The underground cavern three-dimensional modeling method based on dynamic parameterized node control according to claim 6 is characterized in that: The section stitching operation is performed on the local three-dimensional surface set, a connection area topology structure is generated based on the joint direction parameters in the geological environment data set, and a topology optimization process is performed on the connection areas of adjacent surfaces to generate an optimized three-dimensional cave model, including: Performing boundary matching processing on the surfaces of adjacent stretching segments in the local three-dimensional surface set to generate a local three-dimensional surface set after boundary matching; wherein the boundary matching processing eliminates the sudden change of curvature in the splicing area through a smoothing algorithm; Based on the cross-cavern location parameters and joint direction parameters in the geological environment data set, a connection area topology structure that forms a preset angle with the joint direction is generated in the cross-cavern area; wherein the connection area topology structure includes a main extension direction parameter and an auxiliary connection line parameter; Meshing the local three-dimensional surface set after boundary matching according to the surrounding rock grade parameter to generate a three-dimensional cave model including an encrypted mesh area; wherein the density parameter of the encrypted mesh area is dynamically adjusted based on the curvature parameter in the node control parameter set; The three-dimensional cave model is converted into construction-adapted format data containing support structure parameters, and the coordinate parameters of the three-dimensional cave model are dynamically corrected according to the real-time displacement parameters collected by the construction monitoring equipment to generate an optimized three-dimensional cave model.

9. The underground cavern three-dimensional modeling method based on dynamic parameterized node control according to claim 8, characterized in that: The method of generating a connection area topology structure that forms a preset angle with the joint direction in the cross-cavern area based on the cross-cavern position parameters and the joint direction parameters in the geological environment data set includes: Extracting the boundary coordinates of the intersection area according to the intersection cave position parameters to generate the intersection area range; Determine the main extension direction of the intersection area based on the joint direction parameters, and generate an auxiliary extension direction with a fixed angle difference from the main extension direction according to the preset angle; The length parameters of the auxiliary connection lines are adjusted based on the tangent direction parameters in the node control parameter set to construct a connection area topology structure including a Voronoi diagram structure.

10. A three-dimensional modeling system for underground caverns based on dynamic parameterized node control, characterized in that: It includes a processor and a memory, the memory is connected to the processor, the memory is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the memory to implement the underground cavern three-dimensional modeling method based on dynamic parameterized node control as described in any one of claims 1 to 9.

Images