Method and system for predicting displacement distribution of existing line and ground surface in undercrossing tunnel construction
By establishing a mapping relationship model and reconstructing the distribution of track and surface displacement caused by tunnel construction using RBF neural networks, the problem of the difficulty in fully reflecting the overall deformation in existing tunnel monitoring technologies has been solved, and high-precision global displacement prediction and safety assessment have been achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGZHOU METRO DESIGN & RES INST CO LTD
- Filing Date
- 2026-02-05
- Publication Date
- 2026-05-15
AI Technical Summary
Existing tunnel underpass construction monitoring technologies are insufficient to fully reflect the overall deformation distribution of existing lines and the ground surface, and their fitting accuracy is limited by empirical parameters, making them unable to adapt to the nonlinear response characteristics under complex geological and construction conditions.
A mapping model between existing railway lines and surface displacement distribution and different coordinates is established. The displacement distribution of the entire railway line and surface path is reconstructed using an RBF neural network. The input-output mapping relationship is established through the coordinate information and geometric feature parameters of monitoring points to achieve displacement prediction in unmonitored areas.
By using a small number of monitoring points, the deformation of the entire line and the ground surface can be reconstructed, reducing the number of monitoring devices, improving spatial resolution, intuitively reflecting the characteristics of uneven settlement, and providing a reliable basis for construction safety assessment.
Smart Images
Figure CN122046715A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of tunnel underpass construction monitoring and deformation analysis technology, specifically involving a method and system for predicting the displacement distribution of existing lines and the ground surface during underpass construction. Background Technology
[0002] As urban rail transit lines continue to expand, new tunnels inevitably need to pass under existing tunnels. Underpass construction causes significant disturbance to existing lines and the ground surface, potentially leading to track deformation and uneven settlement. Furthermore, the construction of new tunnels repeatedly disturbs adjacent strata and surrounding rock of existing tunnels, causing excessive displacement and deformation pressure, which can trigger surface subsidence, collapse, and other disasters. Therefore, understanding the uneven settlement curves of existing lines and the ground surface is crucial.
[0003] Current monitoring technologies mostly rely on measured data from fixed cross-sections or a small number of monitoring points, which can only obtain local information and are difficult to comprehensively and intuitively reflect the overall deformation distribution of existing lines and the ground surface. Although traditional fitting methods (such as the Peck equation) can be used to describe settlement curves, their accuracy is highly dependent on empirical parameters (settlement trough width parameter), making it difficult to adapt to the nonlinear response characteristics under complex geological and construction conditions. Summary of the Invention
[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and system for predicting the displacement distribution of existing lines and the ground surface during the construction of underpass tunnels. This solves the problems of existing existing line and ground surface displacement monitoring technologies, which are difficult to fully reflect the overall deformation distribution, have limited fitting accuracy, and rely on empirical parameters.
[0005] To achieve the above objectives, the technical solution of the present invention is as follows:
[0006] In a first aspect, the present invention provides a method for predicting the displacement distribution of existing railway lines and ground surfaces during underpass tunnel construction, characterized in that it includes:
[0007] Establish a mapping model between existing railway lines and surface displacement distribution and different coordinates;
[0008] Displacement data of monitoring points are obtained, and the displacement distribution of the entire line and the ground path is reconstructed using the mapping relationship model.
[0009] Optionally, the mapping relationship model is established in the following way:
[0010] The existing lines and surface paths are divided into several nodes, and the nodes corresponding to several locations on the existing lines and surface are selected as monitoring points.
[0011] The coordinate information of the monitoring point and the geometric characteristic parameters of the cross section where it is located are used as input variables, and the displacement data of the monitoring point is used as output variables to establish a mapping relationship model between input and output.
[0012] Optionally, monitoring points can be selected from the nodes located directly above and on both sides of the existing railway line and the midpoint of the newly constructed underpass.
[0013] Optionally, the method further includes:
[0014] By substituting the spatial location of the unmonitored node and the geometric characteristic parameters of its cross section into the mapping relationship model, the displacement distribution results of the continuous path are obtained, forming a complete description of the settlement morphology of the existing line and the ground surface.
