Methods for constructing digital models of automated mining faces

Abstract

This application relates to the field of intelligent mining technology and discloses a method for constructing a digital model of an automated mining face. The method includes: collecting the coordinates of roadway targets and the dip control points of the cutting face; selecting three points to construct a reference plane reflecting the strike and dip of the coal seam; calculating the normal offset of the dip control points relative to the reference plane, constructing an interpolation function and performing spatial mapping to generate a coal seam floor surface; recursively generating multi-segment surfaces along the advancing direction and constructing a three-dimensional solid model in conjunction with the coal thickness; during the mining process, calculating the error between the measured data and the model, decomposing the error into trend and fluctuation terms, and introducing spatial attenuation weights to dynamically correct unmined areas. This invention, by combining a reference plane and offset mapping, takes into account both macroscopic occurrence and local undulation characteristics, and achieves closed-loop updating of the model through error decomposition, thereby improving the accuracy and environmental adaptability of the geological model.

Description

Technical Field

[0001] This invention relates to the field of intelligent mining technology, specifically a method for constructing digital models of automated mining faces. Background Technology

[0002] The intelligent and unmanned operation of fully mechanized coal mining faces is the core trend in the current coal industry. The key lies in realizing the autonomous and coordinated control of fully mechanized mining equipment such as coal mining machines and hydraulic supports. Constructing a high-precision geological digital model of the working face is the basic prerequisite for fully mechanized mining equipment to autonomously plan cutting paths and support strategies. This model needs to accurately reflect the spatial geometry and thickness changes of the coal seam roof and floor to provide a navigation map for the equipment.

[0003] Currently, the construction of geological models for working faces mainly relies on geological exploration data or discrete measurement points obtained during the tunnel excavation stage. A common method is to use measurement data from the two transport roadways and return airway to generate simple planar or regular curved surface models through linear interpolation. However, the actual occurrence conditions of coal seams are often quite complex. Affected by geological structures and sedimentary environments, there are often nonlinear undulations inside the working face, such as a wave-like floor. Existing linear interpolation methods are difficult to accurately describe these local undulation characteristics, resulting in a large deviation between the generated model and the actual coal seam. When the coal mining machine cuts according to the model with large deviations, it is easy to cause rock cutting or coal loss, which not only reduces the recovery rate but also accelerates equipment wear and tear.

[0004] Furthermore, the geological conditions of coal seams are concealed and uncertain. As the working face advances, the prediction accuracy of the initially constructed static model will gradually decrease. Existing model correction techniques usually adopt simple overall translation or dip adjustment. This approach often ignores the difference between local geological changes and overall trends and lacks an effective spatial weighting mechanism. This leads to geometric discontinuities or over-correction in the corrected model in the transition area, which cannot meet the requirements of fully automated mining for continuous high accuracy of the model. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method for constructing a digital model of an automated mining face, which solves the problems that existing geological model construction methods cannot accurately describe the local undulations of coal seams, and that static models cannot perform high-precision, smooth dynamic closed-loop corrections as mining progresses.

[0006] The first aspect of this invention provides a method for constructing a digital model of an automated mining face, comprising the following steps: The data acquisition steps involve setting target control points along the bottom plate of the return airway and transport roadway of the working face under the unified coordinate system of the entire mine, and setting dip control points at the opening position to obtain the three-dimensional coordinate data of the target control points and the dip control points. The reference plane construction step involves selecting three non-collinear points from the target control points of the transport roadway and the return air roadway as reference points, calculating the plane vector formed by the reference points, determining the reference plane that reflects the strike and dip of the coal seam in the local area, and obtaining the normal vector and plane equation of the reference plane. The initial surface fitting step involves calculating the normal offset of the dip control point relative to the reference plane, constructing an interpolation function that describes the change of the offset with the dip position parameter, and spatially mapping the reference plane along the normal vector direction according to the interpolation function to generate the coal seam floor surface. The model generation steps involve sequentially executing the reference plane construction step and the initial surface fitting step along the working face advancement direction to generate multiple sections of coal seam floor surfaces and smoothly connect them. Combined with coal seam thickness data, solid stretching is performed to establish a fully automated mining working face digital model. The dynamic update step involves collecting the measured coordinates of the newly exposed working face during the mining process, calculating the error between the model's predicted value and the measured value, decomposing the error into a trend term and a local fluctuation term, and correcting the model data for subsequent unmined areas.

