A refined evaluation method for the stability of tunnel face

Abstract

The present invention relates to a refined evaluation method for the stability of a tunnel face, comprising the following steps: S1. A camera is arranged on each side of the tunnel, and binocular photography technology is used to collect the face images; S2. Based on the collected multiple face images, the three-dimensional spatial deformation of the face is obtained; S3. The face image is divided into grids, and based on the deformation of the center points of each grid, each grid is merged to obtain multiple face regions and the surrounding rock grades of each face region; S4. Based on the linear relationship pre-established between the equivalent deformation of the face and the deformation of each region, the corresponding equivalent deformation of the current face is calculated, and the linear relationship is constructed based on the idea of equivalent plastic strain energy; S5. The relationship between the corresponding equivalent deformation and the allowable deformation and the ultimate deformation is judged to obtain the evaluation result of the face stability, and the allowable deformation and the ultimate deformation are obtained by means of machine learning. Compared with the prior art, the present invention has the advantages of accurate and reasonable determination results, etc.

Description

Technical Field

[0001] The present invention belongs to the technical field of tunnel engineering construction safety analysis, and relates to a method for determining the stability of a tunnel face, in particular to a refined evaluation method for the stability of a tunnel face. Background Art

[0002] In actual tunnel engineering, constantly mastering the stability state of the tunnel is a key task to ensure the safe construction of tunnel propulsion. At present, there are various methods for judging tunnel instability, such as the judgment of plastic zone penetration, the judgment of the size of the plastic zone, the judgment of plastic strain energy, etc. However, in actual engineering, considering the feasibility, on-site construction mainly uses the deformation data of construction monitoring measurement as the judgment index, and currently mainly judges through whether the indexes such as the deformation amount and the deformation rate exceed the limit. For example, Patent Application No. 202110366110.X discloses a method for interpreting the stability of a tunnel based on three-dimensional deformation monitoring of the tunnel face, which monitors the crown settlement amount and the horizontal convergence amount within the distance of one excavation cycle on the tunnel face in real time, and at the same time uses an inclinometer to monitor the extrusion displacement in front of the tunnel face, realizing the three-dimensional monitoring of the deformation of the tunnel face; the stability of the tunnel face is judged through three deformation indexes: the displacement change rate of the measuring point, the displacement change amount of the measuring point, and the extrusion deformation of the tunnel face.

[0003] The main operation processes and technical solutions of the existing methods are as follows:

[0004] 1) Advanced geological prediction and preliminary evaluation of surrounding rock. Before tunnel excavation, first use advanced horizontal drilling to conduct advanced geological prediction of the tunnel face, timely obtain the geological information of the tunnel surrounding rock, and make a preliminary judgment on the stability of the surrounding rock in front of the tunnel face according to the detection results.

[0005] 2) Classification of the surrounding rock grade of the tunnel face. After tunnel excavation, engineering technicians conduct geological sketching of the tunnel face in combination with the state of the tunnel face exposed by excavation, and then classify the surrounding rock of the tunnel face.

[0006] 3) Construction monitoring measurement. After tunnel support, the crown settlement amount and the horizontal convergence amount within the distance of one excavation cycle on the tunnel face are monitored in real time, and at the same time, an inclinometer is used to monitor the extrusion displacement in front of the tunnel face by drilling surface holes at a fixed distance in front of the tunnel face, measuring the extrusion deformation of the surrounding rock in front of the tunnel face, so as to realize the three-dimensional monitoring of the deformation of the tunnel face.

[0007] 4) Determination of stability criterion. According to the actual working conditions, establish a corresponding numerical calculation model, adopt the strength reduction method, gradually reduce the surrounding rock parameters until the calculation does not converge, and use the corresponding deformation value as the control index.

[0008] 5) Stability judgment. The displacement change amount index of the measuring point is used to judge the stability of the tunnel face. When the index exceeds the control value, it is determined that the tunnel face is in an unstable state.

[0009] Using the above existing technologies, the prominent problems faced are as follows:

[0010] (1) The area of the heading face region is relatively large and often shows non-uniformity, that is, there are obvious differences in the surrounding rock grades between local regions. At present, the means of grading the surrounding rock of the heading face are based on the determination of the entire heading face, ignoring the non-uniformity of the heading face, and the determination results cannot accurately represent the surrounding rock grades of local non-uniform parts. However, during the construction process, larger deformations will occur in the weakened area of the heading face, which will in turn cause local collapses of the tunnel; in a small number of studies on the zonal grading of the heading face, the data of the drill-in holes are used for local grading, and the time to obtain the data is relatively long, and the timeliness is poor.

