A method for intelligent site selection and line orientation search of long and large mountain railway tunnels under terrain driving

Abstract

The present application relates to railway route selection technical field, disclose a kind of topographic driving mountainous area railway long and large tunnel intelligent site selection and line direction search method.The search method integrates topographic data preprocessing, topographic classification and portal search and line direction search three major blocks, specifically: three-dimensional topographic data is converted into spatial wave, the characteristics of topographic relief are obtained by Fourier transform to spatial wave processing;Based on representative tunnel length design low-pass filter filtering preprocessing;After the topographic data processed, the feasible portal position is searched after topographic region is classified by block;Potential tunnel portal position is grouped;With engineering cost as main consideration factor, connect tunnel portal using bidirectional A algorithm, generate several feasible schemes, obtain optimal path by comparison and selection;The optimal path generates specific three-dimensional railway line scheme.The search method is high in search efficiency and suitable for complex mountainous area railway line design.

Description

Technical Field

[0001] This invention relates to the field of railway route selection technology, and in particular to a method for intelligent site selection and route search of long railway tunnels in mountainous areas driven by terrain. Background Technology

[0002] Railway line design is an important part of railway engineering planning and design. Its core objective is to determine the optimal route and generate specific spatial alignment by considering multiple objectives and constraints.

[0003] The site selection of railway tunnels is one of the most complex and challenging aspects of railway construction, directly impacting the feasibility, safety, economy, and operational efficiency of the project. As a crucial structure for railways traversing complex terrain, tunnel site selection requires consideration of constraints such as geological issues, environmental factors, and elevation differences. Since potential portal locations meeting these conditions may be scattered throughout the entire terrain area, it is essential to select the optimal solution best suited to the overall alignment. However, traditional manual tunnel site selection and alignment design methods are time-consuming and labor-intensive, and easily miss potential feasible site options. Therefore, computer-aided railway line optimization design has gradually become a hot topic.

[0004] Existing computer-aided railway line design methods typically generate tunnel portals and alignments through exhaustive search based on a digital elevation model of the research area. This search method is inefficient when the pre-selected design area is large or the terrain is complex. Therefore, current railway line design methods for complex mountainous areas still have shortcomings, and it is necessary to develop a more efficient and accurate tunnel site selection and intelligent alignment optimization method specifically for railways in complex mountainous areas. Summary of the Invention

[0005] The main objective of this invention is to provide a terrain-driven intelligent site selection and route alignment search method for long railway tunnels in mountainous areas, aiming to solve the problems of low efficiency and unsuitability for complex mountain railway line design in existing search methods. The specific technical solution is as follows:

[0006] A terrain-driven intelligent site selection and route alignment search method for long railway tunnels in mountainous areas includes the following steps:

[0007] Step S1: Establish a high-resolution digital elevation model (DEM) containing topographic data for the selected study area;

[0008] Step S2: Perform filtering preprocessing on the terrain data obtained in step S1. Specifically, after abstracting the terrain data into a three-dimensional spatial wave, fit and expand the spatial wave into a two-dimensional Fourier series to obtain the characteristics of terrain undulation; design a low-pass filter based on the representative tunnel length to retain important low-frequency terrain data.

[0009] Step S3: Post-process the terrain data after filtering and preprocessing in Step S2. Specifically, classify the terrain of the study area by the natural slope between the start and end points; identify valley units using the terrain identification method based on Geographic Information System (GIS) as potential tunnel portal locations and group them according to terrain type; and filter the potential portal locations found in the preliminary search based on constraints.

[0010] Step S4: Connect the potential tunnel portal locations obtained from the search to the railway line alignment search module; construct a weighted directed graph with tunnel portals as nodes, tunnel connections as edges, and edge costs as weights, and then modify the bidirectional... Algorithm, searching for the optimal path;

[0011] Step S5: Optimize the optimal path obtained in step S4 based on the particle swarm optimization algorithm, and output the railway line scheme.

[0012] Preferably, step S2 specifically includes the following steps:

[0013] Step S2-1: Treat the terrain data obtained in S1 as a set of data from... × Each with uniform grid spacing and Discrete dataset consisting of arranged measurements This discrete dataset is abstracted as a three-dimensional spatial wave, where the valley spacing is the wavelength and the reciprocal of the valley spacing is the frequency.