[0007] In the data acquisition step, the setting of the target control points includes: From the opening of the working face towards the stop line, physical markers are laid at the stable rock strata on the bottom of the roadway at preset intervals; The coordinate sequence of the target control points of the two roadways is obtained using measuring equipment, and the target control points of the transport roadway and the return airway are located on the same vertical cross section.

[0008] The reference plane construction steps include: Two adjacent target control points are selected on one side of the transport roadway, and a target control point corresponding to one of the target control points on the side of the return air roadway is selected on the other side to form three base points. Calculate the first vector along the direction and the second vector along the dip based on the three base points; The normal vector of the reference plane is calculated using the cross product of the first vector and the second vector; Based on the normal vector and the coordinates of any base point, establish the point normal equation of the reference plane.

[0009] The initial surface fitting step includes: Define the line connecting the base point on the side of the transport roadway and the base point on the side of the return air roadway as the parameterized baseline, and determine the normalized position parameters of each dip control point on the parameterized baseline. Substitute the coordinates of each dip control point into the plane equation of the reference plane, and calculate the algebraic distance from each dip control point to the reference plane as the normal offset. Using Lagrange interpolation or spline interpolation, an offset function that tends towards undulating shapes is constructed based on the normalized position parameters and the normal offset.

[0010] In the initial surface fitting step, the specific method for generating the coal seam floor surface is as follows: The reference plane is discretized into a grid to obtain the basic coordinates of the grid points on the reference plane; For each grid point, the corresponding correction value is calculated using the offset function; The correction value is multiplied by the normal vector of the reference plane and superimposed onto the basic coordinates to obtain the point coordinates on the curved surface of the coal seam floor.

[0011] The model generation steps include: The coal seam floor surface generated in adjacent sections is geometrically continuous at the joint, and the normal vector of the grid points near the boundary is smoothed by weighted averaging. Obtain coal seam thickness distribution data, and based on the smoothed coal seam floor surface, superimpose the coal seam thickness value along the normal vector direction of the surface to generate the roof surface. A closed three-dimensional solid model is formed by using the curved surface of the coal seam floor, the curved surface of the roof, and the side mesh.

[0012] In the dynamic update step, the specific method for calculating the error is as follows: Acquire local scanning data of newly exposed feature points in the longwall face, and combine it with the positioning data of the current coal mining equipment or support equipment to convert the local scanning data into measured coordinates in a unified coordinate system for the entire mine; Based on the measured coordinates of the planar position, the corresponding model elevation is retrieved in the digital model of the fully automated mining face; The spatial error field is obtained by calculating the difference between the measured elevation value and the model elevation.

[0013] In the dynamic update step, the specific methods for decomposing the error into trend terms and local fluctuation terms include: The spatial error field is fitted using a polynomial surface fitting method to obtain a trend term function characterizing the systematic deviation; Calculate the residual between the spatial error field and the trend term function; The residuals are spatially interpolated using a radial basis function interpolation algorithm to establish a local fluctuation term model that characterizes local detail corrections.

[0014] The correction of model data for subsequently unexploited areas includes: A spatial decay weighting function is constructed, the value of which decreases as the distance between the point to be corrected and the current measured coal wall increases; The sum of the trend term value obtained by substituting the coordinates of the position to be corrected into the trend term function and the local fluctuation term value obtained by substituting them into the local fluctuation term model is multiplied by the corresponding spatial decay weight function value to obtain the correction amount. The correction amount is then superimposed on the predicted elevation of the original model to complete the closed-loop update of the model.