[0015] Optionally, an RBF neural network can be used to establish a mapping relationship model between the coordinate information of the monitoring point, the geometric feature parameters of the cross section, and the displacement data of the monitoring point.
[0016] Optionally, a Gaussian kernel function can be used as the basis function for the RBF neural network;
[0017] in, For the input vector, ; Indicates the input dimension; For the first The center of each hidden layer neuron; This is a parameter that controls the width of the function; Represents a nonlinear transformation unit; This represents the Euclidean norm.
[0018] Optionally, the output of the RBF neural network is as shown in the formula;
[0019] in, This represents the output of the RBF neural network; For the first One input sample; Indicates the first The width of each nonlinear transformation unit; For the first The hidden neuron and the first Connection weights between output neurons; This represents the number of hidden layer units; For output dimensions.
[0020] Secondly, the present invention provides a system for predicting the displacement distribution of existing railway lines and ground surfaces during underpass tunnel construction, comprising:
[0021] The mapping relationship establishment module is used to establish a mapping relationship model between existing lines and surface displacement distribution and different coordinates;
[0022] The displacement distribution prediction module acquires displacement data from monitoring points and uses the mapping relationship model to reconstruct the displacement distribution of the entire route and the ground path.
[0023] Compared with the prior art, the advantages of this invention are as follows:
[0024] This invention establishes a mapping relationship model between the existing line and surface displacement distribution and different coordinates. It can reconstruct the deformation of the entire line and surface using a small number of displacement monitoring points. This not only reduces the number of monitoring devices and installation workload, but also displays the continuous settlement pattern of the existing line and surface with higher spatial resolution, and intuitively reflects the uneven settlement characteristics caused by the construction of underpass tunnels. Attached Figure Description
[0025] 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 embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0026] Figure 1 This is a flowchart of the fitting method for the displacement distribution of existing lines and ground surfaces during the close-range construction of the underpass tunnel in this embodiment;
[0027] Figure 2 This is a case overview of the fitting method for the displacement distribution of existing lines and ground surfaces during the close-range construction of the underpass tunnel in this embodiment, where (a) is a planar satellite image, (b) is a planar image, and (c) is a longitudinal profile.
[0028] Figure 3 A diagram illustrating the construction steps of a new underpass tunnel in the fitting method for the displacement distribution of existing lines and ground surfaces during close-proximity construction.
[0029] Figure 4 The numerical model is established, where (a) is the overall model, (b) is the relationship between the existing line and the newly built tunnel, and (c) is the existing line and the surface path;
[0030] Figure 5 To verify the fitting results of existing lines and ground displacements through numerical simulation;
[0031] Figure 6 Experimental diagrams for fitting the displacement distribution of existing railway lines and ground surfaces during close-range construction of underpass tunnels;
[0032] Figure 7 The model test was conducted to verify the fitting results of the existing line and ground displacement. Detailed Implementation
[0033] 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.
[0034] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0035] See Figure 1 As shown in the figure, the method for predicting the displacement distribution of existing lines and ground surfaces during underpass construction provided in this embodiment mainly includes the following steps:
[0036] Establish a mapping model between existing railway lines and surface displacement distribution and different coordinates;
[0037] Displacement data of monitoring points are obtained, and the displacement distribution of the entire line and the ground path is reconstructed using the mapping relationship model.
[0038] The mapping model reconstructs the displacement distribution of the entire route and surface path, enabling displacement prediction at unmonitored locations and outputting the overall displacement distribution response of the existing route and surface caused by the construction of the underpass tunnel. Thus, by collecting displacement change data from monitoring points and combining it with the established mapping model, the displacement distribution of the existing route and surface path under the construction conditions of the new underpass tunnel is fitted and reconstructed, enabling displacement prediction in unmonitored areas.
[0039] Therefore, this method can reconstruct the deformation of the entire line and the ground surface by establishing a mapping relationship model between the existing line and the ground surface displacement distribution and different coordinates, using a small number of displacement monitoring points. This not only reduces the number of monitoring devices and installation workload, but also displays the continuous settlement pattern of the existing line and the ground surface with higher spatial resolution, and intuitively reflects the uneven settlement characteristics caused by the construction of the underpass tunnel.