[0014] Step S2-2: Generate a spectrum containing amplitude, frequency, and phase information through discrete Fourier transform;

[0015] Step S2-3: Design a low-pass filter to filter the terrain spectrum data. Specifically, based on Fourier transform spectral feature analysis, select a representative tunnel length. The terrain wavelength threshold is used as the low-pass frequency. The frequency corresponding to the terrain wavelength threshold is used as the low-pass frequency. A low-pass filter is designed based on the low-pass frequency, and the terrain spectrum data is filtered.

[0016] Step S2-4: Obtain the filtered and preprocessed terrain elevation data by using inverse Fourier transform to obtain the filtered spectrum.

[0017] Preferably, a low-pass filter It can be expressed as follows:

[0018] ;

[0019] in: The cutoff frequency, = ; The spectrum in the spectrum diagram;

[0020] The Fourier transform of a superimposed low-pass filter is expressed as:

[0021] ;

[0022] The inverse Fourier transform of a superimposed low-pass filter is expressed as:

[0023] ;

[0024] in: and The wave number in the spectrum; and It is a discrete dataset Index in , ; Yes The complex numerical matrix obtained by Fourier transform is the digital representation of topographic elevation data in the frequency domain. It is the terrain elevation data after filtering and preprocessing; It is a natural constant; It is an imaginary number.

[0025] Preferably, step S3 specifically includes the following steps:

[0026] Step S3-1: Classify the terrain of the study area according to the natural slope between the starting and ending points of the railway line; identify valley units using GIS-based terrain identification methods and use them as potential tunnel portal areas; group the potential tunnel portals according to terrain type.

[0027] Step S3-2: Perform feasibility screening on the potential portal locations found through the search. Specifically, consider the constraints in the horizontal and longitudinal directions to screen the potential portal locations found through the preliminary search.

[0028] Preferably, in step S3-2:

[0029] Pick Indicates the first The coordinates of the tunnel portal. The coordinates of the starting point of the railway line; Let the coordinates be the coordinates of the end point of the railway line, then:

[0030] In the planar direction, the position of the portal is constrained and filtered by limiting the development coefficient, and its expression is as follows:

[0031] ;

[0032] in: The coefficient of performance; The air distance from the starting point to the destination;

[0033] In the longitudinal direction, constraints are selected based on the natural slope, and the expression is as follows:

[0034] ;

[0035] in: Indicates the coordinates of the first set of tunnel portals; Indicates the coordinates of the second tunnel portal that is adjacent to the first tunnel portal; This represents the maximum permissible slope.

[0036] Preferably, step S4 includes the following steps:

[0037] Step S4-1: Construct a directed graph with each group of tunnel portals as nodes and each tunnel connection as an edge. Each edge is divided into two categories: tunnel segments and open line segments, and different unit prices are assigned according to the category.

[0038] Step S4-2: Perform forward and reverse searches based on the starting and ending points of the railway line to obtain the forward optimal path Pathforward, the forward optimal path total cost Costforward, the reverse optimal path Pathbackward, and the reverse optimal path total cost Costbackward, respectively.

[0039] Step S4-3: Compare the total cost of the forward optimal path Costforward and the total cost of the reverse optimal path Costbackward obtained in step S4-2. If the total cost of the forward optimal path Costforward ≤ the total cost of the reverse optimal path Costbackward, then output the forward optimal path Pathforward as the optimal path; otherwise, output the reverse optimal path Pathbackward as the optimal path.

[0040] Preferably, in step S4-1: the construction cost of each edge, i.e., the product of the distance between tunnel entrances and the unit price, is used as the weight of each edge, specifically expressed as follows:

[0041] For the tunnel section: ;

[0042] For open line segments: ;

[0043] in: Let be the weight of the edge; The distance between tunnel entrances; This represents the average unit price of tunnel section construction costs; The average unit price for the construction cost of the Ming Line section.

[0044] Preferably, in the forward search: the starting point of the railway line is taken as the starting point of the algorithm, and the ending point of the railway line is taken as the ending point of the algorithm;

[0045] In reverse search: the starting point of the railway line is used as the end point of the algorithm, and the end point of the railway line is used as the starting point of the algorithm;

[0046] The search process includes the following steps:

[0047] Step 1: Connect the start and end points of the railway line and draw vertical lines at fixed intervals to divide the entire search area into N vertical grids. Each grid represents a stage of the railway line's progress, i.e., a group. Add a key constraint: During the path's progress, each group can only visit at most one node.