[0015] A second aspect of the present invention provides a computer device including a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, it implements the method described in the present invention.

[0016] This invention provides a method for constructing a digital model of an automated mining face. It has the following beneficial effects: 1. This invention adopts a modeling strategy of reference plane and offset interpolation mapping. First, a reference plane is constructed based on the target control points of the transport roadway and return airway to lock the overall direction and dip of the working face. Then, by calculating the normal offset of the incision dip control point relative to the reference plane, an interpolation function describing the undulation pattern is constructed. This method combines macroscopic planar geometric constraints with microscopic nonlinear undulation characteristics, so that the generated coal seam floor surface can strictly pass through the control points on both sides of the roadway and accurately restore the wave-like undulation pattern inside the working face. This solves the problem that traditional planar modeling methods are difficult to describe complex geological details and significantly improves the geometric accuracy of the initial model.

[0017] 2. This invention introduces a dynamic update step based on error decomposition during the mining process. The deviation between the measured value and the model value is decomposed into a trend term representing systematic changes and a local fluctuation term representing local disturbances. The trend term is processed by polynomial fitting, and the fluctuation term is processed by radial basis function interpolation. Targeted corrections are made for each term. This mechanism can distinguish between the overall gradual change and the local abrupt change in geological conditions, effectively avoiding data oscillation or overfitting that may be caused by a single correction method. It realizes the continuous optimization of the model for unmined areas as mining progresses, and solves the technical problem of the static model accumulating deviations as the advance distance increases.

[0018] 3. In the model update process, this invention constructs a spatial decay weight function, which makes the weight of the correction amount decrease linearly or non-linearly as the distance between the point to be corrected and the measured coal wall increases. In the region close to the measured point, the model fully absorbs the latest measured error for full correction. In the deep region far from the measured point, the correction effect gradually weakens until it disappears. This processing method conforms to the spatial autocorrelation law of geological variables, prevents unreasonable and drastic changes to the distant model caused by local measurement data, and ensures the spatial continuity and smoothness of the digital model in the dynamic update process, providing a stable and reliable data foundation for the cutting path planning of the coal mining machine. Attached Figure Description

[0019] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a plan view of the two lanes selected for control targets according to the present invention; Figure 3 This is a schematic diagram illustrating the perpendicular distance between the calculated curve and the plane according to the present invention. Figure 4 This is a curve fitting diagram of the planar and incision eye control target of the present invention; Figure 5 This is a schematic diagram of the digital model of the coal seam in the working face of the present invention; Figure 6 This is a schematic diagram illustrating the changes in coal seam occurrence conditions according to the present invention; Figure 7 This is a schematic diagram of the computer device structure of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Example: Please see the appendix Figure 1 -Appendix Figure 6 This invention provides a method for constructing a digital model of an automated mining face, comprising the following steps: S100. After the working face is connected, a unified three-dimensional coordinate system is established for the entire mine. From the opening cut to the stop line, target control points are set on the stable floor of the return airway and transport roadway at a preset interval, and the three-dimensional coordinate data of each target control point is obtained using measuring equipment. S200. Select three non-collinear points from the target control points of the two roadways as base points, calculate the plane vector formed by the base points, determine the reference plane that reflects the dip and strike of the coal seam in the area, and obtain the normal vector and plane equation of the reference plane. S300. Select several control points representing the dip of the coal seam on the open cutting face, calculate the offset of each control point relative to the reference plane, use the interpolation algorithm to construct the functional relationship between the offset and the position parameters, and spatially map the reference plane along the normal vector direction according to the functional relationship to generate the coal seam floor surface. S400, along the advancing direction, the reference plane construction and surface fitting steps are executed sequentially to calculate the surface of the subsequent section, the surface of each section is smoothly connected, and the solid is stretched in combination with the coal seam thickness data to establish a digital model of the fully automated mining face. S500: During the mining process, the measured coordinates of the newly exposed working face are collected according to the preset advance distance. The error between the model prediction value and the measured value is calculated. The error is decomposed into trend term and local fluctuation term, and the model data of the subsequent unmined area is corrected.