[0040] In a specific implementation, the mapping relationship model between the existing railway line and the ground surface displacement distribution and different coordinates is established in the following way:
[0041] The existing line and surface path are divided into several nodes, and the nodes corresponding to several locations of the existing line and surface are selected as monitoring points. For example, the monitoring points are selected at the midpoints of the existing line and surface located directly above and on both sides of the newly built underpass.
[0042] The coordinate information of the monitoring point and the geometric characteristic parameters of the cross section where it is located are used as input variables, and the displacement data of the monitoring point is used as output variables to establish a mapping relationship model between input and output.
[0043] In this way, by using a limited number of monitoring points and combining their coordinate information with cross-sectional geometric features, a mapping relationship model between node displacement and spatial location is established, thereby realizing the reconstruction of continuous displacement distribution of the line and surface path.
[0044] In a preferred embodiment, the method for predicting the displacement distribution of existing railway lines and ground surfaces during underpass construction further includes:
[0045] By substituting the spatial location and cross-sectional information of unmonitored nodes into the mapping relationship model, the displacement distribution results of the continuous path are obtained, thereby forming a complete description of the settlement morphology of the existing line and the ground surface.
[0046] In a specific embodiment, a radial basis function (RBF) model is used to establish the mapping relationship between existing railway lines and surface displacement distribution with different coordinates. A Gaussian kernel function is used as the basis function for the RBF neural network.
[0047] in, For the input vector, ; Indicates the input dimension; For the first The center of each hidden layer neuron; This is a parameter that controls the width of the function; Represents a nonlinear transformation unit; This represents the Euclidean norm.
[0048] The output of the RBF neural network is shown in the following formula;
[0049] in, This represents the output of the RBF neural network; For the first One input sample; Indicates the first The width of each nonlinear transformation unit; For the first The hidden neuron and the first Connection weights between output neurons; This represents the number of hidden layer units; For output dimensions.
[0050] In this way, the above method can accurately and efficiently establish a mapping relationship model between existing lines and surface displacement distribution and different coordinates.
[0051] Accordingly, this embodiment provides a system for predicting the displacement distribution of existing railway lines and ground surfaces during underpass tunnel construction, including:
[0052] The mapping relationship establishment module is used to establish a mapping relationship model between existing lines and surface displacement distribution and different coordinates.
[0053] The displacement distribution prediction module is used to reconstruct the displacement distribution of the entire route and the ground path using the mapping relationship model.
[0054] It should be noted that the specific working principles of the mapping relationship establishment module and the displacement distribution prediction module are the same as the methods and steps described above, and will not be repeated here.
[0055] The following examples will further verify and illustrate the method for predicting the displacement distribution of existing lines and ground surfaces during the construction of this underpass tunnel.
[0056] Application Example 1:
[0057] The new tunnel, constructed using the mining method, connects the new station with the adjacent section. It is 39.55 m long, with the upper portion situated in a layer of sand and gravel, and the lower portion in moderately weathered mudstone. The vertical clearance between the new tunnel and the existing subway shield tunnel section is 8.4 m.
[0058] The simulated excavation steps are as follows: Figure 3 As shown in the figure, ①-③ represent pilot tunnels, which include the following steps:
[0059] In the initial stages of construction of the new tunnel and adjacent sections, advanced support measures were implemented to ensure tunnel construction safety. These measures included installing double-layer Φ194 pipe roofs within the tunnel arch area. The initial support for the upper section of the new tunnel consisted of steel arch frames and 350 mm thick early-strength C25 concrete sprayed onto the surface. The first layer of secondary lining in the upper section was 400 mm thick C35 cast-in-place concrete, integrally cast with the capping beam. After the temporary partition walls were removed, waterproofing was carried out, and a second layer of secondary lining, 500 mm thick C35 cast-in-place concrete, was installed.