[0048] Step 2: Construct the open table to store nodes to be evaluated, sorted by f(n); construct the close table to store nodes that have been evaluated and whose optimal parent node has been determined; construct the path array, which is a dynamic array that records the known optimal path from the starting point to a certain node in the current stage in each iteration.

[0049] f(n) is calculated using the following formula:

[0050] f(n) = g(n) + h(n);

[0051] Where: g(n) is the total cost from the starting point through the nodes stored in the path array to node n; h(n) is... The heuristic term of the algorithm represents the estimated minimum cost from node n to the destination; f(n) is the total cost of the global path passing through node n;

[0052] Step 3: Add the starting point of the algorithm to the open list, calculate f(Q) of the starting point, and set k=0 for the current stage;

[0053] Step 4: Take the node with the smallest f(n) value from the open list and call it the parent node P; move the parent node P from the open list to the close list;

[0054] Step 5: Based on the grid group to which the parent node P belongs, determine the next grid group to be visited, k+1; treat all tunnel portal nodes in grid group k+1 as candidate child nodes Ci; calculate the temporary actual cost for each candidate child node Ci. ,in: The weights connecting parent node P and candidate child node Ci;

[0055] Step 6: Check the open and close tables;

[0056] Step ⑦: Traverse all the nodes in the open list that belong to stage k + 1. For each such node, it represents a complete path from the starting point to the current node. Determine the node B with the smallest f(n) in the current open list, compare its path cost with the path cost recorded in the path array for reaching stage k + 1, update the path array, and use node B as the end point of the current optimal path.

[0057] Step ⑧: Clear the open list, and use the end point of the current optimal path recorded in the path array as the new parent node P, and add it to the open list.

[0058] Step ⑨: Make a judgment. Specifically: If the new parent node P is not the end point of the algorithm, then set stage k = k + 1 and return to Step ④; if the new parent node P is the end point of the algorithm, then the search ends, and the optimal path and its total cost are obtained.

[0059] Preferably, in Step ⑥, checking the open list and the close list includes the following three cases:

[0060] Case 1: If the candidate child node Ci is not in the open list and the close list, then calculate h(Ci), calculate f(Ci) = g’ + h(Ci), set the parent node of the candidate child node Ci as P, and add Ci to the open list.

[0061] Case 2: If the candidate child node Ci is in the open list, then compare the new g’ with the g(Ci) recorded in the open list. If g’ < g(Ci), update g(Ci) = g’, update f(Ci) = g’ + h(Ci), update the parent node of the candidate child node Ci as P, and re - sort the open list.

[0062] Case 3: If the candidate child node Ci is in the close list, then the path is fixed and no adjustment is needed.

[0063] Preferably, in Step ⑦: If there is no record in the path array, or the cost of node B is lower, then update the path array with node B and its path information. At this moment, the path array records the current optimal path for reaching stage k + 1.

[0064] Applying the technical solution of the present invention, the effects are:

[0065] The terrain-driven intelligent site selection and route search method for long railway tunnels in mountainous areas disclosed in this invention integrates three major components: terrain data preprocessing, terrain classification and tunnel portal search, and route search. Specifically, it first converts three-dimensional terrain data into spatial waves, then processes the spatial waves through a two-dimensional Fourier transform, fitting and expanding the spatial waves into a two-dimensional Fourier series to obtain the frequency, period, amplitude, and phase characteristics of terrain undulations. Subsequently, a low-pass filter is designed based on the representative tunnel length to filter out and exclude small valleys and hills that are unsuitable as tunnel portal locations, retaining only important low-frequency terrain data, such as high mountains and deep valleys. Next, it integrates tunnel site selection optimization methods from complex mountain railway line design, using the pre-processed terrain data to classify the terrain region into blocks and search for feasible tunnel portal locations. Simultaneously, potential tunnel portal locations are grouped with the line connecting the origin and destination as the axis and the representative tunnel length as the interval, and engineering cost is taken as the main consideration, utilizing bidirectional... The algorithm connects tunnel portals, generating several feasible solutions. By comparing these solutions, the optimal path is obtained. Finally, the optimal path is used to generate a specific 3D railway line plan. This search method is highly efficient and suitable for designing railway lines in complex mountainous areas. Attached Figure Description

[0066] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0067] Figure 1 This is a flowchart of the terrain-driven intelligent site selection and route search method for long railway tunnels in mountainous areas, as described in this embodiment of the invention.