[0022] Step S100 mainly involves the discretization and acquisition of spatial data of the working surface and the establishment of a coordinate system, providing accurate boundary constraints and spatial references for the subsequent geometric reconstruction of the digital model. The specific implementation process of this step is as follows: Establish a unified three-dimensional Cartesian coordinate system for the entire mine. After the fully mechanized mining face is completed, establish a spatial coordinate system based on the mine's existing measurement and control network. ,in, The axis is defined as the direction along the coal seam strike, i.e., the direction of face advance. An axis is defined as being along the dip of the coal seam, i.e., perpendicular to the direction of advance. The axis is defined as the vertical upward elevation direction. All control point coordinates, subsequently generated plane equations, and surface mesh nodes are described in the same coordinate system to ensure spatial consistency of multi-source data.

[0023] Implement the layout of two control points, starting from the opening position of the working face, along... In the positive direction of the axis (i.e., towards the stop line), target control points are set up in the transport roadway and return air roadway, with the spacing between them set as follows: meters, this spacing This is a variable parameter, typically ranging from 1 to 10 meters. The specific value is determined based on the complexity of the geological structure of the coal seam. In areas with simple geological structures, the value can be appropriately increased. The value decreases in areas with faults or folds. To improve control density, the location of target control points follows the principle of geological stability, avoiding the fractured zone of the floor or the area of ​​floating coal accumulation. Stable rock strata of the roadway floor are selected as the base. When laying out the points, laser rangefinders are used to assist in positioning so that the corresponding control points of the transport roadway and the return airway are located on the same cross section perpendicular to the strike, thereby ensuring that the subsequent construction of the reference plane can accurately reflect the coal seam occurrence near the cross section.

[0024] The control points are physically marked and solidified. Loose coal and debris are cleared from the selected base plate location until the original rock layer is exposed. A positioning hole is drilled vertically downward at the marked point using a pneumatic drill or electric drill. The drilling depth is set to 50mm to 100mm and the hole diameter is 10mm to 20mm. Special expansion bolts or grouting anchors are installed in the drilled hole, and a measuring base with a standardized interface is set on the top of the bolt. This physical marking method can resist the vibration interference caused by the deformation of the roadway bottom and the operation of equipment, ensuring that the spatial position of the control point remains relatively static during the service cycle.

[0025] Collect initial cross-section and roadway control point data, and on the cut-out cross-section, along the dip of the coal seam. Axial direction according to preset spacing For example, a row of initial control points is set up every 2 to 5 meters to capture the undulating shape of the coal seam floor at the starting position of the working face. A total station or high-precision electronic theodolite is used for data measurement. During measurement, the instrument is set up at a station with known coordinates, and an azimuth benchmark is established through backsight orientation. Then, prisms are placed on the positioning bolts of each target control point. The instrument is aimed at the center of the prism and observed. The three-dimensional coordinate data of each control point are calculated and recorded one by one. The operation procedures, error control, and adjustment processing of the total station are standard techniques in engineering surveying, ensuring that the measurement accuracy meets the requirements for millimeter-level modeling. Finally, the coordinate sequence of the control point set for the two roadways and the dipping control point set for the opening is obtained, recorded in the following format: This serves as the input boundary condition for subsequent surface fitting algorithms.

[0026] Step S200 mainly involves constructing a reference plane that reflects the local coal seam attitude characteristics using the collected discrete control points, and providing a reference basis for subsequent surface fitting by establishing the plane geometric equation. The specific implementation process of this step is as follows: Select the feature control points for constructing the reference plane. In the control point dataset obtained in step S100, the length of the working surface segment to be modeled is... Three non-collinear spatial points are selected as the base points for defining the plane. The specific selection strategy is as follows: Two adjacent control points are selected on one side of the transport roadway, denoted as . and Their coordinates are respectively and Select one side of the return airway with The corresponding control point is denoted as Its coordinates are The triangular area formed by these three points can cover the main spatial range of the working face in this section, where and The lines connecting them reflect the strike characteristics of the coal seam. and The lines connecting them reflect the dip characteristics of the coal seam.