[0011] (2) Due to joint cutting or local soft failure, the deformation of the heading face also shows regional characteristics. The current stability determination method selects the extrusion deformation of the core soil part of the heading face as the evaluation index. For the heading face of homogeneous surrounding rock, the extrusion deformation of the core soil part is the maximum deformation of the whole section; however, the heading face often shows non-uniformity, and individual characteristic points cannot represent the deformation state of the whole section. The maximum extrusion deformation does not necessarily occur in the middle area of the heading face. Therefore, the extrusion deformation of the core soil area obviously cannot represent the overall deformation distribution of the actual heading face, and there are defects in the existing evaluation system. Summary of the Invention

[0012] The purpose of the present invention is to provide a refined evaluation method for the stability of tunnel heading faces with accurate and reasonable determination results to overcome the defects of the above existing technologies.

[0013] The purpose of the present invention can be achieved through the following technical solutions:

[0014] A refined evaluation method for the stability of tunnel heading faces includes the following steps:

[0015] S1. Install a camera on each side of the tunnel, and use binocular photography technology to collect the images of the heading face;

[0016] S2. Obtain the three-dimensional space deformation of the heading face based on the collected multiple images of the heading face;

[0017] S3. Divide the images of the heading face into grids, and merge the grids based on the deformation of the center points of each grid to obtain multiple heading face regions and the surrounding rock grades of each heading face region;

[0018] S4. Calculate the corresponding equivalent deformation of the current heading face based on the linear relationship between the equivalent deformation of the heading face and the deformation of each region pre-constructed, and the linear relationship is constructed based on the idea of equivalent plastic strain energy;

[0019] S5. Determine the relationship between the corresponding equivalent deformation, allowable deformation, and ultimate deformation to obtain the evaluation result of the tunnel face stability. The allowable deformation and ultimate deformation are obtained based on machine learning means.

[0020] Further, in step S1, the distance between the camera and the tunnel face is 30m to 50m.

[0021] Further, the three-dimensional space deformation of the tunnel face is calculated through the following steps:

[0022] Based on the synchronous images taken by two cameras, use the SGBM algorithm to calculate the disparity of the two synchronous images;

[0023] According to the geometric relationship of parallel binocular vision, calculate the three-dimensional space information of the image;

[0024] Based on the comparison of multiple pictures, calculate the three-dimensional space deformation of the tunnel face.

[0025] Further, the specific merging of each grid is as follows:

[0026] Based on the deformation values of the center points of each grid obtained from multiple observations, form a deformation matrix, use the empirical orthogonal analysis method to calculate the first eigenvector of the deformation matrix, use the density clustering algorithm to automatically cluster the deformation data in the first eigenvector, and merge each grid based on the clustering results.

[0027] Further, the steps for obtaining the first eigenvector include:

[0028] Perform standardization processing on the original deformation matrix X, and solve the covariance matrix A = XX T , use the Jacobi method to solve all the eigenvalues and corresponding eigenvectors of the covariance matrix A, and select to obtain the first eigenvector.

[0029] Further, the surrounding rock grades of each tunnel face area are determined based on the following method:

[0030] Use image processing technology to extract the joint occurrence, tunnel face roughness, surrounding rock color information, and seepage water area information in the corresponding area, and determine the surrounding rock grades of each area.

[0031] Further, the construction of the linear relationship is specifically as follows:

[0032] 1) Establish a partitioned tunnel face calculation model, marked as model 1, and the parameters of each area are taken according to the surrounding rock grade;

[0033] 2) Use the strength reduction method to synchronously reduce the model parameters until the calculation does not converge, and statistically calculate the plastic strain energy of each divided area;

[0034] 3) Establish a homogeneous tunnel face calculation model, marked as model 2, calculate the BQ value of the entire section based on the surrounding rock grade of each area, and select model parameters;

[0035] 4) Use the strength reduction method to reduce the model parameters synchronously until the calculation does not converge, and calculate the total strain plastic properties of the model;