[0068] Figure 2 The images show grayscale terrain images before and after filtering, where: (a) is the original terrain image; and (b) is the terrain image after low-pass filtering.

[0069] Figure 3 This is a schematic diagram of the global route search process, where: (a) is the location of potential portals; (b) is a schematic diagram of feasible route connections; and (c) is a schematic diagram of the optimal path.

[0070] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0071] 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 a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0072] Example:

[0073] A terrain-driven intelligent site selection and route alignment search method for long railway tunnels in mountainous areas, see details. Figure 1 This includes the following steps:

[0074] Step S1: Establish a study area including topographic data for the selected route (see details). Figure 2 The high-resolution digital elevation model (DEM) shown in (a) is illustrated. Those skilled in the art can establish a high-resolution digital elevation model (DEM) by referring to existing technologies.

[0075] Step S2: Perform filtering preprocessing on the obtained terrain data. Specifically, after abstracting the terrain data into a three-dimensional spatial wave, fit and expand the spatial wave into a two-dimensional Fourier series to obtain the frequency, period, amplitude and phase characteristics of the terrain undulations. Design a low-pass filter based on the representative tunnel length to retain important low-frequency terrain data, such as high mountains and deep valleys.

[0076] In this preferred embodiment, step S2 specifically includes the following steps:

[0077] Step S2-1: Treat the terrain data obtained in S1 as a set of data from... × Each with uniform grid spacing and Discrete dataset consisting of arranged measurements This discrete dataset is abstracted as a three-dimensional spatial wave, where the valley spacing is the wavelength and the reciprocal of the valley spacing is the frequency.

[0078] Step S2-2: Generate a spectrum containing amplitude, frequency, and phase information through Discrete Fourier Transform. The processing method for Discrete Fourier Transform can be implemented with reference to existing technologies.

[0079] Step S2-3: Design a low-pass filter to filter the terrain spectrum data. Specifically, based on Fourier transform spectral feature analysis, select a representative tunnel length. The terrain wavelength threshold is used as the threshold frequency; the frequency corresponding to this threshold frequency is the low-pass frequency. A low-pass filter is designed based on the low-pass frequency, and the terrain spectrum data is filtered.

[0080] In this embodiment, since the fully expanded high-order Fourier series is very complex, it is not suitable for application in engineering analysis. The topographic data processed by Fourier transform exhibits the following characteristics in its topographic wave spectrum: small elevation differences are usually densely distributed throughout the area, with high frequencies; large elevation differences often have dramatically undulating landforms (such as high mountains and deep valleys), with large spacing and low frequencies. Therefore, based on the spectral characteristic analysis of Fourier transform, a representative tunnel length ( L 0 The low-pass frequency is defined as the wavelength threshold for the terrain. The frequency corresponding to this threshold is the low-pass frequency; frequencies exceeding this threshold are defined as high-frequency regions, i.e., areas with relatively small elevation differences. Based on the low-pass frequency, a low-pass filter is designed to exclude small valleys and hills, retaining only critical terrain data sensitive to tunnel site selection, such as high mountains and deep valleys. By superimposing the low-pass filter onto the aforementioned Fourier transform, the spectrum is processed. After filtering, the original Fourier expansion is truncated at the low-pass frequency, and its higher-order expansion terms are removed.

[0081] Step S2-4: Obtain the filtered preprocessed terrain elevation data by performing an inverse Fourier transform on the filtered spectrum. After the inverse Fourier transform, excessively minor terrain variations reflected in the terrain elevation data have been filtered out, resulting in a smoother overall terrain change (e.g., ...). Figure 2 (as shown in (b)).

[0082] In this embodiment, the low-pass filter It can be expressed as follows:

[0083] ;

[0084] in: The cutoff frequency, = ; The spectrum in the spectrum diagram;

[0085] The Fourier transform of a superimposed low-pass filter is expressed as:

[0086] ;

[0087] The inverse Fourier transform of a superimposed low-pass filter is expressed as:

[0088] ;

[0089] in: and The wave number in the spectrum; and It is a discrete dataset Index in , ; Yes The complex numerical matrix obtained by Fourier transform is the digital representation of topographic elevation data in the frequency domain. It is the terrain elevation data after filtering and preprocessing; It is a natural constant; It is the imaginary unit, and its mathematical definition is .

[0090] Step S3: Post-process the terrain data after filtering and preprocessing in step S2. Specifically, classify the terrain of the study area by the natural slope between the start and end points; identify valley units using a terrain identification method based on Geographic Information System (GIS) as potential tunnel portal locations and group them according to terrain type; and filter the potential portal locations found in the preliminary search based on constraints.