[0027] Construct planar feature vectors by calculating two linearly independent vectors defining the plane based on the coordinates of three selected feature points. The first vector... Following the alleyway, from point to The second vector Along the dip of the working face, by point to The formula for vector calculation is as follows: ; ; Calculate the normal vector to the reference plane, and use the cross product principle to calculate the vector perpendicular to the reference plane. and The normal vector of the determined plane This normal vector represents the average normal direction of the coal seam floor in this section. Each component , , The calculation is as follows: ; ; ; In the above formula, Representing the normal vector in Projected components on the axis, Indicates in Projected components on the axis, Indicates in Projected components on the axis.

[0028] Establish the point normal equation of the reference plane, based on the calculated normal vector components. , , And any point on the plane, selected here. Given the coordinates, establish the algebraic equation of the reference plane: ; Among them, the constant term By point Substituting into the equation and solving, we get: ; This plane equation uniquely determines a plane that extends infinitely in space, and this plane passes through... , , Thirdly, in the subsequent model building steps, this plane not only serves as the projection reference for calculating the incised eye dip curve, but also as the geometric skeleton for generating the final coal seam floor surface, used to carry the local undulation details obtained later through interpolation algorithms.

[0029] Step S300 mainly involves transforming a simple geometric reference plane into a complex curved surface that conforms to the actual coal seam occurrence morphology. By introducing measured dip data at the cut-in point, an initial geological model is constructed using an interpolation mapping algorithm. The specific implementation process of this step is as follows: The dip direction of the open-cut working face is parameterized, and the control points connecting the transport roadways are... Control point of return airway The straight line segment is defined as the parametric baseline, and a normalization parameter is introduced. Its value range is ,in, Corresponding to the side of the transport roadway point, Corresponding to the return airway side Point, for the first point laid out on the cut-off section Individual control points (Total number is) (number), based on its in The projection position on the connection line determines the corresponding parameter value. When the control points are equidistantly distributed, the parameters The calculation is as follows: ; Calculate the algebraic offset of the dip control point relative to the reference plane, based on the reference plane equation determined in step S200. Calculate each tendency control point The perpendicular distance to this plane is denoted as This offset includes not only the distance but also directional information (positive or negative sign), used to characterize the undulation of the actual coal seam floor relative to the ideal plane. The calculation formula is as follows: ; In the formula, the denominator Normal vector The modulus, the numerator is the point coordinate, and the value is substituted into the plane equation. If This indicates that the control point is located on the side pointed to by the normal vector of the reference plane, and vice versa.

[0030] An interpolation function for the dip undulation morphology of the coal seam is constructed. To describe the continuous undulation changes across the entire dip section, a discrete offset dataset is used. Constructing smooth functions This embodiment uses the Lagrange interpolation method, first for each node. Constructing Lagrange basis functions : ; Then, each basis function is associated with its corresponding offset. By performing a linear combination, we obtain an offset function that can pass through all measured control points. : ; This function Describes the normalization parameters The elevation correction of the actual coal seam floor relative to the reference plane at the corresponding dip position.

[0031] The coal seam floor surface is generated based on normal mapping. The reference plane is parameterized into a two-dimensional mesh, and the offset function is used. The mesh points are stretched along the plane normal to reconstruct a 3D surface. Let the parameters be... Indicates along the direction ( Normalized position (direction), parameters Represents the normalized position along the dip, defining the parametric point on the reference plane. for: ; in For vectors , For vectors .