[0036] 5) Based on the energy equivalence principle, determine whether E=E1+E2+…+En exists, where n is the number of partitions. If so, execute step 6); if not, return to step 3);

[0037] 6) Record the deformation of each region of model 1 Δ1, Δ2, ..., Δn as the characteristic value of the deformation of each region, record the maximum extrusion deformation Δ of model 2 as the characteristic value of the deformation of the tunnel face, and construct a set of samples {Δ1, Δ2, L, Δn, Δ};

[0038] 7) Repeat steps 1) to 6) multiple times to build a sample library, and use the fitting method to obtain the linear relationship between the equivalent deformation of the tunnel face and the deformation value of each partition:

[0039] Δ=β1·Δ1+β2·Δ2+L+β n ·Δn

[0040] In the formula, β1, β2, L, β n is the deformation weight value.

[0041] Furthermore, the extreme deformation is obtained by calculating a tunnel face extreme deformation prediction model, and the model parameters of the tunnel face extreme deformation prediction model are determined by training using machine learning methods. The tunnel face extreme deformation prediction model characterizes the correspondence between the tunnel burial depth, cross-sectional dimensions, and tunnel face surrounding rock strength distribution and the extreme deformation when the tunnel face becomes unstable.

[0042] Furthermore, the training sample data used in training the extreme deformation prediction model of the tunnel face include tunnel burial depth values, tunnel rise-to-span ratio values, deformation characteristic values of surrounding rock regions of various grades, area percentages of surrounding rock regions of various grades, and extreme deformation values.

[0043] Furthermore, the allowable deformation is calculated based on the limit deformation, and the calculation formula is:

[0044]

[0045] Where Δ′ is the limit deformation, [Δ] is the allowable deformation, and [σ] is the safety factor.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] 1) The present invention uses binocular photography technology to obtain deformation data of the tunnel face observed multiple times and establish a deformation matrix; the empirical orthogonal analysis method is combined with a density clustering algorithm to automatically cluster the deformation data, thereby merging and dividing the tunnel face area, and marking the regional surrounding rock grade, thereby achieving rapid and refined grading of the tunnel face, accurately characterizing the distribution state of the surrounding rock of the entire section of the tunnel face, and thus achieving accurate and refined stability evaluation. The present invention can guide engineering practice and reinforce weak surrounding rock areas in a timely manner.

[0048] 2) Based on the idea of plastic strain energy equivalence, the present invention calculates the corresponding equivalent deformation of the face, and uses equivalent deformation to replace the traditional deformation of the core area of the face, which can better characterize the deformation state of the face; at the same time, the equivalent deformation is used as a criterion for the stability of the surrounding rock, and the judgment result is more reasonable. The stability judgment method for uneven faces proposed by the present invention is more reasonable.

[0049] 3) Based on a large number of finite element calculation results, the present invention focuses on analyzing the influence of section burial depth, section span ratio, and face surrounding rock distribution on the ultimate deformation of the face. It uses machine learning to train the face ultimate deformation prediction model, calculates the allowable displacement, and provides stability judgment indicators to establish a more reasonable stability analysis system to ensure construction safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a structural schematic diagram of the present invention;

[0051] Figure 2 Schematic diagram of mesh division of the tunnel face in the embodiment;

[0052] Figure 3 This is a schematic diagram of the marking of the tunnel face area in the embodiment;

[0053] Figure 4 Schematic diagram of the division of the tunnel face area and the determination of the level in the embodiment;

[0054] Figure 5 Schematic diagram of a partitioned tunnel face calculation model established in the embodiment;

[0055] Figure 6 Schematic diagram of the homogeneous tunnel face calculation model established in the embodiment. DETAILED DESCRIPTION

[0056] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0057] Explanation of terms

[0058] Tunnel face: Tunnel face, also known as tunnel face, is a term used in tunnel construction. It refers to the working face that continuously advances in tunnel excavation (coal mining, mining or tunnel engineering).

[0059] Binocular photogrammetry: Use two cameras for positioning. For a feature point on an object, use two cameras fixed at different positions to take images of the object and obtain the coordinates of the point on the image planes of the two cameras. As long as the precise relative positions of the two cameras are known, the coordinates of the feature point in the coordinate system of a fixed camera can be obtained by geometric methods, that is, the position of the feature point is determined.