[0091] In this embodiment, step S3 specifically includes the following steps:

[0092] Step S3-1: Classify the terrain of the study area according to the natural slope between the starting and ending points of the railway line; identify valley units using GIS-based terrain identification methods and use them as potential tunnel portal areas; group the potential tunnel portals according to terrain type.

[0093] Step S3-2: Perform feasibility screening on the potential portal locations found through the search. Specifically, consider the constraints in the horizontal and longitudinal directions to screen the potential portal locations found through the preliminary search.

[0094] In this preferred embodiment, the following is taken: Indicates the first The coordinates of the tunnel portal. The coordinates of the starting point of the railway line; Let the coordinates be the coordinates of the end point of the railway line, then:

[0095] In the planar direction, the position of the portal is constrained and filtered by limiting the development coefficient, and its expression is as follows:

[0096] ;

[0097] in: The coefficient of performance; The air distance from the starting point to the destination;

[0098] In the longitudinal direction, constraints are selected based on the natural slope, and the expression is as follows:

[0099] ;

[0100] in: Indicates the coordinates of the first set of tunnel portals; Indicates the coordinates of the second tunnel portal that is adjacent to the first tunnel portal; This represents the maximum permissible slope.

[0101] Step S4: Connect the potential tunnel portal locations obtained from the search to the railway line alignment search module; construct a weighted directed graph with tunnel portals as nodes, tunnel connections as edges, and edge costs as weights, and then modify the bidirectional... Algorithm, searching for the optimal path.

[0102] In this preferred embodiment, step S4 includes the following steps:

[0103] Step S4-1: Construct a directed graph with each group of tunnel portals as nodes and the tunnel connections as edges. Each edge is divided into two categories: tunnel segments and open-line segments, and different unit prices are assigned to each category (e.g., ...). Figure 3 (as shown in (b)).

[0104] In this preferred embodiment, the construction cost of each edge, i.e., the product of the distance between tunnel entrances and the unit price, is used as the weight of each edge, specifically expressed as follows:

[0105] For the tunnel section: ;

[0106] For open line segments: ;

[0107] in: Let be the weight of the edge; The distance between tunnel entrances; This represents the average unit price of tunnel section construction costs; The average unit price for the construction cost of the Ming Line section.

[0108] Step S4-2: Perform forward and reverse searches based on the starting and ending points of the railway line to obtain the forward optimal path Pathforward, the forward optimal path total cost Costforward, the reverse optimal path Pathbackward, and the reverse optimal path total cost Costbackward, respectively.

[0109] In this preferred forward search embodiment, the starting point of the railway line is used as the starting point of the algorithm, and the ending point of the railway line is used as the ending point of the algorithm. See details... Figure 3 As shown in (a)-(b); in the reverse search: the starting point of the railway line is taken as the end point of the algorithm, and the end point of the railway line is taken as the starting point of the algorithm.

[0110] The further optimization search process includes the following steps:

[0111] Step 1: Connect the start and end points of the railway line and draw vertical lines at fixed intervals to divide the entire search area into N vertical grids. Each grid represents a stage of the railway line's progress, i.e., a group. Add a key constraint: During the path's progress, each group can only visit at most one node.

[0112] Step 2: Construct the open table to store nodes to be evaluated, sorted by f(n); construct the close table to store nodes that have been evaluated and whose optimal parent node has been determined; construct the path array, which is a dynamic array that records the known optimal path from the starting point to a certain node in the current stage in each iteration.

[0113] f(n) is calculated using the following formula:

[0114] f(n) = g(n) + h(n);

[0115] Where: g(n) is the total cost from the starting point through the nodes stored in the path array to node n; h(n) is... The heuristic term of the algorithm represents the estimated minimum cost from node n to the destination; f(n) is the total cost of the global path passing through node n;

[0116] Step 3: Add the starting point of the algorithm to the open list, calculate f(Q) of the starting point, and set k=0 for the current stage;

[0117] Step 4: Take the node with the smallest f(n) value from the open list and call it the parent node P; move the parent node P from the open list to the close list;

[0118] Step 5: Based on the grid group to which the parent node P belongs, determine the next grid group to be visited, k+1; treat all tunnel portal nodes in grid group k+1 as candidate child nodes Ci; calculate the temporary actual cost for each candidate child node Ci. ,in: The weights connecting parent node P and candidate child node Ci;

[0119] Step 6: Check the open and close tables;

[0120] Step 7: Traverse all nodes in the open table that belong to stage k+1. For each such node, it represents a complete path from the starting point to the current node. Determine the node B with the smallest f(n) in the current open table, and compare its path cost with the path cost to stage k+1 recorded in the path array. Update the path array and set node B as the endpoint of the current optimal path.