[0032] The final generated coal seam floor surface The formula for calculating the coordinates of any point on the graph is: ; In the formula, The normalized plane normal vector, i.e. Through this mapping transformation, the originally flat reference plane is transformed into a continuous smooth surface that passes through the boundary control points of the two lanes and accurately fits the undulation characteristics inside the cut eye, thus completing the modeling process from discrete points to digital surfaces.

[0033] Step S400 mainly involves segmental recursive modeling along the working face advance direction and the conversion of two-dimensional curved surfaces into three-dimensional solid models, aiming to construct a complete geological digital model covering the entire mining area. The specific implementation process of this step is as follows: Perform recursive calculations along the strike direction, dividing the entire area to be mined into several continuous geometric segments along the strike, defining the first segment... The starting boundary of the first segment is the first... The termination boundary of each section. Specifically, after completion of the control point... , After fitting the surface of the current segment, with and Subsequent control points in the direction Based on this, select a new set of control points. In the next loop, the process of constructing the reference plane in step S200 and the surface fitting process in step S300 is repeated to generate a series of continuously arranged local surface patches. ,in This represents the total number of segments.

[0034] Implementing smooth connection processing between curved surface patches is necessary because the reference plane normal vectors of adjacent sections may differ due to roadway undulations. Direct splicing would result in angled joints, affecting the smoothness of the coal mining machine's cutting path. This requires addressing issues in adjacent curved surface patches. and The boundary is the parameter and At this point, geometric continuity constraints are applied because the two surface patches share boundary control points. and Positional continuity ( The continuity of the tangent vector is naturally satisfied, but to achieve the continuity of the tangent vector ( (Continuous), in the joint region, a weighted average method is used to smooth the transition of the normal vectors of grid points near the boundary, or a spline surface transition algorithm is used to ensure the generation of the overall coal seam floor surface. Maintain geometric smoothness globally.

[0035] Based on the coal seam thickness, a roof surface is generated, and coal seam thickness data exposed in geological exploration reports or roadways is obtained to establish a coal seam thickness distribution function. If the coal seam thickness varies little, it can be It is considered a constant; if there is thickness variation, interpolation is set based on drilling data, based on the constructed base plate surface. Thickness stretching is performed along the normal vector direction of the surface to calculate the top plate surface. The coordinates of the corresponding points on the top plate, and the equation of the top plate surface, are expressed as: ; In the formula, This is the unit normal vector of the bottom plate surface at the corresponding position. Through this step, the upper boundary (roof) and lower boundary (bottom plate) of the coal mining operation are determined.

[0036] Construct a closed 3D solid model and utilize the generated base plate surface. Top plate curved surface The side grid connecting the edges of the two includes the cutting face, the stop line face, and the coal wall of the two roadways, which enclose a closed volume space. The geometry of this space is discretized into a three-dimensional triangular mesh or voxel model, which serves as the basic digital base of the fully automated mining system. This digital model contains the absolute coordinate information of each spatial node, providing direct data support for adjusting the cutting height of the coal mining machine drum, deciding the support height of the hydraulic support, and planning the moving path of the scraper conveyor.

[0037] Step S500 mainly involves dynamically correcting the digital model of the unmined area ahead using newly revealed measured data during the fully mechanized mining operation. By establishing an error decomposition and closed-loop feedback mechanism, the problem of cumulative deviation in the static model as the advance distance increases is solved. The specific implementation process of this step is as follows: Acquire real-time measured data of the longwall face, and as the longwall face advances, proceed according to the preset advance step distance. For example, every 5 to 10 meters of advance, a machine vision measurement device or 3D laser scanner deployed on the hydraulic support or coal mining machine is activated. The measurement device scans the feature points of the newly exposed coal wall, roof, or floor to obtain a set of measured control points, which is recorded as the first... The three-dimensional coordinates of the measured control points are: ,in , For planar position coordinates, These are measured elevation values, and the measured data reflect the latest geological conditions.

[0038] The elevation error between the model's predicted and measured values ​​is calculated based on the plane coordinates of the measured points. Retrieve or interpolate the corresponding model-predicted elevation from the currently stored digital model. Calculate the elevation deviation at each control point. The formula is: ; This set of deviation values This constitutes the error field sample of the current mining location.