[0060] Example

[0061] This embodiment provides a refined evaluation method for tunnel face stability. Based on binocular photogrammetry technology, after calculating the deformation data of the tunnel face, the empirical orthogonal analysis method and the adaptive clustering algorithm are used to automatically divide the tunnel face area. Based on the area division, with the help of machine learning, an indicator that can better represent the changes in the tunnel face - the equivalent deformation of the tunnel face is constructed; and based on the equivalent deformation, construction control indicators and stability judgment indicators are proposed to realize the judgment of tunnel face stability. Figure 1 As shown, the method specifically comprises the following steps:

[0062] S1. Camera deployment and face image acquisition. Using binocular photography technology, a camera is deployed on each side of the tunnel for synchronous acquisition; the camera is optimally placed 30m to 50m away from the face to ensure that a complete face image can be acquired. In a preferred embodiment, fill-in lights are turned on during the shooting process, and fill-in lighting equipment is provided when necessary to ensure that the entire face is clearly visible and free of shadows.

[0063] S2. Image processing and calculation of tunnel face deformation. The collected tunnel face images are first preprocessed to eliminate image noise caused by the dark and dusty environment of the tunnel; based on the images taken by the left and right cameras at the same time, the SGBM algorithm is used to calculate the disparity of the two images; based on the geometric relationship of parallel binocular vision, the three-dimensional spatial information of the image is calculated; based on the comparison of multiple images, the three-dimensional spatial deformation of the tunnel face is calculated.

[0064] S3. Division and grading of the tunnel face area. The tunnel face image is divided into grids, and the deformation of each grid center point is calculated; a deformation matrix is constructed based on multiple observed deformation values, and the first eigenvector of the deformation matrix is calculated using the empirical orthogonal analysis method; a density clustering algorithm is used to automatically cluster the deformation data in the first eigenvector to merge and divide the tunnel face area; at the same time, an image processing algorithm is used to quantify parameters such as water seepage, joints, and strength in the tunnel face area, and the regional surrounding rock grade is calculated and marked.

[0065] The large-section tunnel face has obvious non-uniformity, which is mainly reflected in the presence of local rock-level weakening areas in the tunnel face surrounding rock, or obvious broken areas due to the cutting effect of joints; the deformation of the tunnel face is mainly reflected in the local deformation mutation, which is inconsistent with the deformation of adjacent areas. This method uses the tunnel face area division method based on empirical orthogonal analysis to realize the division of the tunnel face area.

[0066] (1) Deformation feature extraction based on empirical orthogonal method

[0067] In order to simplify the calculation, the tunnel face is divided into grids, and the grid centroid is used as the regional representative. Figure 2 As shown, point A is the displacement calculation point for grid division; at the same time, the denser the grid division, the greater the amount of calculation, but the more accurate the subsequent partition results.

[0068] Taking the tunnel face image after blasting as the benchmark, tunnel face images are collected at regular intervals to compare and calculate the displacement changes of each point.

[0069] The data is recorded in matrix form:

[0070]

[0071] Where: m——is the observation point number;

[0072] n——is the time series, that is, the number of observations.

[0073] The empirical orthogonal method is used to decompose the above matrix into two parts: time function Z and space function V:

[0074] X=VZ (2-1)

[0075]

[0076]

[0077] The main processing steps are as follows:

[0078] 1) According to the analysis purpose, the original data matrix X is standardized;

[0079] 2) Solve the covariance matrix A=XX from the data matrix T ;

[0080] 3) Using Jacobi method, solve all eigenvalues and corresponding eigenvectors of symmetric matrix A;

[0081] 4) Select the first eigenvector as the representative vector of the spatial deformation feature of the observation point.

[0082] (2) Tunnel face area division based on density clustering

[0083] Based on the first eigenvector data values, combined with the deformation direction data as classification samples, the density clustering algorithm is used to automatically cluster and divide each deformation data. On the tunnel face grid, it means marking and merging the representative areas of the same type of observation points, and finally dividing the tunnel face areas with different deformation characteristics. As Figure 3 shown, the tunnel face in this embodiment is divided into 4 areas.

[0084] (3) Judgment of the surrounding rock grades of each area based on image recognition

[0085] Based on the above-mentioned merged and divided tunnel face areas, using image processing technology, extract the joint occurrence, tunnel face roughness, surrounding rock color information, and leakage area information in the corresponding areas, automatically calculate the BQ value, and judge the surrounding rock grade of the corresponding area according to the current "Highway Tunnel Design Code". As Figure 4 shown, different grades of the 4 areas are marked.