[0060] like Figure 4 As shown in (a), a numerical model with a geometric similarity ratio of 1:1 is constructed; as Figure 4As shown in (b), the relationship between the existing line and the newly built tunnel is illustrated; Figure 4 As shown in (c), the existing railway line and surface path are represented. To effectively eliminate boundary effects, the finite element model size is set to 60 m × 40 m × 60 m (length × width × height). Axial displacement is constrained at the front and rear boundaries, horizontal displacement is constrained at the left and right boundaries, and triaxial displacement is constrained at the bottom. The surface is free and a uniformly distributed load of 20 kPa is applied. The soil is an ideal elastoplastic material, conforming to the Mohr-Coulomb yield criterion, and the initial stress field of the stratum originates from its own weight. The construction process is simulated numerically, and the calculation steps are consistent with the actual construction sequence. The soil, existing station structure, existing shield tunnel structure, and new tunnel structure are all modeled using first-order eight-node hexahedral solid elements (C3D8), while the temporary support structure and the central partition wall are modeled using four-node shell elements (S4). The "birth and death element" function is used to simulate soil excavation and support construction. Soil reinforcement is achieved by increasing material parameters. The support piles of the new tunnel are converted into continuous walls according to the principle of equivalent stiffness.
[0061] like Figure 5 As shown, numerical simulations validate the fitting results for existing line and surface displacement. Using the proposed method, displacement data from both ends and the midpoint of the path were used to fit the existing line (L1–L2) and the surface (L3–L4). The maximum fitting errors for L1–L2 were 5.8% and 5.6%, respectively, both occurring in step 6 (excavation of the lower half of the cross-section), near the shield-station junction; the average errors were below 1.7% and 1.6%, validating the reliability of the method. The maximum fitting errors for L3–L4 were 6.1% and 5.6%, both occurring above the station; the average errors were below 1.6% and 1.5%. The errors mainly stemmed from the mitigating effect of the station structure on surface settlement. The overall errors were within acceptable limits.
[0062] Application Example 2:
[0063] The new tunnel, constructed using the mining method, connects the new station with the adjacent section. It is 39.55 m long, with the upper portion situated in a layer of sand and gravel, and the lower portion in moderately weathered mudstone. The vertical clearance between the new tunnel and the existing subway shield tunnel section is 8.4 m.
[0064] The excavation steps of the experiment are as follows Figure 3As shown, in the initial stage of construction of the new tunnel and adjacent sections, advanced support measures were adopted to ensure tunnel construction safety. These measures included installing double-layer Φ194 pipe roofs within the arch area of the new tunnel. The initial support for the upper part of the new tunnel used steel arch frames and sprayed with 350 mm thick early-strength C25 concrete. The first layer of secondary lining in the upper part was 400 mm thick C35 cast-in-place concrete, integrally cast with the capping beam. After the temporary partition walls were removed, waterproofing was carried out, and a second layer of secondary lining, namely 500 mm thick C35 cast-in-place concrete, was installed.
[0065] like Figure 6 As shown, a model box with a geometric similarity ratio of 1:40 and geometric dimensions of 2 m × 1 m × 1.3 m was constructed, using undisturbed soil from the construction site. The construction process was simplified into 6 stages, including the following steps:
[0066] Laser displacement gauges were deployed at the midpoint of paths L1 and L2 and 0.5 m on both sides to monitor vertical displacement. Displacement of the existing track was monitored throughout the simulated excavation process. Fiber Bragg grating (FBG) sensors were deployed at the bottom of the track to acquire bending strain distribution, and the global displacement was calculated using the conjugate beam method as a control.
[0067] like Figure 7 As shown, based on the displacement data of points D1, D2, and D3 along path L1–L4, an RBF neural network is used to fit the overall displacement distribution of the path and the surface. The maximum error for path L1 is 5.43%, and the average error is 0.75%; the maximum error for path L2 is 4.83%, and the average error is 0.67%. The fitting effect is good; under limited monitoring points, the maximum error is less than 6%, and the average error is less than 3.5%. The proposed method can intuitively reflect the overall displacement evolution, especially in characterizing uneven settlement features.