[0039] To accurately distinguish between systematic changes in coal seam occurrence, such as overall dip angle changes, and local random disturbances, such as local undulations or measurement noise, the spatial error field is decomposed into components. Decomposed into trend terms and local fluctuation term Two parts, namely: ; Trend Item Characterizing large-scale geological trend deviations, local fluctuation terms Characterize local detail corrections.

[0040] A trend term model is constructed and solved, and a bivariate polynomial surface fitting method is used to describe the trend term. By setting the polynomial order to 2nd or 3rd order to balance fitting ability and computational stability, the bivariate quadratic polynomial model is represented as: ; In the formula, Given the coefficient vector to be determined, the coefficient vector is solved using the least squares method based on the deviation sample data obtained in the previous steps. The goal is to minimize the sum of squared fitting residuals at all measured points: ; in The number of measured points involved in the calculation is used to solve for the function. This represents the systematic deviation trend between the model and reality within that region.

[0041] Spatial interpolation is performed on the local fluctuation term to calculate the residual sequence after removing the trend term. A local fluctuation term model is constructed using a radial basis function interpolation algorithm. To accurately characterize nonlinear local fluctuations, the interpolation function takes the form: ; In the formula, This indicates the calculation of planar Euclidean distance; The kernel function is a radial basis function, preferably a Gaussian kernel function. , For shape parameters; These are the weighting coefficients, which are obtained by solving a system of linear equations. Obtain, among which It is an interpolation matrix composed of kernel function values.

[0042] The model is weighted and updated, utilizing the trend terms obtained from the decomposition. and fluctuation term The model elevation of the unmined area in front of the working face is corrected. To prevent overfitting caused by data extrapolation, a spatial decay weighting function is introduced. ,in This represents the distance from the point to be corrected to the currently measured coal face. The updated model elevation. The calculation formula is: ; Weight With distance By increasing and decreasing the size of the model, full corrections are made in areas close to the measured points, while only trend corrections or a gradual transition to the original model are retained in areas far from the measured points. This achieves dynamic approximation between the digital model and the real geological environment, forming a closed-loop update system.

[0043] Please see the appendix Figure 7 This application also provides a computer device, including: a processor and a memory, wherein the memory stores a computer program executable by the processor, and the computer program performs the method described above when executed by the processor.

Claims

1. A method for constructing a digital model of an automated mining face, characterized in that, Includes the following steps: The data acquisition steps involve setting target control points along the bottom plate of the return airway and transport roadway of the working face under the unified coordinate system of the entire mine, and setting dip control points at the opening position to obtain the three-dimensional coordinate data of the target control points and the dip control points. The reference plane construction step involves selecting three non-collinear points from the target control points of the transport roadway and the return air roadway as reference points, calculating the plane vector formed by the reference points, determining the reference plane that reflects the strike and dip of the coal seam in the local area, and obtaining the normal vector and plane equation of the reference plane. The initial surface fitting step involves calculating the normal offset of the dip control point relative to the reference plane, constructing an interpolation function that describes the change of the offset with the dip position parameter, and spatially mapping the reference plane along the normal vector direction according to the interpolation function to generate the coal seam floor surface. The model generation steps involve sequentially executing the reference plane construction step and the initial surface fitting step along the working face advancement direction to generate multiple sections of coal seam floor surfaces and smoothly connect them. Combined with coal seam thickness data, solid stretching is performed to establish a fully automated mining working face digital model. The dynamic update step involves collecting the measured coordinates of the newly exposed working face during the mining process, calculating the error between the model's predicted value and the measured value, decomposing the error into a trend term and a local fluctuation term, and correcting the model data for subsequent unmined areas.