[0086] S4. Calculation of the equivalent deformation of the tunnel face. Based on the idea of equivalent plastic strain energy, calculate the corresponding equivalent deformation of the tunnel face, and construct the deformation samples of each area of the tunnel face and the equivalent deformation of the tunnel face; based on a large number of data samples, use data fitting processing means to obtain the linear relationship between the equivalent deformation and the deformation of each area; use this linear relationship as the equivalent deformation calculation formula to calculate the equivalent deformation of each tunnel face.

[0087] Tunnel excavation will cause stress redistribution, the stress state of the surrounding rock changes, and energy dissipation and release occur during the entire stress adjustment process. The displacement and deformation of the tunnel face are an external manifestation of the result of energy dissipation.

[0088] The deformation results of different areas of the tunnel face correspond to the plastic strain energy of different areas. Here, the method of equivalent plastic strain energy is used to establish an equivalent deformation calculation model of the tunnel face.

[0089] The main calculation process is as follows:

[0090] 1) Use finite element calculation software to establish a partitioned tunnel face calculation model (Model 1), as Figure 5 shown, the constitutive model of the model adopts the Mohr-Coulomb constitutive model, and the parameters of each area are taken according to the calculated BQ value in accordance with the current "Highway Tunnel Design Code";

[0091] 2) Use the strength reduction method to synchronously reduce the model parameters until the calculation does not converge; count and calculate the plastic strain energy (E1, E2, E3, E4) of each divided area;

[0092] The strain energy calculation method is as follows:

[0093]

[0094] Where: E——total plastic strain energy of the area;

[0095] E I ——plastic strain energy of a single plastic element;

[0096] V I ——volume of a single plastic element;

[0097] σ ij ——stress tensor of a single plastic element;

[0098] ε ij ——strain tensor of a single plastic element;

[0099] 3) Establish a homogeneous heading face calculation model (Model 2) using finite element calculation software, as Figure 6 shown. The constitutive model of the model adopts the Mohr-Coulomb constitutive model. Calculate the BQ value according to the above full-section BQ value calculation method, and select the parameter values according to the current "Highway Tunnel Design Code";

[0100] 4) Use the strength reduction method to synchronously reduce the model parameters until the calculation does not converge, and calculate the total strain plastic energy E of the calculation model;

[0101] 5) Based on the energy equivalence principle, judge the magnitude relationship between E and E1+E2+E3+E4; if they are not equal, readjust the parameters of the homogeneous model and repeat steps 3) and 4) until E = E1+E2+E3+E4;

[0102] 6) Record the deformations Δ1, Δ2, Δ3, Δ4 of each area of Model 1 as the characteristic values of the deformations of each area;

[0103] 7) Record the maximum extrusion deformation Δ of Model 2 as the characteristic value of the heading face deformation;

[0104] 8) Use the above data to construct a set of samples {Δ1, Δ2,..., Δ4, Δ};

[0105] 9) Repeat steps 1) to 8) multiple times to construct a sample library, and use the fitting method to obtain the linear relationship between the equivalent deformation of the heading face and the deformation values of each partition, and construct a calculation model.

[0106] Δ = β1·Δ1 + β2·Δ2 + β3·Δ3 + β4·Δ4 (4)

[0107] Where: Δ——equivalent deformation value of the heading face;

[0108] Δ1——characteristic value of the deformation of Area ①;

[0109] Δ2——characteristic value of the deformation of Area ②;

[0110] Δ3——characteristic value of the deformation of Area ③;

[0111] Δ4——deformation characteristic value of area ④;

[0112] β1——Deformation weight value of region ①;

[0113] β2——Deformation weight value of area ②;

[0114] β3——Deformation weight value of area ③;

[0115] β4——Deformation weight value of area ④.

[0116] S5. Evaluation of tunnel face stability. Based on a large number of finite element calculation results, the influence of section depth, section span ratio, and tunnel face surrounding rock distribution on the tunnel face limit deformation is analyzed in detail; machine learning is used to train the tunnel face limit deformation prediction model, and the allowable displacement, i.e., allowable deformation, is calculated at the same time; when the tunnel face equivalent deformation is greater than the allowable displacement, an early warning needs to be issued and response measures need to be taken to deal with it; when the tunnel face equivalent deformation is greater than the limit deformation, the tunnel face is unstable.