[0068] In summary, the method of the present invention can intuitively and accurately reconstruct the displacement distribution of existing lines and ground surfaces during the underpass construction process, even with a limited number of monitoring points, providing a reliable quantitative basis for construction safety assessment and deformation control.
[0069] The above embodiments are merely illustrative of the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made based on the essence of the content of the present invention should be covered within the scope of protection of the present invention.
Claims
1. A method for predicting the displacement distribution of existing railway lines and ground surfaces during underpass tunnel construction, characterized in that, include: Establish a mapping model between existing railway lines and surface displacement distribution and different coordinates; Displacement data of monitoring points are obtained, and the displacement distribution of the entire line and the ground path is reconstructed using the mapping relationship model.
2. The method for predicting the displacement distribution of existing railway lines and ground surfaces during underpass construction as described in claim 1, characterized in that, The mapping relationship model is established in the following way: The existing lines and surface paths are divided into several nodes, and the nodes corresponding to several locations on the existing lines and surface are selected as monitoring points. The coordinate information of the monitoring point and the geometric characteristic parameters of the cross section where it is located are used as input variables, and the displacement data of the monitoring point is used as output variables to establish a mapping relationship model between input and output.
3. The method for predicting the displacement distribution of existing lines and ground surfaces during underpass tunnel construction as described in claim 2, characterized in that, The nodes corresponding to the locations of the existing railway line and the midpoints on both sides of the newly constructed underpass tunnel were selected as monitoring points.
4. The method for predicting the displacement distribution of existing railway lines and ground surfaces during underpass construction as described in claim 2 or 3, characterized in that, The method further includes: By substituting the spatial location of the unmonitored node and the geometric characteristic parameters of its cross section into the mapping relationship model, the displacement distribution results of the continuous path are obtained, forming a complete description of the settlement morphology of the existing line and the ground surface.
5. The method for predicting the displacement distribution of existing railway lines and ground surfaces during underpass tunnel construction as described in claim 2, characterized in that, An RBF neural network is used to establish a mapping relationship model between the coordinate information of the monitoring point, the geometric feature parameters of the cross section, and the displacement data of the monitoring point.
6. The method for predicting the displacement distribution of existing railway lines and ground surfaces during underpass tunnel construction as described in claim 5, characterized in that, A Gaussian kernel function is used as the basis function for the RBF neural network; ； in, For the input vector, ; Indicates the input dimension; For the first The center of each hidden layer neuron; This is a parameter that controls the width of the function; Represents a nonlinear transformation unit; This represents the Euclidean norm.
7. The method for predicting the displacement distribution of existing railway lines and ground surfaces during underpass construction as described in claim 6, characterized in that, The output of the RBF neural network is as shown in the formula; ； in, This represents the output of the RBF neural network; For the first One input sample; Indicates the first The width of each nonlinear transformation unit; For the first The hidden neuron and the first Connection weights between output neurons; This represents the number of hidden layer units; For output dimensions.
8. A system for predicting the displacement distribution of existing railway lines and ground surfaces during underpass tunnel construction, characterized in that, include: The mapping relationship establishment module is used to establish a mapping relationship model between existing lines and surface displacement distribution and different coordinates; The displacement distribution prediction module acquires displacement data from monitoring points and uses the mapping relationship model to reconstruct the displacement distribution of the entire route and the ground path.
9. The system for predicting the displacement distribution of existing railway lines and ground surfaces during underpass construction as described in claim 8, characterized in that, The mapping relationship model is established in the following way: The existing lines and surface paths are divided into several nodes, and the nodes corresponding to several locations on the existing lines and surface are selected as monitoring points. The coordinate information of the monitoring point and the geometric characteristic parameters of the cross section where it is located are used as input variables, and the displacement data of the monitoring point is used as output variables to establish a mapping relationship model between input and output.
10. The system for predicting the displacement distribution of existing railway lines and ground surfaces during underpass construction as described in claim 9, characterized in that, Monitoring points were selected from the nodes located directly above and on both sides of the existing railway line and the midpoint of the newly constructed underpass.