2. The method for constructing a digital model of an automated mining face according to claim 1, characterized in that, In the data acquisition step, the setting of the target control points includes: From the opening of the working face towards the stop line, physical markers are laid at the stable rock strata on the bottom of the roadway at preset intervals; The coordinate sequence of the target control points of the two roadways is obtained using measuring equipment, and the target control points of the transport roadway and the return airway are located on the same vertical cross section.

3. The method for constructing a digital model of an automated mining face according to claim 1, characterized in that, The reference plane construction steps include: Two adjacent target control points are selected on one side of the transport roadway, and a target control point corresponding to one of the target control points on the side of the return air roadway is selected on the other side to form three base points. Calculate the first vector along the direction and the second vector along the dip based on the three base points; The normal vector of the reference plane is calculated using the cross product of the first vector and the second vector; Based on the normal vector and the coordinates of any base point, establish the point normal equation of the reference plane.

4. The method for constructing a digital model of an automated mining face according to claim 1, characterized in that, The initial surface fitting step includes: Define the line connecting the base point on the side of the transport roadway and the base point on the side of the return air roadway as the parameterized baseline, and determine the normalized position parameters of each dip control point on the parameterized baseline. Substitute the coordinates of each dip control point into the plane equation of the reference plane, and calculate the algebraic distance from each dip control point to the reference plane as the normal offset. Using Lagrange interpolation or spline interpolation, an offset function that tends towards undulating shapes is constructed based on the normalized position parameters and the normal offset.

5. The method for constructing a digital model of an automated mining face according to claim 4, characterized in that, In the initial surface fitting step, the specific method for generating the coal seam floor surface is as follows: The reference plane is discretized into a grid to obtain the basic coordinates of the grid points on the reference plane; For each grid point, the corresponding correction value is calculated using the offset function; The correction value is multiplied by the normal vector of the reference plane and superimposed onto the basic coordinates to obtain the point coordinates on the curved surface of the coal seam floor.

6. The method for constructing a digital model of an automated mining face according to claim 1, characterized in that, The model generation steps include: The coal seam floor surface generated in adjacent sections is geometrically continuous at the joint, and the normal vector of the grid points near the boundary is smoothed by weighted averaging. Obtain coal seam thickness distribution data, and based on the smoothed coal seam floor surface, superimpose the coal seam thickness value along the normal vector direction of the surface to generate the roof surface. A closed three-dimensional solid model is formed by using the curved surface of the coal seam floor, the curved surface of the roof, and the side mesh.

7. The method for constructing a digital model of an automated mining face according to claim 1, characterized in that, In the dynamic update step, the specific method for calculating the error is as follows: Acquire local scanning data of newly exposed feature points in the longwall face, and combine it with the positioning data of the current coal mining equipment or support equipment to convert the local scanning data into measured coordinates in a unified coordinate system for the entire mine; Based on the measured coordinates of the planar position, the corresponding model elevation is retrieved in the digital model of the fully automated mining face; The spatial error field is obtained by calculating the difference between the measured elevation value and the model elevation.

8. The method for constructing a digital model of an automated mining face according to claim 1, characterized in that, In the dynamic update step, the specific methods for decomposing the error into a trend term and a local fluctuation term include: The spatial error field is fitted using a polynomial surface fitting method to obtain a trend term function characterizing the systematic deviation; Calculate the residual between the spatial error field and the trend term function; The residuals are spatially interpolated using a radial basis function interpolation algorithm to establish a local fluctuation term model that characterizes local detail corrections.

9. The method for constructing a digital model of an automated mining face according to claim 8, characterized in that, The correction of the model data for subsequently unexploited areas includes: A spatial attenuation weighting function is constructed, the value of which decreases as the distance between the point to be corrected and the current measured coal wall increases; The sum of the trend term value obtained by substituting the coordinates of the position to be corrected into the trend term function and the local fluctuation term value obtained by substituting them into the local fluctuation term model is multiplied by the corresponding spatial decay weight function value to obtain the correction amount. The correction amount is then superimposed on the predicted elevation of the original model to complete the closed-loop update of the model.

10. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1-9.

Images