[0121] Step ⑧: Clear the open list, and take the end point of the currently optimal path recorded in the path array as the new parent node P, and add it to the open list;

[0122] Step ⑨: Make a judgment. Specifically, if the new parent node P is not the end point of the algorithm, then set the stage k = k + 1, and return to Step ④; if the new parent node P is the end point of the algorithm, then the search ends, and the optimal path and its total cost are obtained.

[0123] Step S4-3: Compare the total cost of the forward optimal path Costforward and the total cost of the reverse optimal path Costbackward obtained in Step S4-2. If the total cost of the forward optimal path Costforward ≤ the total cost of the reverse optimal path Costbackward, then output the forward optimal path Pathforward as the optimal path; otherwise, output the reverse optimal path Pathbackward as the optimal path.

[0124] In a further preferred embodiment of the present invention, the inspection of the open list and the close list in Step ⑥ includes the following three cases:

[0125] Case 1: If the candidate child node Ci is not in the open list and the close list, then calculate h(Ci), calculate f(Ci) = g' + h(Ci), set the parent node of the candidate child node Ci as P, and add the candidate child node Ci to the open list;

[0126] Case 2: If the candidate child node Ci is in the open list, then compare the new g' and the g(Ci) recorded in the open list. If g' < g(Ci), update g(Ci) = g', update f(Ci) = g' + h(Ci), update the parent node of the candidate child node Ci as P, and re-sort the open list;

[0127] Case 3: If the candidate child node Ci is in the close list, then the path is fixed and no adjustment is required.

[0128] In a further preferred embodiment of the present invention, in Step ⑦: If there is no record in the path array, or the cost of node B is lower, then update the path array with node B and its path information. At this moment, the path array records the currently optimal path to stage k + 1.

[0129] Step S5: Optimize the optimal path obtained in Step S4 based on the particle swarm optimization algorithm (as shown in (c) in Figure 3 ), and output the railway line plan. The method that can be adopted here is: construct a railway line optimization model considering the line engineering cost, and optimize the above railway path direction based on the particle swarm optimization algorithm. For details, reference can be made to the prior art.

[0130] The effect of applying the technical solution of this invention is as follows:

[0131] This invention proposes a high-efficiency railway alignment method suitable for complex mountainous areas by integrating three parts: terrain data preprocessing, terrain classification and tunnel portal search, and route alignment search. Specifically: First, the three-dimensional terrain data is converted into spatial waves, and then processed using a two-dimensional Fourier transform to fit and expand the spatial waves into a two-dimensional Fourier series, obtaining the frequency, period, amplitude, and phase characteristics of the terrain undulations. Next, a low-pass filter is designed based on the representative tunnel length to filter out and exclude small valleys and hills unsuitable as tunnel portal locations, retaining only important low-frequency terrain data, such as mountains and deep valleys. Then, by integrating tunnel site selection optimization methods from complex mountainous railway line design, the terrain regions are divided into blocks using the pre-processed terrain data, and feasible tunnel portal locations are searched. Simultaneously, potential tunnel portal locations are grouped using the line connecting the origin and destination as the axis and the representative tunnel length as the interval, with engineering cost as the main consideration, and using bidirectional... The algorithm connects the tunnel portals, generates several feasible solutions, compares and selects the optimal path, and finally inputs the optimal path into the line optimization module to generate a specific three-dimensional railway line plan.

[0132] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural transformations made using the contents of the present invention's specification and drawings under the inventive concept of the present invention, or direct / indirect applications in other related technical fields, are included within the patent protection scope of the present invention.