[0117] On the basis of the above finite element calculation results, the influence of tunnel depth, cross-section size and strength distribution of surrounding rock of the tunnel face on the ultimate extrusion deformation when the tunnel face becomes unstable is analyzed in detail.

[0118] The SVM method is used to mine the quantitative relationship between the above indicators:

[0119] 1) Construct a training sample set, where the sample data includes tunnel depth value, tunnel span ratio value, deformation characteristic value of surrounding rock area of each grade, area percentage of surrounding rock area of each grade and limit deformation value. A single sample can be expressed as follows:

[0120] {H,k,Δ1,S1,Δ2,S2,Δ3,S3,Δ4,S4,Δ5,S5,Δ′}

[0121] Where: H is the tunnel depth value;

[0122] K——the ratio of tunnel rise to span;

[0123] Δ1——is the deformation characteristic value of the surrounding rock area of level I;

[0124] S1——area percentage of Grade I surrounding rock area;

[0125] Δ2——is the deformation characteristic value of the II-level surrounding rock area;

[0126] S2——area percentage of Grade II surrounding rock area;

[0127] Δ3——deformation characteristic value of the surrounding rock area of level III;

[0128] S3——Percentage of the area of grade Ⅲ surrounding rock area;

[0129] Δ4——Deformation characteristic value of grade Ⅳ surrounding rock area;

[0130] S4——Percentage of the area of grade Ⅳ surrounding rock area;

[0131] Δ5——Deformation characteristic value of grade Ⅴ surrounding rock area;

[0132] S5——Percentage of the area of grade Ⅴ surrounding rock area;

[0133] Δ′——Ultimate deformation value;

[0134] 2) Divide the samples into a training set and a test set, use the training set data as input parameters to train a machine learning model; use a testing machine to test the model, and if the accuracy reaches 90%, it is considered to meet the requirements.

[0135] Using the above machine learning model, the ultimate deformation value Δ′ of the tunnel face under various burial depths, cross-sections, and geological conditions can be obtained. To meet the engineering safety requirements, the allowable deformation value is used as the construction control index, and the calculation method of the allowable deformation value is as follows:

[0136]

[0137] In the formula: Δ′——Ultimate deformation;

[0138] [Δ]——Allowable deformation;

[0139] [σ]——Safety factor.

[0140] In summary:

[0141] 1) Δ′ is the ultimate deformation. When the equivalent deformation of the tunnel face reaches Δ′, the tunnel face becomes unstable;

[0142] 2) [Δ] is the allowable deformation. When the equivalent deformation of the tunnel face reaches [Δ], corresponding control measures need to be taken.

[0143] The above stability evaluation method based on the zonal deformation of the tunnel face, compared with the traditional tunnel face stability analysis method, realizes the rapid and refined classification of the tunnel face, accurately characterizes the distribution state of the surrounding rock of the entire cross-section of the tunnel face. Based on a large number of finite element calculation results, a machine learning method is used to train a prediction model for the ultimate deformation of the tunnel face, and more effective and reliable ultimate deformation can be obtained, thus establishing a more reasonable stability analysis system, which can ensure construction safety.

[0144] When the above method is implemented in the form of software functional units and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs.

[0145] The preferred specific embodiments of the present invention have been described in detail above. It should be understood that those of ordinary skill in the art can make many modifications and variations based on the concept of the present invention without creative efforts. Therefore, all technical solutions that can be obtained by those skilled in the art in the technical field based on the concept of the present invention through logical analysis, reasoning, or limited experiments on the basis of the prior art should be within the protection scope determined by the claims.

Claims

1. A refined evaluation method for the stability of a tunnel heading face, characterized in that It includes the following steps: S1. Install one camera on each side of the tunnel, and use binocular photography technology to collect the face images; S2. Obtain the three-dimensional spatial deformation of the face based on the collected multiple face images; S3. Divide the face images into grids, and based on the deformation of the center points of each grid, merge each grid to obtain multiple face regions and the surrounding rock grades of each face region; S4. Calculate the corresponding equivalent deformation of the current face based on the linear relationship between the equivalent deformation of the face and the deformation of each region pre-established, and the linear relationship is constructed based on the idea of equivalent plastic strain energy; S5. Judge the relationship between the corresponding equivalent deformation and the allowable deformation and the ultimate deformation to obtain the face stability evaluation result, and the allowable deformation and the ultimate deformation are obtained based on machine learning means; The construction of the linear relationship is specifically as follows: 1) Establish a partitioned face calculation model, marked as Model 1, and the parameters of each region are taken according to the surrounding rock grade; 2) Adopt the strength reduction method to synchronously reduce the model parameters until the calculation does not converge, and statistically calculate the plastic strain energy of each divided region; 3) Establish a homogeneous face calculation model, marked as Model 2, calculate the full-section BQ value based on the surrounding rock grades of each region, and select the model parameters; 4) Adopt the strength reduction method to synchronously reduce the model parameters until the calculation does not converge, and calculate the total strain plastic energy of the model; 5) Based on the energy equivalence principle, judge whether there is E = E1 + E2 + … + En, where n is the number of partitions. If so, execute step 6). If not, return to step 3); 6) Record the deformations Δ1, Δ2, …, Δn of each region of Model 1 as the characteristic values of the deformations of each region, record the maximum extrusion deformation Δ of Model 2 as the face deformation characteristic value, and construct a set of samples {Δ1, Δ2, …, Δn, Δ}; 7) Repeat steps 1) to 6) multiple times to construct a sample library, and use the fitting method to obtain the linear relationship between the equivalent deformation of the face and the deformation values of each partition; Δ = β1·Δ1 + β2·Δ2 + … + β n ·Δn where β1, β2, …, β n are deformation weight values.

2. The refined evaluation method for the stability of the tunnel heading face according to claim 1, wherein In step S1, the distance between the camera and the face is 30m to 50m.

3. The refined evaluation method for the stability of the tunnel face according to claim 1, wherein The three-dimensional spatial deformation of the face is calculated through the following steps: Based on the synchronous images taken by two cameras, use the SGBM algorithm to calculate the disparity of the two synchronous images; According to the geometric relationship of parallel binocular vision, calculate the three-dimensional spatial information of the images; Based on the comparison of multiple pictures, calculate the three-dimensional spatial deformation of the face.

4. The refined evaluation method for the stability of tunnel heading face according to claim 1, wherein The specific merging of each grid is as follows: Based on the deformation values of the center points of each grid observed multiple times, form a deformation matrix, use the empirical orthogonal analysis method to calculate the first eigenvector of the deformation matrix, use the density clustering algorithm to automatically cluster each deformation data in the first eigenvector, and merge each grid based on the clustering result.

5. The refined evaluation method for the stability of the tunnel face according to claim 4, characterized in that The steps for obtaining the first eigenvector include: Standardize the original deformation matrix X and solve the covariance matrix A = XX T , and use the Jacobi method to solve all the eigenvalues and corresponding eigenvectors of the covariance matrix A, and select the first eigenvector obtained.

6. The refined evaluation method for the stability of tunnel face according to claim 1, characterized in that The surrounding rock grades of each face region are determined based on the following method: Adopt image processing technology to extract the joint attitude, face roughness, surrounding rock color information and water leakage area information in the corresponding region, and determine the surrounding rock grades of each region.

7. The refined evaluation method for the stability of the tunnel face according to claim 1, wherein The ultimate deformation is obtained by calculating the ultimate deformation prediction model of the tunnel face. The model parameters of the ultimate deformation prediction model of the tunnel face are determined by training using machine learning methods. The ultimate deformation prediction model of the tunnel face characterizes the corresponding relationship between the tunnel burial depth, cross-sectional dimensions, and strength distribution of the tunnel face surrounding rock and the ultimate deformation when the tunnel face becomes unstable.

8. The refined evaluation method for the stability of tunnel face according to claim 7, characterized in that The training sample data used in training the extreme deformation prediction model of the tunnel face include tunnel burial depth value, tunnel rise-to-span ratio value, deformation characteristic values of surrounding rock areas of various grades, area percentages of surrounding rock areas of various grades, and extreme deformation values.

9. The refined evaluation method for the stability of tunnel face according to claim 1, characterized in that The allowable deformation is calculated based on the limit deformation, and the calculation formula is: Where Δ′ is the limit deformation, [Δ] is the allowable deformation, and [σ] is the safety factor.

Images