Claims

1. A terrain-driven intelligent site selection and route alignment search method for long railway tunnels in mountainous areas, characterized in that, Includes the following steps: Step S1: Establish a high-resolution digital elevation model (DEM) containing topographic data for the selected study area; Step S2: Perform filtering preprocessing on the terrain data obtained in step S1. Specifically, after abstracting the terrain data into a three-dimensional spatial wave, fit and expand the spatial wave into a two-dimensional Fourier series to obtain the characteristics of terrain undulation; design a low-pass filter based on the representative tunnel length to retain low-frequency terrain data. Step S3: Post-process the terrain data after filtering and preprocessing in Step S2. Specifically, classify the terrain of the study area by the natural slope between the start and end points; identify valley units using the terrain identification method based on Geographic Information System (GIS) as potential tunnel portal locations and group them according to terrain type; and filter the potential portal locations found in the preliminary search based on constraints. Step S4: Connect the potential tunnel portal locations obtained from the search to the railway line alignment search module; construct a weighted directed graph with tunnel portals as nodes, tunnel connections as edges, and edge costs as weights, and then modify the bidirectional... Algorithm, searching for the optimal path; Step S5: Optimize the optimal path obtained in step S4 based on the particle swarm optimization algorithm, and output the railway line scheme; Step S2 specifically includes the following steps: Step S2-1: Treat the terrain data obtained in S1 as a set of data from... × Each with uniform grid spacing and Discrete dataset consisting of arranged measurements This discrete dataset is abstracted as a three-dimensional spatial wave, where the valley spacing is the wavelength and the reciprocal of the valley spacing is the frequency. Step S2-2: Generate a spectrum containing amplitude, frequency, and phase information through discrete Fourier transform; Step S2-3: Design a low-pass filter to filter the terrain spectrum data. Specifically, based on Fourier transform spectral feature analysis, select a representative tunnel length. The terrain wavelength threshold is used as the low-pass frequency. The frequency corresponding to the terrain wavelength threshold is used as the low-pass frequency. A low-pass filter is designed based on the low-pass frequency, and the terrain spectrum data is filtered. Step S2-4: Obtain the filtered and preprocessed terrain elevation data by performing inverse Fourier transform on the filtered spectrum; Step S4 includes the following steps: Step S4-1: Construct a directed graph with each group of tunnel portals as nodes and each tunnel connection as an edge. Each edge is divided into two categories: tunnel segments and open line segments, and different unit prices are assigned according to the category. Step S4-2: Perform forward and reverse searches based on the starting and ending points of the railway line to obtain the forward optimal path Pathforward, the forward optimal path total cost Costforward, the reverse optimal path Pathbackward, and the reverse optimal path total cost Costbackward, respectively. Step S4-3: Compare the total cost of the forward optimal path Costforward and the total cost of the reverse optimal path Costbackward obtained in step S4-2. If the total cost of the forward optimal path Costforward ≤ the total cost of the reverse optimal path Costbackward, then output the forward optimal path Pathforward as the optimal path; otherwise, output the reverse optimal path Pathbackward as the optimal path.

2. The intelligent site selection and route search method for long railway tunnels in mountainous areas under terrain-driven conditions as described in claim 1, characterized in that, low-pass filter Expressed as follows: ； in: The cutoff frequency, = ; The spectrum in the spectrum diagram; The Fourier transform of a superimposed low-pass filter is expressed as: ； The inverse Fourier transform of a superimposed low-pass filter is expressed as: ； in: and The wave number in the spectrum; and It is a discrete dataset Index in , ; Yes The complex numerical matrix obtained by Fourier transform is the digital representation of topographic elevation data in the frequency domain. It is the terrain elevation data after filtering and preprocessing; is a natural constant; i is the imaginary unit.

3. The terrain-driven intelligent site selection and route search method for long railway tunnels in mountainous areas as described in claim 2, characterized in that, Step S3 specifically includes the following steps: Step S3-1: Classify the terrain of the study area according to the natural slope between the starting and ending points of the railway line; identify valley units using GIS-based terrain identification methods and use them as potential tunnel portal areas; group the potential tunnel portals according to terrain type. Step S3-2: Perform feasibility screening on the potential portal locations found through the search. Specifically, consider the constraints in the horizontal and longitudinal directions to screen the potential portal locations found through the preliminary search.

4. The intelligent site selection and route search method for long railway tunnels in mountainous areas under terrain-driven conditions as described in claim 3, characterized in that, In step S3-2: Pick Indicates the first The coordinates of the tunnel portal. The coordinates of the starting point of the railway line; Let the coordinates be the coordinates of the end point of the railway line, then: In the planar direction, the position of the portal is constrained and filtered by limiting the development coefficient, and its expression is as follows: ； in: The coefficient of performance; The air distance from the starting point to the destination; In the longitudinal section direction, constraint screening is carried out through the natural slope, and its expression is as follows: ； in: Indicates the coordinates of the first set of tunnel portals; Indicates the coordinates of the second tunnel portal that is adjacent to the first tunnel portal; This represents the maximum permissible slope.

5. The terrain-driven intelligent site selection and route search method for long railway tunnels in mountainous areas as described in any one of claims 1-4, characterized in that, In step S4-1: The construction cost of each edge, that is, the product of the distance between the tunnel entrances and the unit price, is used as the weight of each edge, and it is specifically expressed as: For the tunnel section: ; For open line segments: ; in: Let be the weight of the edge; The distance between tunnel entrances; This represents the average unit price of tunnel section construction costs; The average unit price for the construction cost of the Ming Line section.

6. The intelligent site selection and route search method for long railway tunnels in mountainous areas under terrain-driven conditions as described in claim 5, characterized in that, In the forward search: The starting point of the railway line is used as the starting point of the algorithm, and the end point of the railway line is used as the end point of the algorithm; In the reverse search: The starting point of the railway line is used as the end point of the algorithm, and the end point of the railway line is used as the starting point of the algorithm; The search process includes the following steps: Step ①: Along the connection line between the starting and ending points of the railway line, vertical lines are made at fixed intervals to divide the entire search area into N vertical grids. Each grid represents a stage of the railway line's advancement, that is, a group; key constraints are added. Specifically: During the advancement of the path, at most only one node can be visited within each group; Step ②: Construct an open table for storing nodes to be evaluated, sorted according to f(n); construct a close table for storing nodes that have been evaluated and whose optimal parent nodes have been determined; construct a path array, which is a dynamic array. In each iteration step, it records the known optimal path from the starting point to a certain node at the current stage; f(n) is calculated using the following formula: f(n)=g(n)+h(n); Where: g(n) is the total cost from the starting point through the nodes stored in the path array to node n; h(n) is... The heuristic term of the algorithm represents the estimated minimum cost from node n to the destination; f(n) is the total cost of the global path passing through node n; Step ③: Add the starting point of the algorithm to the open table, calculate f(Q) of the starting point, and set the current stage k = 0; Step ④: Take out the node with the smallest f(n) value from the open table, and call it the parent node P; move the parent node P from the open table to the close table; Step 5: Based on the grid group to which the parent node P belongs, determine the next grid group to be visited, k+1; treat all tunnel portal nodes in grid group k+1 as candidate child nodes Ci; calculate the temporary actual cost for each candidate child node Ci. ,in: The weights connecting parent node P and candidate child node Ci; Step ⑥: Check the open table and the close table; Step ⑦: Traverse all nodes belonging to stage k + 1 in the open table. For each such node, it represents a complete path from the starting point to the current node; determine the node B with the smallest f(n) in the current open table, and compare its path cost with the path cost recorded in the path array for reaching stage k + 1, update the path array and use the node B as the end point of the current optimal path; Step ⑧: Clear the open table, and use the end point of the current optimal path recorded in the path array as the new parent node P and add it to the open table; Step ⑨: Make a judgment. Specifically: If the new parent node P is not the end point of the algorithm, then take k = k + 1 and return to step ④; if the new parent node P is the end point of the algorithm, the search ends, and the optimal path and its total cost are obtained.

7. The intelligent site selection and route search method for long railway tunnels in mountainous areas under terrain-driven conditions as described in claim 6, characterized in that, The inspection of the open table and the close table in step ⑥ includes the following three situations: Situation 1: If the candidate child node Ci is not in the open and close tables, then calculate h(Ci), calculate f(Ci)=g’+h(Ci), set the parent node of the candidate child node Ci as P, and add the candidate child node Ci to the open table; Situation 2: If the candidate child node Ci is in the open table, then compare the new g’ and the g(Ci) recorded in the open table. If g’<g(Ci), update g(Ci)=g’, update f(Ci)=g’+h(Ci), update the parent node of the candidate child node Ci as P, and re-sort the open table; Scenario 3: If the candidate child node Ci is in the close table, then the path is fixed and no adjustment is needed.

8. The intelligent site selection and route search method for long railway tunnels in mountainous areas under terrain-driven conditions as described in claim 6, characterized in that, In step ⑦: If there is no record in the path array, or the cost of node B is lower, then update the path array with node B and its path information. At this moment, the path array records the current optimal path to reach stage k+1.

Images